marketplace, reseller, or hybrid: strategic analysis of an
TRANSCRIPT
Marketplace Reseller or Hybrid Strategic Analysis ofan Emerging E-Commerce Model
Lin TianSchool of Management Fudan University Shanghai 200433 China ltian11fudaneducn
Asoo J VakhariaDepartment of Information Systems amp Operations Management Warrington College of Business Administration University of Florida
Gainesville Florida 32611-7169 USA asoovufledu
Yinliang (Ricky) TanA B Freeman School of Business Tulane University New Orleans Louisiana 70118 USA ytan2tulaneedu
Yifan XuSchool of Management Fudan University Shanghai 200433 China yfxufudaneducn
T raditionally online retailers have acted as product resellers Recently these retailers have also started to serve asonline marketplaces by providing a platform to directly connect sellers with buyers Over and above re-shaping the
traditional e-commerce market conventional wisdom suggests that this new format will mitigate the double-marginaliza-tion effect and benefit both the intermediary and suppliers through a revenue sharing scheme However we find thatupstream competition between suppliers critically moderates this possibility We also find that the interaction of order-ful-fillment costs and upstream competition intensity moderates the selection of an optimal mode for the intermediary Morespecifically when order-fulfillment costs are large and when the supplier product offerings are similar (ie competitionintensity is high) the pure reseller mode is the preferred choice when order-fulfillment costs are small and the supplierproduct offerings are highly differentiated (ie low competition intensity) the pure marketplace mode is the preferredchoice Finally the hybrid mode is preferred when order-fulfillment costs are moderate and suppliersrsquo products are some-what similar (ie competition intensity is moderate) The intuition behind these results hinges on the trade-off betweentransfer of pricing rights and the responsibility for order fulfillment Our findings not only complement the emergingonline marketplace literature but also provide testable empirical questions concerning the relationship and magnitude ofdifferent factors steering the mode choice
Key words online marketplace channel structure upstream competition order-fulfillment costHistory Received December 2017 Accepted April 2018 by Subodha Kumar after 3 revisions
1 Introduction
Online retailers who have traditionally acted as resel-lers are starting to offer online markeplace servicesproviding a direct connection between buyers andsellers The success of online marketplaces is welldocumented For example direct sales from suppliersaccounted for 45 of units sold on Amazoncom inthe second quarter of 2015 and leading Chineseonline retailer JDcom reports that sales through itsonline marketplace grew at a rate higher than its pro-jection at 40 Perhaps this is what has driven onlineretailers who were originally resellers to add anonline marketplace (eg Amazon JDcom Gome) toexpand their product offering options for the con-sumer However note that neither Amazon nor JDcom is moving forward to the pure marketplace
mode For example Amazon and its marketplace sell-ers offered a total of 353710754 different productsand Amazon itself carried 12231203 in May 20161
Further some online retailers (eg BestBuycom) arestill working as pure resellers More interestinglythere are also examples of retailers (such as Zapposcom) that started as online markeplaces but havenow switched completely to the reseller mode Theunderlying motivation for these alternative modes isnot always clearIn the reselling mode online retailers purchase
products from suppliers for a wholesale price thendetermine retail prices for consumers In contrast inthe online marketplace mode suppliers determineretail prices and share revenue with the online retai-ler The online retailer does not incur any directinventory or delivery costs but it loses the flexibility
1595
Vol 27 No 8 August 2018 pp 1595ndash1610 DOI 101111poms12885ISSN 1059-1478|EISSN 1937-5956|18|2708|1595 copy 2018 Production and Operations Management Society
of setting market prices Of course it is also possiblefor the online retailer to adopt a hybrid configurationmode that is for some products they act as resellersand for others they serve as an online marketplaceRegardless of the chosen channel mode online retail-ers often offer products from competing supplierswhich necessitates an integration of upstream compe-tition in the chain From an efficiency perspectivethere is also a need to consider costs related to inven-tory storage and transportation we refer to these asldquoorder-fulfillment costsrdquo According to a recent reportin the Wall Street Journal (Kapner 2014) the cost of ful-fillment operations can run as high as 25 of the salesrevenue The adopted channel mode also affectswhether the online retailer bears these costs (resellermode) or whether they are borne by the suppliers (on-line marketplace mode)These observations motivate the research of this
study which is to address the following general ques-tion Should online retailers operate as resellersonline marketplaces or adopt both modes of opera-tion Which mode would be preferred for suppliersThe focus of this study is to examine the desirabilityof these platform choices for the online intermediaryMore specifically we investigate the impact ofupstream (supplier) competition and fulfillment (in-ventory and delivery) costs on intermediary and sup-plier profitability across three mutually exclusiveplatform modes (a) Reseller Mode for both suppliers(b) Hybrid Mode where one supplier operates underthe reseller format while the other operates as a directseller through an online marketplace and (c) OnlineMarketplace Mode for both suppliers Our studyhelps to shed light on why different online platformsadopt different selling modes and determine the keydrivers of different proportional fees based on pro-duct categoryThere is an emerging stream of literature that dis-
cusses the importance of the strategic contract choicesbetween resellers and online marketplaces with afocus on trade-offs between these two distinct puremodes Prior research addressing this general ques-tion has focused on information asymmetry (Jianget al 2011) positive or negative cross-channel effects(Abhishek et al 2016) complementary relationshipsbetween products and devices (Hao and Fan 2014)online product reviews (Young et al 2014) controlrights of marketing activity (Hagiu and Wright 2015)and downstream competition (Tan et al 2016) Ourfocus is completely different as we provide insightsinto how both upstream competition and order-fulfill-ment costs moderate the choice of channel mode forthe online intermediaryThe key findings of our study can be summarized
as follows As opposed to the monopoly settingwhere the marketplace mode will dominate the
reseller mode our analysis reveals that upstreamcompetition intensity and order-fulfillment costs aretwo interacting forces steering the mode choice of theonline retailer Essentially in the reseller mode theintermediary can function as a moderator to alleviatethe fierce price competition in the online marketplaceSpecifically when the order-fulfillment cost is rela-tively small and the competition intensity is low thepure marketplace mode will become the preferredchoice when the order-fulfillment cost is relativelyhigh or the competition is sufficiently intense thenthe pure reseller mode is the preferred choice Thehybrid mode is the preferred choice when order-ful-fillment costs and competition intensity are both atmoderate levels The intuition hinges on the trade-offbetween the intensity of the price competition due tothe transfer of the pricing right and the responsibilityof bearing the order-fulfillment cost accompanied bythe mode changeOur findings also suggest that the revenue-sharing
scheme adopted by many online platforms (ie onlinemarketplace mode) does not always outperform thetraditional reseller model This result holds when theupstream competition is very intense or the order-ful-fillment cost is high This implies that depending uponupstream competition intensity and magnitude oforder fulfillment costs online intermediaries shoulddesign different contract terms for distinct product cat-egories For example instead of allowing the suppliersto sell on its website Amazon can become the resellerin certain highly competitive product categories suchas electronics On the other hand it is more profitablefor Amazon to serve as a marketplace for other lesscompetitive product categories such as fine art (ielong-tail product category) For print books (with highorder-fulfillment costs) Amazon should serve as areseller while for e-books (with low order-fulfillmentcosts) Amazon should choose to operate as an onlinemarketplace Thus the revenue sharing proportion(and associated bounds) are useful to encourage ordiscourage the adoption of the online marketplacemode by suppliersThe remainder of this study is organized as follows
In the next section we review and examine how ourstudy relates to relevant literature This is followed bya description of our model and analysis in sections 3and 4 respectively Section 5 presents the maininsights of our results Section 6 concludes the paperwith a summary and possible avenues for futureresearch All proofs can be found in the OnlineAppendix
2 Relevant Literature
This study contributes to the emerging research litera-ture that focuses on the strategic mode choice
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1596 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
between suppliers and online retailers and aims toreveal the trade-offs between the two distinct puremodes (ie reseller and online marketplace) One ofthe key aspects in this context is the cross-channeleffect This effect is a result of suppliers offeringgoods through multiple channels potentially leadingsales in one channel to impact sales in another chan-nel In the presence of spillovers between channelsdifferent selling modes will elicit different reactionsfrom suppliers because the sales through new chan-nels would impact suppliersrsquo sales in the traditionalchannel (Brynjolfsson et al 2009) Abhishek et al(2016) study a setting with one supplier and twoonline intermediaries and find that the online inter-mediary prefers to use the marketplace mode if thesales through the intermediary lead to a negativeeffect on demand of the supplierrsquos traditional channeland vice versa They also show that such preferencesare moderated by competition among intermediariesIn contrast in our study we examine competitionbetween suppliersHagiu and Wright (2015) focus on a single decision
variable the level of demand-enhancing marketingactivity which is controlled by suppliers in the puremarketplace mode but by the intermediary in thepure reseller mode They assume information asym-metry such that each supplier has information on theimpact of market activities on demand while theintermediary and other suppliers do not Their resultsindicate that the pure marketplace mode is preferredif and only if the variance of the private informationexceeds the squared value of spillovers from market-ing activities across products In their model suppli-ers are only charged a fixed membership fee in thepure marketplace mode However in reality there isusually a small fixed participation fee plus a largeproportional revenue sharing fee By contrast weendogenize the revenue sharing fee and examinehow the online intermediary can manipulate this feeto encourage or discourage the adoption of the onlinemarketplace mode for the suppliersJiang et al (2011) also consider the impacts of
demand-enhancing marketing activities and informa-tion asymmetry They focus on the strategic under-selling behavior of sellers on platforms rather thanthe strategic mode choice we consider in this studyOther issues of interest have been explored in priorresearch on online marketplaces Hao and Fan (2014)focus on the publishing industry and show that e-book prices are lower in the pure reseller mode due tothe existence of a complementary market (ie e-bookreaders) Young et al (2014) show that channel modechoice can be used as a strategic tool to benefit fromthird-party product reviews More recently Tan et al(2016) and Tan and Carrillo (2017) illustrate that themechanism of the online marketplace can benefit both
the upstream supplier and the retailer in the digitalpublishing industryOur study differs from the aforementioned studies
in four key aspects First we explicitly considerupstream supplier competition in identifying an equi-librium configuration mode for the online retailer Webelieve this is a critical aspect since most online retail-ers simultaneously offer substitute products fromcompeting suppliers Second since the configurationmode impacts the order-fulfillment costs borne by dif-ferent supply chain members we also integrate thiscritical aspect in our setting As indicated earlierthese costs are a significant portion of the total chan-nel costs Third practice indicates that some onlineretailers operate as both resellers and online market-places for the same product category This drives usto explicitly include such a hybrid configurationmode in our comparisons Fourth we endogenize theequilibrium mode choice within the interactionsbetween the suppliers and the online intermediaryand also the proportional fee for the suppliers underthe online marketplace modeOur study also contributes to the research literature
on channel structure Early work on market channelsconcerned whether suppliers are better off using inde-pendent intermediaries instead of vertical integration(eg Choi 1991 Coughlan 1985 McGuire and Staelin1983 Trivedi 1998) A common result is that the exis-tence of independent intermediaries is able to miti-gate the competition among suppliers For exampleMcGuire and Staelin (1983) consider the partial sub-stitutability between two products from two suppliersselling through exclusive retailers They conclude thatsuppliers prefer to use independent intermediarieswhen the degree of substitutability is high and prefercompany-owned stores otherwise Coughlan (1985)discusses the problem of choosing a vertical market-ing channel in a product-differentiated duopoly mar-ket and shows that integration of the marketingfunction results in greater price competition andlower prices than the use of independent marketingmiddlemenAnother stream of research on channel structure
examines the impact of introducing an online directchannel in addition to an existing retail channel Chi-ang et al (2003) have argued that the introduction ofthe direct channel may benefit the retailer through areduction of the wholesale price Tsay and Agrawal(2004a) have confirmed that result in a different set-ting by considering various sources of inefficiencyCai (2010) and Cai et al (2012) explore the impact ofmultiple channel structures and revenue sharingschemes to the supplier the retailer and the supplychain Recent research has considered the rationale ofselling through online platforms For example Ryanet al (2012) consider a setting with a supplier who
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1597
sells goods through its own direct sale website andhas the option of selling its goods through an onlineplatform They assume that selling through the plat-form system can attract more customers to the sup-plierrsquos own website but at some expense eg a fixedparticipation fee or proportional fee charged by theonline platform Interested readers are referred toCattani et al (2004) and Tsay and Agrawal (2004b) formore comprehensive reviewsOur study on strategic mode choice between sup-
pliers and online retailers differs from the stream ofmarket channel literature in three ways First in priorresearch the online retailerintermediary is assumedto function as a reseller In our study under the onlinemarketplace mode the intermediary acts as an agentcharging a proportional fee for each sale Shy andWang (2011) suggest that such a fee could be regardedas a revenue sharing mechanism that may mitigatedouble marginalization Second the focus of ourstudy is identifying which selling mode is more bene-ficial for online intermediaries rather than a compar-ison of directindirect selling activities Third weexplicitly examine how the order-fulfillment cost andupstream competition two significant but unexploredfactors determine the mode configuration choiceOur research also complements the stream of litera-
ture on store-within-store in an offline setting (egJerath and Zhang 2010 Netemeyer et al 2012) whereretailers essentially rent out their retail space to sup-pliers and give them complete autonomy over retaildecisions like pricing and in-store services Our studyof online marketplaces differs from this literaturestream in two ways First in store-within-store con-tracts suppliers typically manage all retail decisionsand the retailers charge them a fixed periodic rent(Jerath and Zhang 2010) Our work suggests that aproportionalrevenue sharing fee could be a viablealternative in this setting Through guidelines on pro-portional fees (a) (ie the revenue sharing proportionthat the online retailers can keep) we show thatonline retailers can set the value of a to encourage ordiscourage the adoption of the online marketplacemode for the suppliers As noted by Cachon and Lari-viere (2005) revenue sharing contract is very difficultto implement in the offline setting However this isnot the case for the online marketplace setting whichis analogous to the store-within-store model Secondwe show that order-fulfillment costs and upstreamcompetition jointly determine the optimal mode con-figuration choice for online retailers To the best ofour knowledge the literature on the store-within-store setting has largely neglected the impacts of thesetwo significant but unexplored factorsIn summary the key aspects of our paper (eg the
proportional-feerevenue-sharing fee the order-ful-fillment cost and the upstream competition between
suppliers) have not been thoroughly explored in theextant store-within-store literature Thus our findingsregarding the online marketplace can also provideuseful insights into store-within-store settings In thenext section we turn to a description of our modelingframework
3 Modeling Framework
We consider a stylized supply chain consisting of twocompeting suppliers (A and B) selling two substi-tutable products (A sells product a B sells product b)through a common online intermediary (I) Given thatthe products are substitutable we follow establishednorms in the marketing and operations literature(eg Birge et al 1998 Choi 1991 Gal-Or et al 2008Garcia-Gallego and Georgantzis 2001 Li and Zhang2008 McGuire and Staelin 1983) and our demandfunction is2
da frac14 h pa thorn cethpb paTHORN eth1THORNdb frac14 h pb thorn cethpa pbTHORN eth2THORN
where di and pi refer to the realized demand andretail prices for product i(=a b) respectively and hdenotes the market potential c gt 0 is the measure ofthe intensity of price competition with higher valuesindicating a greater degree of product substitutionand greater intensity of price competition3 Ourentire analysis without loss of generality assumesthat production costs for suppliers are normalizedto be equal to zero For ease of exposition we usethe pronoun ldquoherdquo to represent the intermediary andldquosherdquo to denote the suppliers in the remainder ofthe studyThree alternative mode choices evaluated are as
follows
bull Reseller Mode RR Under mode RR the inter-mediary acts as a reseller for both suppliersSince many online intermediaries (eg Ama-zon and JDcom) were historically pure resel-lers we set RR mode to be the benchmark(current) mode
bull Hybrid Mode PR This mode represents a set-ting where the intermediary acts as a market-place for one supplier (suppose Supplier A) byspecifying a proportional fee a and acts as areseller for the other supplier (Supplier B)
bull Online Marketplace Mode PP In direct con-trast to mode RR under this mode the inter-mediary specifies a proportional fee a to bothsuppliers and acts as a marketplace for bothsuppliers
Following industry practice the intermediary actsas a Stackelberg leader in terms of whether she will
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1598 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
act as a reseller or marketplace for one or both suppli-ers The suppliers as followers accept or reject theintermediaryrsquos offer In determining the equilibriumoutcomes for each alternative mode the conditionsunder which suppliers will accept the intermediaryrsquosmode choice are explicitly incorporatedBased on this discussion it is apparent that the pos-
sible existence of each mode of operation is driven bythe intermediaryrsquos operational offering (eg the pro-portional fee a) and the suppliersrsquo willingness toaccept this offer Assuming the existence of eachmode the interaction between the intermediary andsuppliers is illustrated in Figure 1 The sequence ofevents for each mode of operation is as follows
bull Mode RR In this benchmark (current) mode theintermediary has decided not to offer the onlinemarketplace service and acts as a reseller for bothsuppliers Suppliers start by simultaneouslyoffering wholesale prices wa and wb to the inter-mediary who in turn simultaneously sets retailprices pa and pb for the consumers
bull Mode PR In this mode the intermediary hasoffered to act as a marketplace for Supplier A byspecifying a proportional fee a and still acts as areseller for Supplier B Assuming Supplier Aaccepts the online marketplace offer and Sup-plier B chooses to continue participating under
the reseller format then based on the quotedwholesale price wb from Supplier B both Sup-plier A and the intermediary simultaneously setthe retail prices pa and pb respectively
bull Mode PP In this mode the intermediary hasoffered the online marketplace service to bothsuppliers by specifying the proportional fee aAssuming both suppliers accept the offer thenboth of them determine the retail prices pa andpb simultaneously
Note that under the marketplace service the inter-mediary yields the pricing power to the supplier Theintermediary receives a commission in proportion tothe supplierrsquos revenue at a rate of a which is deter-mined by the intermediary All three modes are ana-lyzed under a complete information settingThe proportion a is specified as a referral fee by
some online intermediaries (Geng et al 2018) Thisproportional fee although different across productcategories is the same for all products within a cer-tain category Empirical data indicates that common avalues range from 6 to 25 of the sale price depend-ing on the product category on Amazon while forJDcom the fee for most product categories rangesfrom 5 to 12 In addition to this proportional feesome intermediaries also charge a fixed subscription
Figure 1 Channel Modes RR PR and PP [Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1599
fee For example Amazon Marketplace charges a$3999 monthly subscription fee for sellers who planto sell more than 40 items a month Searscom alsocharges a $3999 fee to suppliers whose sales are over$400 during one month Because this fee is relativelysmall compared with the suppliersrsquo sales volumewithout loss of generality we normalize this fixed feepart to zero Previous literature has also adopted thisassumption (Abhishek et al 2016)Order fulfillment that is delivering physical goods
to the customer is commonly cited as one of the mostexpensive and critical operations of online sellers(Agatz et al 2008) The cost of order fulfillment canrun as high as 25 of sales (Kapner 2014) To fulfill anorder firms take on costs such as warehouse build-ingrenting costs hiring staff to handle packages anddelivering the products to customers The costs of stor-age and hiring staff are quite significant and can beviewed as fixed cost Meanwhile the delivery cost foronline shopping is usually undertaken by the cus-tomers For ease of exposition we assume that order-fulfillment costs are fixed since we find that the keyinsights stemming from our analysis are similar evenif these costs are a function of the number of ordersreceived4 Essentially if the intermediary functions asa reseller he will incur a fixed cost FI gt 0 to fulfill theorder for per product category (product a or b) if theintermediary functions as an online marketplace eachsupplier must be responsible for a fixed cost FS gt 0 tofulfill orders Through analysis we find that if theintermediaryrsquos order-fulfillment cost is lower than thesupplierrsquos order-fulfillment cost (ie FI lt FS) the equi-librium mode is more likely to be the pure resellermode and vice versa To eliminate the possibility thatthe equilibrium mode is driven by the asymmetric coststructure we assume FI = FS = F in our base modelMore specifically in mode RR the intermediary bearsorder-fulfillment costs for both product categories (ie2F) in mode PR Supplier A bears the order-fulfillmentcost (F) for its product and the intermediary bears theorder-fulfillment cost (F) for reselling Supplier Brsquos pro-duct in mode PP both Supplier A and Supplier B bearthe order-fulfillment cost (F) of their own product cate-gory5 Recently some online intermediaries (egAmazon Sears JD etc) and 3rd party online order ful-fillment service providers have offered an option tosuppliers that allows them to outsource their order ful-fillment In this case the supplier will pay some feesfor order fulfillment that is even if the supplier out-sources the order fulfillment he still needs to bear therelated costs For ease of exposition and without lossof generality we focus on the case that suppliers willfulfill the order in-house in our base model6
In the next section the equilibrium outcomes ofeach channel mode are structurally characterized
4 Equilibrium Analysis
The first set of results described below is our bench-mark (current) setting where the intermediary acts asa reseller for both suppliers (mode RR) Next we con-sider the setting where one supplier switches to theonline marketplace mode while the other supplier stilladopts the reseller mode (model PR) Finally wefocus on the case in which both suppliers choose toaccept the intermediary terms and switch to theonline marketplace mode (mode PP) In derivingthese results the necessary conditions for the exis-tence of each operating mode (ie RR PR and PP)are that the equilibrium profits for each supplier areat least as large as the profits each supplier could real-ize through exercising an outside option (denoted bypO)
41 Mode RRUnder this setting the suppliers simultaneouslyquote wholesale prices (wa and wb) and then the inter-mediary (who bears the order-fulfillment costs forboth products) determines the retail prices pa and pbThe profits for the suppliers and the intermediary areas follows
pAethRRTHORN frac14 wafrac12h pa thorn cethpb paTHORNpBethRRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethRRTHORN frac14 ethpa waTHORNfrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN 2F
We solve this game by backward induction Forany given wholesale prices wa and wb we first charac-terize the equilibrium retail prices that would maxi-mize pI(RR) We then determine the wholesale pricesfor the suppliers by simultaneously maximizing theirindividual profit functions
LEMMA 1 There exists a unique equilibrium for modeRR The equilibrium prices and demands are as follows
wRRa frac14 wRR
b frac14 1
2thorn ch
pRRa frac14 pRRb frac14 3thorn c2eth2thorn cTHORN h
dRRa frac14 dRRb frac14 1thorn c2eth2thorn cTHORN h
and the corresponding optimal profits are given by
pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2
pRRI frac14 eth1thorn cTHORN22eth2thorn cTHORN2 h
2 2F
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1600 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
It is straightforward to note that an increase incompetition intensity (c) would lead to an increasein the intermediaryrsquos profit and a simultaneousdecrease in supplier profits since an increase in creduces the suppliersrsquo pricing power relative tothat of the intermediary The profit for each sup-plier under this RR benchmark mode serves as abase for evaluating whether a supplier wouldchoose the online marketplace if it was offered bythe intermediaryThe technical condition for the existence of the
RR mode and its corresponding equilibrium is thatthe supplierrsquos profit under this mode is at least aslarge as the profit that each supplier could realizeby exercising the outside option that ispO pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2 Given that the reseller
mode is adopted by a large number of suppliersselling through online intermediaries this exis-tence condition is assumed to hold for the remain-der of this study
42 Mode PRIn this setting only one supplier will be offeredthe option of the online marketplace mode fromthe intermediary Hence one supplier adopts theonline marketplace mode if accepted while theother retains the reseller mode Without loss ofgenerality we assume that the online marketplaceoption is accepted by Supplier A and the interme-diary operates as a reseller for Supplier B For agiven proportion a the suppliersrsquo and intermedi-aryrsquos profits are
pAethPRTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethPRTHORN frac14 apafrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN F
We solve this game by backward induction as fol-lows For any given wb we first characterize thesimultaneous pricing decision where Supplier A deci-des the retail price of product a (ie pa) to maximizepA(PR) and the intermediary decides the retail priceof product b (ie pb) to maximize pI(PR) Next wedetermine Supplier Brsquos wholesale price wb to maxi-mize pB(PR) Finally using this wholesale price wbthe optimal retail prices pa and pb can be set Theresults of this analysis are presented in the lemmabelow
LEMMA 2 Given a proportion a and Supplier A accept-ing the offer of an online marketplace while the intermedi-ary serves as a reseller for Supplier B there exists aunique equilibrium with prices and demands are asfollows
wPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc
2eth1thorn cTHORNeth2thorn 4cthorn c2THORN h
pPRa frac14 eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORNh
pPRb frac14 eth3thorn 6cthorn 2c2THORNeth2thorn 3cTHORNthorn aeth1thorn c c2THORNcfrac124eth1thorn cTHORN2eth1thorn aTHORNc2eth2thorn 4cthorn c2THORN h
dPRa frac14 eth4thorn 9cthorn 3c2THORNeth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn cTHORNeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORN h
dPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2 h
and the corresponding equilibrium profits are
pPRA frac14
eth1aTHORNeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2F
pPRB frac14 frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc22eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc2eth4thorn8cthorn2c2THORNh
2
pPRI frac14aeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22
frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2
thorn
frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc2eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN
Mh2F
where
M frac14 frac12eth1thorn cTHORNeth2thorn 3cTHORNeth2thorn 4cthorn c2THORN thorn aeth1thorn cTHORNeth2thorn 3cTHORNc2thorn aeth3thorn 4cTHORNeth2thorn 4cthorn c2THORNc a2eth1thorn 2cTHORNc3
For the equilibrium in Lemma 2 to exist the follow-ing proposition characterizes the sufficient conditionsunder which Supplier A will accept the intermedi-aryrsquos marketplace offer and Supplier B will continueto participate under the reseller mode
PROPOSITION 1 Supplier A will accept the intermediaryrsquosoffer of an online marketplace provided pPRA pRRA andSupplier B will continue to participate under the resellermode provided pPRB pO This leads to the following suffi-cient conditions for the existence of the PR mode
bull F le FPR where FPR frac14 eth4thorn 9cthorn 3c2THORN2eth1thorn cTHORNeth2thorn cTHORN2eth4thorn 8cthorn 2c2THORN2 h
2 1thorn c
2eth2thorn cTHORN2 h2 and
bull a aPR frac14 minfrac12aPRA aPRB where aPRA is theunique solution of eth1 aPRA THORNeth1thorn cTHORNfrac12eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aPRA eth1thorn 2cTHORNc22
frac124eth1thorn cTHORN2 eth1thorn aPRATHORNc22eth4thorn 8cthorn 2c2THORN2 h2 F frac14 1thorn c
2eth2thorn cTHORN2
h2 and aPRB is the unique solution offrac12eth1thorn cTHORNeth2thorn 3cTHORN aPRB eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPRBTHORNc2eth4thorn 8cthorn 2c2THORN h
2 frac14 pO
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1601
Proposition 1 characterizes the conditions underwhich Supplier A will accept the intermediaryrsquos mar-ketplace offer and Supplier B will not pursue the out-side option but continue to participate under thereseller mode The first condition relates to the order-fulfillment cost which should not exceed a certainthreshold Recall that by choosing the online market-place service the supplier gains pricing power butbears the order-fulfillment cost at the same timeThus this threshold value of the fulfillment costreflects the trade-off between gains from the pricingpower and the costs (in terms of order fulfillment) ofrealizing such gainsAssuming the first condition is satisfied the propo-
sition then specifies a second condition related to themaximum proportion a which can be charged to thesupplier by the intermediary Interesting insightsregarding this proportion a are shown in Lemma 3
LEMMA 3 (i) pPRI increases in a whereas both pPRA andpPRB decrease in a (ii) aPR decreases in F
Lemma 3(i) indicates that an increase in a willincrease the intermediaryrsquos profit pPRI and simultane-ously decrease Supplier Arsquos profit pPRA Hence ashould be small (ie a aPRA ) to ensure pPRA pRRA sothat Supplier A will accept the intermediaryrsquos market-place offer Meanwhile a counterintuitive result isthat an increase in a would also lead to a reduction inthe profit of the other supplier (B) who has not beenoffered the online marketplace service The reason forthis negative spillover effect is that when a isincreased the supplier (A) tends to increase the priceas the intermediary will keep a higher proportion ofthe revenue More specifically when the intermediaryraises the proportion a she can keep Supplier Aresponds by charging higher retail prices In responsethe intermediary sets a higher retail price for SupplierBrsquos product leading to lower demand for the sameproduct The net result is lower profit for Supplier BThus a needs to be small enough (a aPRB ) so thatSupplier B will not pursue the outside option but con-tinue to participate under the reseller mode orpPRB pO7 In equilibrium assuming the conditionrelated to the order-fulfillment cost is met (ieF le FPR) then the intermediary will set a frac14 aPR frac14minfrac12aPRA aPRB since this will allow the intermediary toextract the maximum profits from the supplier Fur-ther Lemma 3(ii) shows that this threshold value ofthe proportional fee (ie aPR) is decreasing in theorder-fulfillment cost (F)In summary from a managerial perspective our
results indicate that the intermediary requires astrategic focus in choosing the parameters for offeringthe online marketplace service This service will onlybe attractive to suppliers provided the order-
fulfillment cost is not significantly high If this is thecase then the intermediary should set the proportiona equal to its upper bound in order to realize the max-imum profits From a supplier perspective the onlinemarketplace offering is preferred over the reseller set-ting when the flexibility gains due to price setting out-weigh the explicit costs of order fulfillment
43 Mode PPIn this setting both suppliers will be offered theoption of online marketplace from the intermediaryIf both suppliers choose the online marketplace ser-vice they will take into consideration the proportiona which they need to share with the intermediary Fora given proportion a the profits for each supplier andthe intermediary under this mode (ie both supplierschoose the online marketplace service) are as follows
pAethPPTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPPTHORN frac14 eth1 aTHORNpbfrac12h pb thorn cethpa pbTHORN F
pIethPPTHORN frac14 apafrac12h pa thorn cethpb paTHORNthorn apbfrac12h pb thorn cethpa pbTHORN
We solve this game by simultaneously determiningthe retail prices pa and pb which maximize pA(PP)and pB(PP) The results are shown in the lemmabelow
LEMMA 4 Assuming both suppliers accept the propor-tion a there exists a unique equilibrium with retail pricesand demands are as follows
pPPa frac14 pPPb frac14 1
2thorn ch
dPPa frac14 dPPb frac14 1thorn c2thorn c
h
The corresponding equilibrium profits are given by
pPPA frac14 pPPB frac14 eth1 aTHORN 1thorn c
eth2thorn cTHORN2 h2 F
pPPI frac14 2a1thorn c
eth2thorn cTHORN2 h2
When is this mode a possible outcome This can beframed using the decision matrix in Table 1 where weare interested in the equilibrium when both supplierswill accept the online marketplace offer This equilib-rium will arise when both of the following conditionshold Given that Supplier B (A) adopts the resellermodel Supplier Arsquos (Brsquos) profit should be higherunder the condition of acceptance than refusal of theintermediaryrsquos marketplace offer which is pPRA pRRA(pRPB pRRB ) given that Supplier B (A) adopts theonline marketplace Supplier Arsquos (Brsquos) profit should
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1602 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
also be higher when it accepts rather than refuses theintermediaryrsquos marketplace offer which is pPPA pRPA(pPPB pPRB ) In addition to ensure this equilibrium isstable both suppliersrsquo profits under the PP modeshould not be lower than the profit under the outsideoption which is pPPA pO (pPPB pO) Based on theresults in Lemma 4 the following proposition identi-fies the conditions under which both suppliers willnot pursue the outside option but choose the onlinemarketplace service offering
PROPOSITION 2 Both suppliers will choose the onlinemarketplace service provided pPRA pRRA pPPA pRPA andpPPA pO This leads to the following sufficient condi-tions for the existence of the PP mode
bull F le FPP where FPP frac14 1thorn ceth2thorn cTHORN2 h
2 eth1thorn cTHORNeth2thorn 3cTHORN2eth2thorn cTHORNeth4thorn 8cthorn 2c2THORN
h2 and
bull a aPP frac14 minfaPRA aPPA aPPB g where aPPA is the
unique solution of eth1 aPPA THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14frac12eth1thorn cTHORNeth2thorn 3cTHORN aPPA eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPPATHORNc2eth4thorn 8cthorn 2c2THORN h
2 and aPPB is the
unique solution of eth1 aPPB THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14 pO
Similar to the hybrid mode both suppliers willaccept the intermediaryrsquos online marketplace offerwhen the order-fulfillment cost and the proportionalfee that the intermediary charges are all below a cer-tain threshold It is worth noting that the threshold inorder-fulfillment cost is more stringent than that statedin Proposition 1 since FPP lt FPR As before the inter-mediaryrsquos profit pPPI increases in a while both pPPA andpPPB decrease in a In equilibrium the intermediary isonce again incentivized to set a = aPP which decreasesin the order-fulfillment cost (F) We also note that sup-pliersrsquo profits will decrease as the upstream competi-tion intensifies (ie a higher value of c)In the next section we compare the equilibrium
solutions for the three modes with a view to provid-ing managerial insights
5 Insights
51 Pricing and DemandFrom a pricing and demand perspective the proposi-tion below provides a comparison across the threechannel modes
PROPOSITION 3 Comparing the equilibrium results ofretail prices and demand we have (i) pRRa frac14pRRb [ pPRb [ pPRa [ pPPa frac14 pPPb (ii) dPRa [ dPPa frac14dPPb [ dRRa frac14 dRRb [ dPRb
Interestingly the results of Proposition 3 are inde-pendent of the proportional fee rate a and hold regard-less of the parameter settings for competitionintensity (c) and maximum market size (h) Essen-tially market prices are highest in the pure resellermode since they are set by the intermediary whoattempts to moderate the competition between thesuppliers On the other hand when the suppliers setthe retail prices by operating in the online market-place mode they compete with each other directlyleading to lower retail prices Because of doublemarginalization in mode PR the retail price of theproduct that does not join the online marketplace ishigher than the price for which the supplier whochooses to do so Corresponding to these pricingstructures demand coverage is higher in the onlinemarketplace mode (PP) as compared to the resellermode (RR)
52 Intermediary ProfitsWe now proceed to examine how intermediary prof-itability across the three modes is affected by two keyparameters competition intensity (c) and order-fulfill-ment cost (F)8 To assess how these parameters affectprofitability of all supply chain members across thethree modes we start by making the followingassumptions First we assume that the order-fulfill-ment cost is small enough to ensure that both modePR and PP are feasible Analytically the potential val-ues of F under this assumption are that F lt FPP(ltFPR)Note that the value of F which satisfies the firstassumption also leads both aPR and aPP to be positiveOur second and final assumption is that the interme-diary chooses to set a = aPR for mode PR and a = aPP
for mode PP which represent equilibrium outcomesProposition 4 structurally characterizes which
mode choice would be preferred by the intermedi-ary as moderated by the competition intensityparameter c
PROPOSITION 4 There exist competition intensity thresh-olds c gt 0 and c(gtc) such that
bull If c lt c the optimal mode is PP
bull If c gt c the optimal mode is RR and
bull If c le c le c mode PR can be the optimal mode
Proposition 4 indicates that when competitionintensity is low the optimal mode is PP when compe-tition intensity is high the optimal mode is RR Inaddition we observe that when competition intensityis medium the optimal mode is PR The rationale
Table 1 Decision Matrix with Resulting Profits
Supplier A option
Supplier B option
Online marketplace Reseller
Online marketplace pPPA pPPB pPRA pPRB
Reseller pRPA pRPB pRRA pRRB
Note Due to symmetry we know that pPRA frac14 pRPB and pPRB frac14 pRPA
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1603
behind these results hinges on the moderating role ofthe intermediary Previous literature has shown thatthe intermediary prefers high levels of upstream com-petition as it strengthens its channel power (Wangand Shin 2015) From a mode perspective supplierscompete more fiercely in the online marketplacemode compared with the reseller mode as they set theprices directly and alternative modes allow the inter-mediary to moderate this competitive effect To elabo-rate when both suppliers are very competitive theintermediary would prefer to operate as a resellersince any other mode will lead to increased price com-petition which in turn negatively impacts the inter-mediary This result validates the conjecture of Hagiu(2007) that strong substitutability between suppliersrsquoproducts would lead to a greater preference for thepure reseller mode When the competition intensitybetween suppliers is low the intermediaryrsquos moderat-ing role is not as important Note that in mode PP theproportional fee could be considered a form of rev-enue sharing which would mitigate the impact ofdouble-marginalization (Geng et al 2018 Shy andWang 2011 Tan and Carrillo 2017) Hence the inter-mediary prefers mode PP when competition intensityis low When the competition intensity between sup-pliers is in the intermediate range the intermediaryrsquosmoderating role is important but not significantenough to alleviate double-marginalization effectsleading to a preference for the hybrid mode PRAnecdotal observations provide face validity for
our results For the household appliance market inChina the similarity between branded products leadsto strong substitutability (or high competition inten-sity) Hence most online retailers (eg JDcom) preferto act as resellers for products in this category On theother hand the media content industry (publishingvideo games) is characterized by low levels of pro-duct substitutability (or low competition intensity)This has motivated online retailers to set the terms forthe online marketplace mode in such a way that sup-pliers have chosen that mode of operation Etsycomthe largest e-commerce website in handmade or vin-tage items serves as a pure online marketplacebecause of low product substitutability FurtherAmazon serves as the reseller in certain high competi-tive product categories such as electronics but servesas a marketplace for other less competitive productcategories such as fine art (ie long-tail productcategory)Next we consider the impact of the order-fulfill-
ment cost on the choice of equilibrium mode for theintermediary Note that Proposition 4 shows thatwhen the market competition intensity is above a cer-tain threshold the intermediary will always prefer toact as a reseller To investigate the impact of theorder-fulfillment cost next we focus on the case that
the market competition intensity is not too high Theproposition below characterizes the mode preferredby the intermediary depending on the magnitude ofthe order-fulfillment cost
PROPOSITION 5 When the market competition intensityis relatively small (ie c cjFfrac140) there exists twothresholds F gt 0 and F(geF) such that
bull when F le F the preferred mode is PP and
bull when F lt F le F the preferred mode is PR and
bull when F gt F the preferred mode is RR
Proposition 5 shows that the magnitude of theorder-fulfillment cost will alter the mode choicesWhen the market competition intensity is relativelylow (c cjFfrac140) as the order-fulfillment costincreases the equilibrium mode will evolve from apure marketplace mode to a hybrid mode then to apure reseller mode The intuition of this result lies inthe fact that the higher the order-fulfillment costbecomes the more difficult it becomes for the inter-mediary to convince the suppliers to adopt the mar-ketplace mode and endure the order-fulfillment costRecall that both aPR and aPP decrease in F That is asthe order-fulfillment cost increases the intermediaryhas to set a much lower proportional fee to ensurethat the suppliers will accept the marketplace offerThis becomes increasingly expensive as the order-ful-fillment cost increases As a result the intermediaryprefers the marketplace mode for both supplierswhen the order-fulfillment cost is low and favors thereseller mode when the order-fulfillment cost issignificantFrom a practical perspective we observe that most
hotel and airline companies use the PP mode on tra-vel sites such as Expedia Travelocity Ctrip andQunar (Ctrip and Qunar are leading travel sites inChina) This is partially driven by the low (almostnegligible) order-fulfillment cost for electronic tick-etsbookings In the publishing industry a transitionto the PP mode from the RR mode is evident for e-books while the RR mode is still preferred for printbooks This could be due in part to the fact that order-fulfillment costs are greater for printed books as com-pared to e-booksHaving analytically examined the impacts of the
competition intensity and order-fulfillment cost indi-vidually we next numerically illustrate how the inter-actions between these two parameters affect the modechoice for the intermediary Figure 2 was generatedby setting h = 10 (ie the market potential equals to10) and pO = 0 (ie the outside option facing thesupplier is 0) Then we simultaneously varyF (F 2 eth0 FPPjcfrac140THORN where FPPjcfrac140 frac14 125) and c(c 2 (0 2)) Note that our results are robust to theparameter changes
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1604 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
of setting market prices Of course it is also possiblefor the online retailer to adopt a hybrid configurationmode that is for some products they act as resellersand for others they serve as an online marketplaceRegardless of the chosen channel mode online retail-ers often offer products from competing supplierswhich necessitates an integration of upstream compe-tition in the chain From an efficiency perspectivethere is also a need to consider costs related to inven-tory storage and transportation we refer to these asldquoorder-fulfillment costsrdquo According to a recent reportin the Wall Street Journal (Kapner 2014) the cost of ful-fillment operations can run as high as 25 of the salesrevenue The adopted channel mode also affectswhether the online retailer bears these costs (resellermode) or whether they are borne by the suppliers (on-line marketplace mode)These observations motivate the research of this
study which is to address the following general ques-tion Should online retailers operate as resellersonline marketplaces or adopt both modes of opera-tion Which mode would be preferred for suppliersThe focus of this study is to examine the desirabilityof these platform choices for the online intermediaryMore specifically we investigate the impact ofupstream (supplier) competition and fulfillment (in-ventory and delivery) costs on intermediary and sup-plier profitability across three mutually exclusiveplatform modes (a) Reseller Mode for both suppliers(b) Hybrid Mode where one supplier operates underthe reseller format while the other operates as a directseller through an online marketplace and (c) OnlineMarketplace Mode for both suppliers Our studyhelps to shed light on why different online platformsadopt different selling modes and determine the keydrivers of different proportional fees based on pro-duct categoryThere is an emerging stream of literature that dis-
cusses the importance of the strategic contract choicesbetween resellers and online marketplaces with afocus on trade-offs between these two distinct puremodes Prior research addressing this general ques-tion has focused on information asymmetry (Jianget al 2011) positive or negative cross-channel effects(Abhishek et al 2016) complementary relationshipsbetween products and devices (Hao and Fan 2014)online product reviews (Young et al 2014) controlrights of marketing activity (Hagiu and Wright 2015)and downstream competition (Tan et al 2016) Ourfocus is completely different as we provide insightsinto how both upstream competition and order-fulfill-ment costs moderate the choice of channel mode forthe online intermediaryThe key findings of our study can be summarized
as follows As opposed to the monopoly settingwhere the marketplace mode will dominate the
reseller mode our analysis reveals that upstreamcompetition intensity and order-fulfillment costs aretwo interacting forces steering the mode choice of theonline retailer Essentially in the reseller mode theintermediary can function as a moderator to alleviatethe fierce price competition in the online marketplaceSpecifically when the order-fulfillment cost is rela-tively small and the competition intensity is low thepure marketplace mode will become the preferredchoice when the order-fulfillment cost is relativelyhigh or the competition is sufficiently intense thenthe pure reseller mode is the preferred choice Thehybrid mode is the preferred choice when order-ful-fillment costs and competition intensity are both atmoderate levels The intuition hinges on the trade-offbetween the intensity of the price competition due tothe transfer of the pricing right and the responsibilityof bearing the order-fulfillment cost accompanied bythe mode changeOur findings also suggest that the revenue-sharing
scheme adopted by many online platforms (ie onlinemarketplace mode) does not always outperform thetraditional reseller model This result holds when theupstream competition is very intense or the order-ful-fillment cost is high This implies that depending uponupstream competition intensity and magnitude oforder fulfillment costs online intermediaries shoulddesign different contract terms for distinct product cat-egories For example instead of allowing the suppliersto sell on its website Amazon can become the resellerin certain highly competitive product categories suchas electronics On the other hand it is more profitablefor Amazon to serve as a marketplace for other lesscompetitive product categories such as fine art (ielong-tail product category) For print books (with highorder-fulfillment costs) Amazon should serve as areseller while for e-books (with low order-fulfillmentcosts) Amazon should choose to operate as an onlinemarketplace Thus the revenue sharing proportion(and associated bounds) are useful to encourage ordiscourage the adoption of the online marketplacemode by suppliersThe remainder of this study is organized as follows
In the next section we review and examine how ourstudy relates to relevant literature This is followed bya description of our model and analysis in sections 3and 4 respectively Section 5 presents the maininsights of our results Section 6 concludes the paperwith a summary and possible avenues for futureresearch All proofs can be found in the OnlineAppendix
2 Relevant Literature
This study contributes to the emerging research litera-ture that focuses on the strategic mode choice
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1596 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
between suppliers and online retailers and aims toreveal the trade-offs between the two distinct puremodes (ie reseller and online marketplace) One ofthe key aspects in this context is the cross-channeleffect This effect is a result of suppliers offeringgoods through multiple channels potentially leadingsales in one channel to impact sales in another chan-nel In the presence of spillovers between channelsdifferent selling modes will elicit different reactionsfrom suppliers because the sales through new chan-nels would impact suppliersrsquo sales in the traditionalchannel (Brynjolfsson et al 2009) Abhishek et al(2016) study a setting with one supplier and twoonline intermediaries and find that the online inter-mediary prefers to use the marketplace mode if thesales through the intermediary lead to a negativeeffect on demand of the supplierrsquos traditional channeland vice versa They also show that such preferencesare moderated by competition among intermediariesIn contrast in our study we examine competitionbetween suppliersHagiu and Wright (2015) focus on a single decision
variable the level of demand-enhancing marketingactivity which is controlled by suppliers in the puremarketplace mode but by the intermediary in thepure reseller mode They assume information asym-metry such that each supplier has information on theimpact of market activities on demand while theintermediary and other suppliers do not Their resultsindicate that the pure marketplace mode is preferredif and only if the variance of the private informationexceeds the squared value of spillovers from market-ing activities across products In their model suppli-ers are only charged a fixed membership fee in thepure marketplace mode However in reality there isusually a small fixed participation fee plus a largeproportional revenue sharing fee By contrast weendogenize the revenue sharing fee and examinehow the online intermediary can manipulate this feeto encourage or discourage the adoption of the onlinemarketplace mode for the suppliersJiang et al (2011) also consider the impacts of
demand-enhancing marketing activities and informa-tion asymmetry They focus on the strategic under-selling behavior of sellers on platforms rather thanthe strategic mode choice we consider in this studyOther issues of interest have been explored in priorresearch on online marketplaces Hao and Fan (2014)focus on the publishing industry and show that e-book prices are lower in the pure reseller mode due tothe existence of a complementary market (ie e-bookreaders) Young et al (2014) show that channel modechoice can be used as a strategic tool to benefit fromthird-party product reviews More recently Tan et al(2016) and Tan and Carrillo (2017) illustrate that themechanism of the online marketplace can benefit both
the upstream supplier and the retailer in the digitalpublishing industryOur study differs from the aforementioned studies
in four key aspects First we explicitly considerupstream supplier competition in identifying an equi-librium configuration mode for the online retailer Webelieve this is a critical aspect since most online retail-ers simultaneously offer substitute products fromcompeting suppliers Second since the configurationmode impacts the order-fulfillment costs borne by dif-ferent supply chain members we also integrate thiscritical aspect in our setting As indicated earlierthese costs are a significant portion of the total chan-nel costs Third practice indicates that some onlineretailers operate as both resellers and online market-places for the same product category This drives usto explicitly include such a hybrid configurationmode in our comparisons Fourth we endogenize theequilibrium mode choice within the interactionsbetween the suppliers and the online intermediaryand also the proportional fee for the suppliers underthe online marketplace modeOur study also contributes to the research literature
on channel structure Early work on market channelsconcerned whether suppliers are better off using inde-pendent intermediaries instead of vertical integration(eg Choi 1991 Coughlan 1985 McGuire and Staelin1983 Trivedi 1998) A common result is that the exis-tence of independent intermediaries is able to miti-gate the competition among suppliers For exampleMcGuire and Staelin (1983) consider the partial sub-stitutability between two products from two suppliersselling through exclusive retailers They conclude thatsuppliers prefer to use independent intermediarieswhen the degree of substitutability is high and prefercompany-owned stores otherwise Coughlan (1985)discusses the problem of choosing a vertical market-ing channel in a product-differentiated duopoly mar-ket and shows that integration of the marketingfunction results in greater price competition andlower prices than the use of independent marketingmiddlemenAnother stream of research on channel structure
examines the impact of introducing an online directchannel in addition to an existing retail channel Chi-ang et al (2003) have argued that the introduction ofthe direct channel may benefit the retailer through areduction of the wholesale price Tsay and Agrawal(2004a) have confirmed that result in a different set-ting by considering various sources of inefficiencyCai (2010) and Cai et al (2012) explore the impact ofmultiple channel structures and revenue sharingschemes to the supplier the retailer and the supplychain Recent research has considered the rationale ofselling through online platforms For example Ryanet al (2012) consider a setting with a supplier who
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1597
sells goods through its own direct sale website andhas the option of selling its goods through an onlineplatform They assume that selling through the plat-form system can attract more customers to the sup-plierrsquos own website but at some expense eg a fixedparticipation fee or proportional fee charged by theonline platform Interested readers are referred toCattani et al (2004) and Tsay and Agrawal (2004b) formore comprehensive reviewsOur study on strategic mode choice between sup-
pliers and online retailers differs from the stream ofmarket channel literature in three ways First in priorresearch the online retailerintermediary is assumedto function as a reseller In our study under the onlinemarketplace mode the intermediary acts as an agentcharging a proportional fee for each sale Shy andWang (2011) suggest that such a fee could be regardedas a revenue sharing mechanism that may mitigatedouble marginalization Second the focus of ourstudy is identifying which selling mode is more bene-ficial for online intermediaries rather than a compar-ison of directindirect selling activities Third weexplicitly examine how the order-fulfillment cost andupstream competition two significant but unexploredfactors determine the mode configuration choiceOur research also complements the stream of litera-
ture on store-within-store in an offline setting (egJerath and Zhang 2010 Netemeyer et al 2012) whereretailers essentially rent out their retail space to sup-pliers and give them complete autonomy over retaildecisions like pricing and in-store services Our studyof online marketplaces differs from this literaturestream in two ways First in store-within-store con-tracts suppliers typically manage all retail decisionsand the retailers charge them a fixed periodic rent(Jerath and Zhang 2010) Our work suggests that aproportionalrevenue sharing fee could be a viablealternative in this setting Through guidelines on pro-portional fees (a) (ie the revenue sharing proportionthat the online retailers can keep) we show thatonline retailers can set the value of a to encourage ordiscourage the adoption of the online marketplacemode for the suppliers As noted by Cachon and Lari-viere (2005) revenue sharing contract is very difficultto implement in the offline setting However this isnot the case for the online marketplace setting whichis analogous to the store-within-store model Secondwe show that order-fulfillment costs and upstreamcompetition jointly determine the optimal mode con-figuration choice for online retailers To the best ofour knowledge the literature on the store-within-store setting has largely neglected the impacts of thesetwo significant but unexplored factorsIn summary the key aspects of our paper (eg the
proportional-feerevenue-sharing fee the order-ful-fillment cost and the upstream competition between
suppliers) have not been thoroughly explored in theextant store-within-store literature Thus our findingsregarding the online marketplace can also provideuseful insights into store-within-store settings In thenext section we turn to a description of our modelingframework
3 Modeling Framework
We consider a stylized supply chain consisting of twocompeting suppliers (A and B) selling two substi-tutable products (A sells product a B sells product b)through a common online intermediary (I) Given thatthe products are substitutable we follow establishednorms in the marketing and operations literature(eg Birge et al 1998 Choi 1991 Gal-Or et al 2008Garcia-Gallego and Georgantzis 2001 Li and Zhang2008 McGuire and Staelin 1983) and our demandfunction is2
da frac14 h pa thorn cethpb paTHORN eth1THORNdb frac14 h pb thorn cethpa pbTHORN eth2THORN
where di and pi refer to the realized demand andretail prices for product i(=a b) respectively and hdenotes the market potential c gt 0 is the measure ofthe intensity of price competition with higher valuesindicating a greater degree of product substitutionand greater intensity of price competition3 Ourentire analysis without loss of generality assumesthat production costs for suppliers are normalizedto be equal to zero For ease of exposition we usethe pronoun ldquoherdquo to represent the intermediary andldquosherdquo to denote the suppliers in the remainder ofthe studyThree alternative mode choices evaluated are as
follows
bull Reseller Mode RR Under mode RR the inter-mediary acts as a reseller for both suppliersSince many online intermediaries (eg Ama-zon and JDcom) were historically pure resel-lers we set RR mode to be the benchmark(current) mode
bull Hybrid Mode PR This mode represents a set-ting where the intermediary acts as a market-place for one supplier (suppose Supplier A) byspecifying a proportional fee a and acts as areseller for the other supplier (Supplier B)
bull Online Marketplace Mode PP In direct con-trast to mode RR under this mode the inter-mediary specifies a proportional fee a to bothsuppliers and acts as a marketplace for bothsuppliers
Following industry practice the intermediary actsas a Stackelberg leader in terms of whether she will
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1598 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
act as a reseller or marketplace for one or both suppli-ers The suppliers as followers accept or reject theintermediaryrsquos offer In determining the equilibriumoutcomes for each alternative mode the conditionsunder which suppliers will accept the intermediaryrsquosmode choice are explicitly incorporatedBased on this discussion it is apparent that the pos-
sible existence of each mode of operation is driven bythe intermediaryrsquos operational offering (eg the pro-portional fee a) and the suppliersrsquo willingness toaccept this offer Assuming the existence of eachmode the interaction between the intermediary andsuppliers is illustrated in Figure 1 The sequence ofevents for each mode of operation is as follows
bull Mode RR In this benchmark (current) mode theintermediary has decided not to offer the onlinemarketplace service and acts as a reseller for bothsuppliers Suppliers start by simultaneouslyoffering wholesale prices wa and wb to the inter-mediary who in turn simultaneously sets retailprices pa and pb for the consumers
bull Mode PR In this mode the intermediary hasoffered to act as a marketplace for Supplier A byspecifying a proportional fee a and still acts as areseller for Supplier B Assuming Supplier Aaccepts the online marketplace offer and Sup-plier B chooses to continue participating under
the reseller format then based on the quotedwholesale price wb from Supplier B both Sup-plier A and the intermediary simultaneously setthe retail prices pa and pb respectively
bull Mode PP In this mode the intermediary hasoffered the online marketplace service to bothsuppliers by specifying the proportional fee aAssuming both suppliers accept the offer thenboth of them determine the retail prices pa andpb simultaneously
Note that under the marketplace service the inter-mediary yields the pricing power to the supplier Theintermediary receives a commission in proportion tothe supplierrsquos revenue at a rate of a which is deter-mined by the intermediary All three modes are ana-lyzed under a complete information settingThe proportion a is specified as a referral fee by
some online intermediaries (Geng et al 2018) Thisproportional fee although different across productcategories is the same for all products within a cer-tain category Empirical data indicates that common avalues range from 6 to 25 of the sale price depend-ing on the product category on Amazon while forJDcom the fee for most product categories rangesfrom 5 to 12 In addition to this proportional feesome intermediaries also charge a fixed subscription
Figure 1 Channel Modes RR PR and PP [Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1599
fee For example Amazon Marketplace charges a$3999 monthly subscription fee for sellers who planto sell more than 40 items a month Searscom alsocharges a $3999 fee to suppliers whose sales are over$400 during one month Because this fee is relativelysmall compared with the suppliersrsquo sales volumewithout loss of generality we normalize this fixed feepart to zero Previous literature has also adopted thisassumption (Abhishek et al 2016)Order fulfillment that is delivering physical goods
to the customer is commonly cited as one of the mostexpensive and critical operations of online sellers(Agatz et al 2008) The cost of order fulfillment canrun as high as 25 of sales (Kapner 2014) To fulfill anorder firms take on costs such as warehouse build-ingrenting costs hiring staff to handle packages anddelivering the products to customers The costs of stor-age and hiring staff are quite significant and can beviewed as fixed cost Meanwhile the delivery cost foronline shopping is usually undertaken by the cus-tomers For ease of exposition we assume that order-fulfillment costs are fixed since we find that the keyinsights stemming from our analysis are similar evenif these costs are a function of the number of ordersreceived4 Essentially if the intermediary functions asa reseller he will incur a fixed cost FI gt 0 to fulfill theorder for per product category (product a or b) if theintermediary functions as an online marketplace eachsupplier must be responsible for a fixed cost FS gt 0 tofulfill orders Through analysis we find that if theintermediaryrsquos order-fulfillment cost is lower than thesupplierrsquos order-fulfillment cost (ie FI lt FS) the equi-librium mode is more likely to be the pure resellermode and vice versa To eliminate the possibility thatthe equilibrium mode is driven by the asymmetric coststructure we assume FI = FS = F in our base modelMore specifically in mode RR the intermediary bearsorder-fulfillment costs for both product categories (ie2F) in mode PR Supplier A bears the order-fulfillmentcost (F) for its product and the intermediary bears theorder-fulfillment cost (F) for reselling Supplier Brsquos pro-duct in mode PP both Supplier A and Supplier B bearthe order-fulfillment cost (F) of their own product cate-gory5 Recently some online intermediaries (egAmazon Sears JD etc) and 3rd party online order ful-fillment service providers have offered an option tosuppliers that allows them to outsource their order ful-fillment In this case the supplier will pay some feesfor order fulfillment that is even if the supplier out-sources the order fulfillment he still needs to bear therelated costs For ease of exposition and without lossof generality we focus on the case that suppliers willfulfill the order in-house in our base model6
In the next section the equilibrium outcomes ofeach channel mode are structurally characterized
4 Equilibrium Analysis
The first set of results described below is our bench-mark (current) setting where the intermediary acts asa reseller for both suppliers (mode RR) Next we con-sider the setting where one supplier switches to theonline marketplace mode while the other supplier stilladopts the reseller mode (model PR) Finally wefocus on the case in which both suppliers choose toaccept the intermediary terms and switch to theonline marketplace mode (mode PP) In derivingthese results the necessary conditions for the exis-tence of each operating mode (ie RR PR and PP)are that the equilibrium profits for each supplier areat least as large as the profits each supplier could real-ize through exercising an outside option (denoted bypO)
41 Mode RRUnder this setting the suppliers simultaneouslyquote wholesale prices (wa and wb) and then the inter-mediary (who bears the order-fulfillment costs forboth products) determines the retail prices pa and pbThe profits for the suppliers and the intermediary areas follows
pAethRRTHORN frac14 wafrac12h pa thorn cethpb paTHORNpBethRRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethRRTHORN frac14 ethpa waTHORNfrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN 2F
We solve this game by backward induction Forany given wholesale prices wa and wb we first charac-terize the equilibrium retail prices that would maxi-mize pI(RR) We then determine the wholesale pricesfor the suppliers by simultaneously maximizing theirindividual profit functions
LEMMA 1 There exists a unique equilibrium for modeRR The equilibrium prices and demands are as follows
wRRa frac14 wRR
b frac14 1
2thorn ch
pRRa frac14 pRRb frac14 3thorn c2eth2thorn cTHORN h
dRRa frac14 dRRb frac14 1thorn c2eth2thorn cTHORN h
and the corresponding optimal profits are given by
pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2
pRRI frac14 eth1thorn cTHORN22eth2thorn cTHORN2 h
2 2F
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1600 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
It is straightforward to note that an increase incompetition intensity (c) would lead to an increasein the intermediaryrsquos profit and a simultaneousdecrease in supplier profits since an increase in creduces the suppliersrsquo pricing power relative tothat of the intermediary The profit for each sup-plier under this RR benchmark mode serves as abase for evaluating whether a supplier wouldchoose the online marketplace if it was offered bythe intermediaryThe technical condition for the existence of the
RR mode and its corresponding equilibrium is thatthe supplierrsquos profit under this mode is at least aslarge as the profit that each supplier could realizeby exercising the outside option that ispO pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2 Given that the reseller
mode is adopted by a large number of suppliersselling through online intermediaries this exis-tence condition is assumed to hold for the remain-der of this study
42 Mode PRIn this setting only one supplier will be offeredthe option of the online marketplace mode fromthe intermediary Hence one supplier adopts theonline marketplace mode if accepted while theother retains the reseller mode Without loss ofgenerality we assume that the online marketplaceoption is accepted by Supplier A and the interme-diary operates as a reseller for Supplier B For agiven proportion a the suppliersrsquo and intermedi-aryrsquos profits are
pAethPRTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethPRTHORN frac14 apafrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN F
We solve this game by backward induction as fol-lows For any given wb we first characterize thesimultaneous pricing decision where Supplier A deci-des the retail price of product a (ie pa) to maximizepA(PR) and the intermediary decides the retail priceof product b (ie pb) to maximize pI(PR) Next wedetermine Supplier Brsquos wholesale price wb to maxi-mize pB(PR) Finally using this wholesale price wbthe optimal retail prices pa and pb can be set Theresults of this analysis are presented in the lemmabelow
LEMMA 2 Given a proportion a and Supplier A accept-ing the offer of an online marketplace while the intermedi-ary serves as a reseller for Supplier B there exists aunique equilibrium with prices and demands are asfollows
wPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc
2eth1thorn cTHORNeth2thorn 4cthorn c2THORN h
pPRa frac14 eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORNh
pPRb frac14 eth3thorn 6cthorn 2c2THORNeth2thorn 3cTHORNthorn aeth1thorn c c2THORNcfrac124eth1thorn cTHORN2eth1thorn aTHORNc2eth2thorn 4cthorn c2THORN h
dPRa frac14 eth4thorn 9cthorn 3c2THORNeth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn cTHORNeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORN h
dPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2 h
and the corresponding equilibrium profits are
pPRA frac14
eth1aTHORNeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2F
pPRB frac14 frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc22eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc2eth4thorn8cthorn2c2THORNh
2
pPRI frac14aeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22
frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2
thorn
frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc2eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN
Mh2F
where
M frac14 frac12eth1thorn cTHORNeth2thorn 3cTHORNeth2thorn 4cthorn c2THORN thorn aeth1thorn cTHORNeth2thorn 3cTHORNc2thorn aeth3thorn 4cTHORNeth2thorn 4cthorn c2THORNc a2eth1thorn 2cTHORNc3
For the equilibrium in Lemma 2 to exist the follow-ing proposition characterizes the sufficient conditionsunder which Supplier A will accept the intermedi-aryrsquos marketplace offer and Supplier B will continueto participate under the reseller mode
PROPOSITION 1 Supplier A will accept the intermediaryrsquosoffer of an online marketplace provided pPRA pRRA andSupplier B will continue to participate under the resellermode provided pPRB pO This leads to the following suffi-cient conditions for the existence of the PR mode
bull F le FPR where FPR frac14 eth4thorn 9cthorn 3c2THORN2eth1thorn cTHORNeth2thorn cTHORN2eth4thorn 8cthorn 2c2THORN2 h
2 1thorn c
2eth2thorn cTHORN2 h2 and
bull a aPR frac14 minfrac12aPRA aPRB where aPRA is theunique solution of eth1 aPRA THORNeth1thorn cTHORNfrac12eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aPRA eth1thorn 2cTHORNc22
frac124eth1thorn cTHORN2 eth1thorn aPRATHORNc22eth4thorn 8cthorn 2c2THORN2 h2 F frac14 1thorn c
2eth2thorn cTHORN2
h2 and aPRB is the unique solution offrac12eth1thorn cTHORNeth2thorn 3cTHORN aPRB eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPRBTHORNc2eth4thorn 8cthorn 2c2THORN h
2 frac14 pO
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1601
Proposition 1 characterizes the conditions underwhich Supplier A will accept the intermediaryrsquos mar-ketplace offer and Supplier B will not pursue the out-side option but continue to participate under thereseller mode The first condition relates to the order-fulfillment cost which should not exceed a certainthreshold Recall that by choosing the online market-place service the supplier gains pricing power butbears the order-fulfillment cost at the same timeThus this threshold value of the fulfillment costreflects the trade-off between gains from the pricingpower and the costs (in terms of order fulfillment) ofrealizing such gainsAssuming the first condition is satisfied the propo-
sition then specifies a second condition related to themaximum proportion a which can be charged to thesupplier by the intermediary Interesting insightsregarding this proportion a are shown in Lemma 3
LEMMA 3 (i) pPRI increases in a whereas both pPRA andpPRB decrease in a (ii) aPR decreases in F
Lemma 3(i) indicates that an increase in a willincrease the intermediaryrsquos profit pPRI and simultane-ously decrease Supplier Arsquos profit pPRA Hence ashould be small (ie a aPRA ) to ensure pPRA pRRA sothat Supplier A will accept the intermediaryrsquos market-place offer Meanwhile a counterintuitive result isthat an increase in a would also lead to a reduction inthe profit of the other supplier (B) who has not beenoffered the online marketplace service The reason forthis negative spillover effect is that when a isincreased the supplier (A) tends to increase the priceas the intermediary will keep a higher proportion ofthe revenue More specifically when the intermediaryraises the proportion a she can keep Supplier Aresponds by charging higher retail prices In responsethe intermediary sets a higher retail price for SupplierBrsquos product leading to lower demand for the sameproduct The net result is lower profit for Supplier BThus a needs to be small enough (a aPRB ) so thatSupplier B will not pursue the outside option but con-tinue to participate under the reseller mode orpPRB pO7 In equilibrium assuming the conditionrelated to the order-fulfillment cost is met (ieF le FPR) then the intermediary will set a frac14 aPR frac14minfrac12aPRA aPRB since this will allow the intermediary toextract the maximum profits from the supplier Fur-ther Lemma 3(ii) shows that this threshold value ofthe proportional fee (ie aPR) is decreasing in theorder-fulfillment cost (F)In summary from a managerial perspective our
results indicate that the intermediary requires astrategic focus in choosing the parameters for offeringthe online marketplace service This service will onlybe attractive to suppliers provided the order-
fulfillment cost is not significantly high If this is thecase then the intermediary should set the proportiona equal to its upper bound in order to realize the max-imum profits From a supplier perspective the onlinemarketplace offering is preferred over the reseller set-ting when the flexibility gains due to price setting out-weigh the explicit costs of order fulfillment
43 Mode PPIn this setting both suppliers will be offered theoption of online marketplace from the intermediaryIf both suppliers choose the online marketplace ser-vice they will take into consideration the proportiona which they need to share with the intermediary Fora given proportion a the profits for each supplier andthe intermediary under this mode (ie both supplierschoose the online marketplace service) are as follows
pAethPPTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPPTHORN frac14 eth1 aTHORNpbfrac12h pb thorn cethpa pbTHORN F
pIethPPTHORN frac14 apafrac12h pa thorn cethpb paTHORNthorn apbfrac12h pb thorn cethpa pbTHORN
We solve this game by simultaneously determiningthe retail prices pa and pb which maximize pA(PP)and pB(PP) The results are shown in the lemmabelow
LEMMA 4 Assuming both suppliers accept the propor-tion a there exists a unique equilibrium with retail pricesand demands are as follows
pPPa frac14 pPPb frac14 1
2thorn ch
dPPa frac14 dPPb frac14 1thorn c2thorn c
h
The corresponding equilibrium profits are given by
pPPA frac14 pPPB frac14 eth1 aTHORN 1thorn c
eth2thorn cTHORN2 h2 F
pPPI frac14 2a1thorn c
eth2thorn cTHORN2 h2
When is this mode a possible outcome This can beframed using the decision matrix in Table 1 where weare interested in the equilibrium when both supplierswill accept the online marketplace offer This equilib-rium will arise when both of the following conditionshold Given that Supplier B (A) adopts the resellermodel Supplier Arsquos (Brsquos) profit should be higherunder the condition of acceptance than refusal of theintermediaryrsquos marketplace offer which is pPRA pRRA(pRPB pRRB ) given that Supplier B (A) adopts theonline marketplace Supplier Arsquos (Brsquos) profit should
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1602 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
also be higher when it accepts rather than refuses theintermediaryrsquos marketplace offer which is pPPA pRPA(pPPB pPRB ) In addition to ensure this equilibrium isstable both suppliersrsquo profits under the PP modeshould not be lower than the profit under the outsideoption which is pPPA pO (pPPB pO) Based on theresults in Lemma 4 the following proposition identi-fies the conditions under which both suppliers willnot pursue the outside option but choose the onlinemarketplace service offering
PROPOSITION 2 Both suppliers will choose the onlinemarketplace service provided pPRA pRRA pPPA pRPA andpPPA pO This leads to the following sufficient condi-tions for the existence of the PP mode
bull F le FPP where FPP frac14 1thorn ceth2thorn cTHORN2 h
2 eth1thorn cTHORNeth2thorn 3cTHORN2eth2thorn cTHORNeth4thorn 8cthorn 2c2THORN
h2 and
bull a aPP frac14 minfaPRA aPPA aPPB g where aPPA is the
unique solution of eth1 aPPA THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14frac12eth1thorn cTHORNeth2thorn 3cTHORN aPPA eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPPATHORNc2eth4thorn 8cthorn 2c2THORN h
2 and aPPB is the
unique solution of eth1 aPPB THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14 pO
Similar to the hybrid mode both suppliers willaccept the intermediaryrsquos online marketplace offerwhen the order-fulfillment cost and the proportionalfee that the intermediary charges are all below a cer-tain threshold It is worth noting that the threshold inorder-fulfillment cost is more stringent than that statedin Proposition 1 since FPP lt FPR As before the inter-mediaryrsquos profit pPPI increases in a while both pPPA andpPPB decrease in a In equilibrium the intermediary isonce again incentivized to set a = aPP which decreasesin the order-fulfillment cost (F) We also note that sup-pliersrsquo profits will decrease as the upstream competi-tion intensifies (ie a higher value of c)In the next section we compare the equilibrium
solutions for the three modes with a view to provid-ing managerial insights
5 Insights
51 Pricing and DemandFrom a pricing and demand perspective the proposi-tion below provides a comparison across the threechannel modes
PROPOSITION 3 Comparing the equilibrium results ofretail prices and demand we have (i) pRRa frac14pRRb [ pPRb [ pPRa [ pPPa frac14 pPPb (ii) dPRa [ dPPa frac14dPPb [ dRRa frac14 dRRb [ dPRb
Interestingly the results of Proposition 3 are inde-pendent of the proportional fee rate a and hold regard-less of the parameter settings for competitionintensity (c) and maximum market size (h) Essen-tially market prices are highest in the pure resellermode since they are set by the intermediary whoattempts to moderate the competition between thesuppliers On the other hand when the suppliers setthe retail prices by operating in the online market-place mode they compete with each other directlyleading to lower retail prices Because of doublemarginalization in mode PR the retail price of theproduct that does not join the online marketplace ishigher than the price for which the supplier whochooses to do so Corresponding to these pricingstructures demand coverage is higher in the onlinemarketplace mode (PP) as compared to the resellermode (RR)
52 Intermediary ProfitsWe now proceed to examine how intermediary prof-itability across the three modes is affected by two keyparameters competition intensity (c) and order-fulfill-ment cost (F)8 To assess how these parameters affectprofitability of all supply chain members across thethree modes we start by making the followingassumptions First we assume that the order-fulfill-ment cost is small enough to ensure that both modePR and PP are feasible Analytically the potential val-ues of F under this assumption are that F lt FPP(ltFPR)Note that the value of F which satisfies the firstassumption also leads both aPR and aPP to be positiveOur second and final assumption is that the interme-diary chooses to set a = aPR for mode PR and a = aPP
for mode PP which represent equilibrium outcomesProposition 4 structurally characterizes which
mode choice would be preferred by the intermedi-ary as moderated by the competition intensityparameter c
PROPOSITION 4 There exist competition intensity thresh-olds c gt 0 and c(gtc) such that
bull If c lt c the optimal mode is PP
bull If c gt c the optimal mode is RR and
bull If c le c le c mode PR can be the optimal mode
Proposition 4 indicates that when competitionintensity is low the optimal mode is PP when compe-tition intensity is high the optimal mode is RR Inaddition we observe that when competition intensityis medium the optimal mode is PR The rationale
Table 1 Decision Matrix with Resulting Profits
Supplier A option
Supplier B option
Online marketplace Reseller
Online marketplace pPPA pPPB pPRA pPRB
Reseller pRPA pRPB pRRA pRRB
Note Due to symmetry we know that pPRA frac14 pRPB and pPRB frac14 pRPA
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1603
behind these results hinges on the moderating role ofthe intermediary Previous literature has shown thatthe intermediary prefers high levels of upstream com-petition as it strengthens its channel power (Wangand Shin 2015) From a mode perspective supplierscompete more fiercely in the online marketplacemode compared with the reseller mode as they set theprices directly and alternative modes allow the inter-mediary to moderate this competitive effect To elabo-rate when both suppliers are very competitive theintermediary would prefer to operate as a resellersince any other mode will lead to increased price com-petition which in turn negatively impacts the inter-mediary This result validates the conjecture of Hagiu(2007) that strong substitutability between suppliersrsquoproducts would lead to a greater preference for thepure reseller mode When the competition intensitybetween suppliers is low the intermediaryrsquos moderat-ing role is not as important Note that in mode PP theproportional fee could be considered a form of rev-enue sharing which would mitigate the impact ofdouble-marginalization (Geng et al 2018 Shy andWang 2011 Tan and Carrillo 2017) Hence the inter-mediary prefers mode PP when competition intensityis low When the competition intensity between sup-pliers is in the intermediate range the intermediaryrsquosmoderating role is important but not significantenough to alleviate double-marginalization effectsleading to a preference for the hybrid mode PRAnecdotal observations provide face validity for
our results For the household appliance market inChina the similarity between branded products leadsto strong substitutability (or high competition inten-sity) Hence most online retailers (eg JDcom) preferto act as resellers for products in this category On theother hand the media content industry (publishingvideo games) is characterized by low levels of pro-duct substitutability (or low competition intensity)This has motivated online retailers to set the terms forthe online marketplace mode in such a way that sup-pliers have chosen that mode of operation Etsycomthe largest e-commerce website in handmade or vin-tage items serves as a pure online marketplacebecause of low product substitutability FurtherAmazon serves as the reseller in certain high competi-tive product categories such as electronics but servesas a marketplace for other less competitive productcategories such as fine art (ie long-tail productcategory)Next we consider the impact of the order-fulfill-
ment cost on the choice of equilibrium mode for theintermediary Note that Proposition 4 shows thatwhen the market competition intensity is above a cer-tain threshold the intermediary will always prefer toact as a reseller To investigate the impact of theorder-fulfillment cost next we focus on the case that
the market competition intensity is not too high Theproposition below characterizes the mode preferredby the intermediary depending on the magnitude ofthe order-fulfillment cost
PROPOSITION 5 When the market competition intensityis relatively small (ie c cjFfrac140) there exists twothresholds F gt 0 and F(geF) such that
bull when F le F the preferred mode is PP and
bull when F lt F le F the preferred mode is PR and
bull when F gt F the preferred mode is RR
Proposition 5 shows that the magnitude of theorder-fulfillment cost will alter the mode choicesWhen the market competition intensity is relativelylow (c cjFfrac140) as the order-fulfillment costincreases the equilibrium mode will evolve from apure marketplace mode to a hybrid mode then to apure reseller mode The intuition of this result lies inthe fact that the higher the order-fulfillment costbecomes the more difficult it becomes for the inter-mediary to convince the suppliers to adopt the mar-ketplace mode and endure the order-fulfillment costRecall that both aPR and aPP decrease in F That is asthe order-fulfillment cost increases the intermediaryhas to set a much lower proportional fee to ensurethat the suppliers will accept the marketplace offerThis becomes increasingly expensive as the order-ful-fillment cost increases As a result the intermediaryprefers the marketplace mode for both supplierswhen the order-fulfillment cost is low and favors thereseller mode when the order-fulfillment cost issignificantFrom a practical perspective we observe that most
hotel and airline companies use the PP mode on tra-vel sites such as Expedia Travelocity Ctrip andQunar (Ctrip and Qunar are leading travel sites inChina) This is partially driven by the low (almostnegligible) order-fulfillment cost for electronic tick-etsbookings In the publishing industry a transitionto the PP mode from the RR mode is evident for e-books while the RR mode is still preferred for printbooks This could be due in part to the fact that order-fulfillment costs are greater for printed books as com-pared to e-booksHaving analytically examined the impacts of the
competition intensity and order-fulfillment cost indi-vidually we next numerically illustrate how the inter-actions between these two parameters affect the modechoice for the intermediary Figure 2 was generatedby setting h = 10 (ie the market potential equals to10) and pO = 0 (ie the outside option facing thesupplier is 0) Then we simultaneously varyF (F 2 eth0 FPPjcfrac140THORN where FPPjcfrac140 frac14 125) and c(c 2 (0 2)) Note that our results are robust to theparameter changes
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1604 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
between suppliers and online retailers and aims toreveal the trade-offs between the two distinct puremodes (ie reseller and online marketplace) One ofthe key aspects in this context is the cross-channeleffect This effect is a result of suppliers offeringgoods through multiple channels potentially leadingsales in one channel to impact sales in another chan-nel In the presence of spillovers between channelsdifferent selling modes will elicit different reactionsfrom suppliers because the sales through new chan-nels would impact suppliersrsquo sales in the traditionalchannel (Brynjolfsson et al 2009) Abhishek et al(2016) study a setting with one supplier and twoonline intermediaries and find that the online inter-mediary prefers to use the marketplace mode if thesales through the intermediary lead to a negativeeffect on demand of the supplierrsquos traditional channeland vice versa They also show that such preferencesare moderated by competition among intermediariesIn contrast in our study we examine competitionbetween suppliersHagiu and Wright (2015) focus on a single decision
variable the level of demand-enhancing marketingactivity which is controlled by suppliers in the puremarketplace mode but by the intermediary in thepure reseller mode They assume information asym-metry such that each supplier has information on theimpact of market activities on demand while theintermediary and other suppliers do not Their resultsindicate that the pure marketplace mode is preferredif and only if the variance of the private informationexceeds the squared value of spillovers from market-ing activities across products In their model suppli-ers are only charged a fixed membership fee in thepure marketplace mode However in reality there isusually a small fixed participation fee plus a largeproportional revenue sharing fee By contrast weendogenize the revenue sharing fee and examinehow the online intermediary can manipulate this feeto encourage or discourage the adoption of the onlinemarketplace mode for the suppliersJiang et al (2011) also consider the impacts of
demand-enhancing marketing activities and informa-tion asymmetry They focus on the strategic under-selling behavior of sellers on platforms rather thanthe strategic mode choice we consider in this studyOther issues of interest have been explored in priorresearch on online marketplaces Hao and Fan (2014)focus on the publishing industry and show that e-book prices are lower in the pure reseller mode due tothe existence of a complementary market (ie e-bookreaders) Young et al (2014) show that channel modechoice can be used as a strategic tool to benefit fromthird-party product reviews More recently Tan et al(2016) and Tan and Carrillo (2017) illustrate that themechanism of the online marketplace can benefit both
the upstream supplier and the retailer in the digitalpublishing industryOur study differs from the aforementioned studies
in four key aspects First we explicitly considerupstream supplier competition in identifying an equi-librium configuration mode for the online retailer Webelieve this is a critical aspect since most online retail-ers simultaneously offer substitute products fromcompeting suppliers Second since the configurationmode impacts the order-fulfillment costs borne by dif-ferent supply chain members we also integrate thiscritical aspect in our setting As indicated earlierthese costs are a significant portion of the total chan-nel costs Third practice indicates that some onlineretailers operate as both resellers and online market-places for the same product category This drives usto explicitly include such a hybrid configurationmode in our comparisons Fourth we endogenize theequilibrium mode choice within the interactionsbetween the suppliers and the online intermediaryand also the proportional fee for the suppliers underthe online marketplace modeOur study also contributes to the research literature
on channel structure Early work on market channelsconcerned whether suppliers are better off using inde-pendent intermediaries instead of vertical integration(eg Choi 1991 Coughlan 1985 McGuire and Staelin1983 Trivedi 1998) A common result is that the exis-tence of independent intermediaries is able to miti-gate the competition among suppliers For exampleMcGuire and Staelin (1983) consider the partial sub-stitutability between two products from two suppliersselling through exclusive retailers They conclude thatsuppliers prefer to use independent intermediarieswhen the degree of substitutability is high and prefercompany-owned stores otherwise Coughlan (1985)discusses the problem of choosing a vertical market-ing channel in a product-differentiated duopoly mar-ket and shows that integration of the marketingfunction results in greater price competition andlower prices than the use of independent marketingmiddlemenAnother stream of research on channel structure
examines the impact of introducing an online directchannel in addition to an existing retail channel Chi-ang et al (2003) have argued that the introduction ofthe direct channel may benefit the retailer through areduction of the wholesale price Tsay and Agrawal(2004a) have confirmed that result in a different set-ting by considering various sources of inefficiencyCai (2010) and Cai et al (2012) explore the impact ofmultiple channel structures and revenue sharingschemes to the supplier the retailer and the supplychain Recent research has considered the rationale ofselling through online platforms For example Ryanet al (2012) consider a setting with a supplier who
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1597
sells goods through its own direct sale website andhas the option of selling its goods through an onlineplatform They assume that selling through the plat-form system can attract more customers to the sup-plierrsquos own website but at some expense eg a fixedparticipation fee or proportional fee charged by theonline platform Interested readers are referred toCattani et al (2004) and Tsay and Agrawal (2004b) formore comprehensive reviewsOur study on strategic mode choice between sup-
pliers and online retailers differs from the stream ofmarket channel literature in three ways First in priorresearch the online retailerintermediary is assumedto function as a reseller In our study under the onlinemarketplace mode the intermediary acts as an agentcharging a proportional fee for each sale Shy andWang (2011) suggest that such a fee could be regardedas a revenue sharing mechanism that may mitigatedouble marginalization Second the focus of ourstudy is identifying which selling mode is more bene-ficial for online intermediaries rather than a compar-ison of directindirect selling activities Third weexplicitly examine how the order-fulfillment cost andupstream competition two significant but unexploredfactors determine the mode configuration choiceOur research also complements the stream of litera-
ture on store-within-store in an offline setting (egJerath and Zhang 2010 Netemeyer et al 2012) whereretailers essentially rent out their retail space to sup-pliers and give them complete autonomy over retaildecisions like pricing and in-store services Our studyof online marketplaces differs from this literaturestream in two ways First in store-within-store con-tracts suppliers typically manage all retail decisionsand the retailers charge them a fixed periodic rent(Jerath and Zhang 2010) Our work suggests that aproportionalrevenue sharing fee could be a viablealternative in this setting Through guidelines on pro-portional fees (a) (ie the revenue sharing proportionthat the online retailers can keep) we show thatonline retailers can set the value of a to encourage ordiscourage the adoption of the online marketplacemode for the suppliers As noted by Cachon and Lari-viere (2005) revenue sharing contract is very difficultto implement in the offline setting However this isnot the case for the online marketplace setting whichis analogous to the store-within-store model Secondwe show that order-fulfillment costs and upstreamcompetition jointly determine the optimal mode con-figuration choice for online retailers To the best ofour knowledge the literature on the store-within-store setting has largely neglected the impacts of thesetwo significant but unexplored factorsIn summary the key aspects of our paper (eg the
proportional-feerevenue-sharing fee the order-ful-fillment cost and the upstream competition between
suppliers) have not been thoroughly explored in theextant store-within-store literature Thus our findingsregarding the online marketplace can also provideuseful insights into store-within-store settings In thenext section we turn to a description of our modelingframework
3 Modeling Framework
We consider a stylized supply chain consisting of twocompeting suppliers (A and B) selling two substi-tutable products (A sells product a B sells product b)through a common online intermediary (I) Given thatthe products are substitutable we follow establishednorms in the marketing and operations literature(eg Birge et al 1998 Choi 1991 Gal-Or et al 2008Garcia-Gallego and Georgantzis 2001 Li and Zhang2008 McGuire and Staelin 1983) and our demandfunction is2
da frac14 h pa thorn cethpb paTHORN eth1THORNdb frac14 h pb thorn cethpa pbTHORN eth2THORN
where di and pi refer to the realized demand andretail prices for product i(=a b) respectively and hdenotes the market potential c gt 0 is the measure ofthe intensity of price competition with higher valuesindicating a greater degree of product substitutionand greater intensity of price competition3 Ourentire analysis without loss of generality assumesthat production costs for suppliers are normalizedto be equal to zero For ease of exposition we usethe pronoun ldquoherdquo to represent the intermediary andldquosherdquo to denote the suppliers in the remainder ofthe studyThree alternative mode choices evaluated are as
follows
bull Reseller Mode RR Under mode RR the inter-mediary acts as a reseller for both suppliersSince many online intermediaries (eg Ama-zon and JDcom) were historically pure resel-lers we set RR mode to be the benchmark(current) mode
bull Hybrid Mode PR This mode represents a set-ting where the intermediary acts as a market-place for one supplier (suppose Supplier A) byspecifying a proportional fee a and acts as areseller for the other supplier (Supplier B)
bull Online Marketplace Mode PP In direct con-trast to mode RR under this mode the inter-mediary specifies a proportional fee a to bothsuppliers and acts as a marketplace for bothsuppliers
Following industry practice the intermediary actsas a Stackelberg leader in terms of whether she will
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1598 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
act as a reseller or marketplace for one or both suppli-ers The suppliers as followers accept or reject theintermediaryrsquos offer In determining the equilibriumoutcomes for each alternative mode the conditionsunder which suppliers will accept the intermediaryrsquosmode choice are explicitly incorporatedBased on this discussion it is apparent that the pos-
sible existence of each mode of operation is driven bythe intermediaryrsquos operational offering (eg the pro-portional fee a) and the suppliersrsquo willingness toaccept this offer Assuming the existence of eachmode the interaction between the intermediary andsuppliers is illustrated in Figure 1 The sequence ofevents for each mode of operation is as follows
bull Mode RR In this benchmark (current) mode theintermediary has decided not to offer the onlinemarketplace service and acts as a reseller for bothsuppliers Suppliers start by simultaneouslyoffering wholesale prices wa and wb to the inter-mediary who in turn simultaneously sets retailprices pa and pb for the consumers
bull Mode PR In this mode the intermediary hasoffered to act as a marketplace for Supplier A byspecifying a proportional fee a and still acts as areseller for Supplier B Assuming Supplier Aaccepts the online marketplace offer and Sup-plier B chooses to continue participating under
the reseller format then based on the quotedwholesale price wb from Supplier B both Sup-plier A and the intermediary simultaneously setthe retail prices pa and pb respectively
bull Mode PP In this mode the intermediary hasoffered the online marketplace service to bothsuppliers by specifying the proportional fee aAssuming both suppliers accept the offer thenboth of them determine the retail prices pa andpb simultaneously
Note that under the marketplace service the inter-mediary yields the pricing power to the supplier Theintermediary receives a commission in proportion tothe supplierrsquos revenue at a rate of a which is deter-mined by the intermediary All three modes are ana-lyzed under a complete information settingThe proportion a is specified as a referral fee by
some online intermediaries (Geng et al 2018) Thisproportional fee although different across productcategories is the same for all products within a cer-tain category Empirical data indicates that common avalues range from 6 to 25 of the sale price depend-ing on the product category on Amazon while forJDcom the fee for most product categories rangesfrom 5 to 12 In addition to this proportional feesome intermediaries also charge a fixed subscription
Figure 1 Channel Modes RR PR and PP [Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1599
fee For example Amazon Marketplace charges a$3999 monthly subscription fee for sellers who planto sell more than 40 items a month Searscom alsocharges a $3999 fee to suppliers whose sales are over$400 during one month Because this fee is relativelysmall compared with the suppliersrsquo sales volumewithout loss of generality we normalize this fixed feepart to zero Previous literature has also adopted thisassumption (Abhishek et al 2016)Order fulfillment that is delivering physical goods
to the customer is commonly cited as one of the mostexpensive and critical operations of online sellers(Agatz et al 2008) The cost of order fulfillment canrun as high as 25 of sales (Kapner 2014) To fulfill anorder firms take on costs such as warehouse build-ingrenting costs hiring staff to handle packages anddelivering the products to customers The costs of stor-age and hiring staff are quite significant and can beviewed as fixed cost Meanwhile the delivery cost foronline shopping is usually undertaken by the cus-tomers For ease of exposition we assume that order-fulfillment costs are fixed since we find that the keyinsights stemming from our analysis are similar evenif these costs are a function of the number of ordersreceived4 Essentially if the intermediary functions asa reseller he will incur a fixed cost FI gt 0 to fulfill theorder for per product category (product a or b) if theintermediary functions as an online marketplace eachsupplier must be responsible for a fixed cost FS gt 0 tofulfill orders Through analysis we find that if theintermediaryrsquos order-fulfillment cost is lower than thesupplierrsquos order-fulfillment cost (ie FI lt FS) the equi-librium mode is more likely to be the pure resellermode and vice versa To eliminate the possibility thatthe equilibrium mode is driven by the asymmetric coststructure we assume FI = FS = F in our base modelMore specifically in mode RR the intermediary bearsorder-fulfillment costs for both product categories (ie2F) in mode PR Supplier A bears the order-fulfillmentcost (F) for its product and the intermediary bears theorder-fulfillment cost (F) for reselling Supplier Brsquos pro-duct in mode PP both Supplier A and Supplier B bearthe order-fulfillment cost (F) of their own product cate-gory5 Recently some online intermediaries (egAmazon Sears JD etc) and 3rd party online order ful-fillment service providers have offered an option tosuppliers that allows them to outsource their order ful-fillment In this case the supplier will pay some feesfor order fulfillment that is even if the supplier out-sources the order fulfillment he still needs to bear therelated costs For ease of exposition and without lossof generality we focus on the case that suppliers willfulfill the order in-house in our base model6
In the next section the equilibrium outcomes ofeach channel mode are structurally characterized
4 Equilibrium Analysis
The first set of results described below is our bench-mark (current) setting where the intermediary acts asa reseller for both suppliers (mode RR) Next we con-sider the setting where one supplier switches to theonline marketplace mode while the other supplier stilladopts the reseller mode (model PR) Finally wefocus on the case in which both suppliers choose toaccept the intermediary terms and switch to theonline marketplace mode (mode PP) In derivingthese results the necessary conditions for the exis-tence of each operating mode (ie RR PR and PP)are that the equilibrium profits for each supplier areat least as large as the profits each supplier could real-ize through exercising an outside option (denoted bypO)
41 Mode RRUnder this setting the suppliers simultaneouslyquote wholesale prices (wa and wb) and then the inter-mediary (who bears the order-fulfillment costs forboth products) determines the retail prices pa and pbThe profits for the suppliers and the intermediary areas follows
pAethRRTHORN frac14 wafrac12h pa thorn cethpb paTHORNpBethRRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethRRTHORN frac14 ethpa waTHORNfrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN 2F
We solve this game by backward induction Forany given wholesale prices wa and wb we first charac-terize the equilibrium retail prices that would maxi-mize pI(RR) We then determine the wholesale pricesfor the suppliers by simultaneously maximizing theirindividual profit functions
LEMMA 1 There exists a unique equilibrium for modeRR The equilibrium prices and demands are as follows
wRRa frac14 wRR
b frac14 1
2thorn ch
pRRa frac14 pRRb frac14 3thorn c2eth2thorn cTHORN h
dRRa frac14 dRRb frac14 1thorn c2eth2thorn cTHORN h
and the corresponding optimal profits are given by
pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2
pRRI frac14 eth1thorn cTHORN22eth2thorn cTHORN2 h
2 2F
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1600 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
It is straightforward to note that an increase incompetition intensity (c) would lead to an increasein the intermediaryrsquos profit and a simultaneousdecrease in supplier profits since an increase in creduces the suppliersrsquo pricing power relative tothat of the intermediary The profit for each sup-plier under this RR benchmark mode serves as abase for evaluating whether a supplier wouldchoose the online marketplace if it was offered bythe intermediaryThe technical condition for the existence of the
RR mode and its corresponding equilibrium is thatthe supplierrsquos profit under this mode is at least aslarge as the profit that each supplier could realizeby exercising the outside option that ispO pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2 Given that the reseller
mode is adopted by a large number of suppliersselling through online intermediaries this exis-tence condition is assumed to hold for the remain-der of this study
42 Mode PRIn this setting only one supplier will be offeredthe option of the online marketplace mode fromthe intermediary Hence one supplier adopts theonline marketplace mode if accepted while theother retains the reseller mode Without loss ofgenerality we assume that the online marketplaceoption is accepted by Supplier A and the interme-diary operates as a reseller for Supplier B For agiven proportion a the suppliersrsquo and intermedi-aryrsquos profits are
pAethPRTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethPRTHORN frac14 apafrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN F
We solve this game by backward induction as fol-lows For any given wb we first characterize thesimultaneous pricing decision where Supplier A deci-des the retail price of product a (ie pa) to maximizepA(PR) and the intermediary decides the retail priceof product b (ie pb) to maximize pI(PR) Next wedetermine Supplier Brsquos wholesale price wb to maxi-mize pB(PR) Finally using this wholesale price wbthe optimal retail prices pa and pb can be set Theresults of this analysis are presented in the lemmabelow
LEMMA 2 Given a proportion a and Supplier A accept-ing the offer of an online marketplace while the intermedi-ary serves as a reseller for Supplier B there exists aunique equilibrium with prices and demands are asfollows
wPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc
2eth1thorn cTHORNeth2thorn 4cthorn c2THORN h
pPRa frac14 eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORNh
pPRb frac14 eth3thorn 6cthorn 2c2THORNeth2thorn 3cTHORNthorn aeth1thorn c c2THORNcfrac124eth1thorn cTHORN2eth1thorn aTHORNc2eth2thorn 4cthorn c2THORN h
dPRa frac14 eth4thorn 9cthorn 3c2THORNeth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn cTHORNeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORN h
dPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2 h
and the corresponding equilibrium profits are
pPRA frac14
eth1aTHORNeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2F
pPRB frac14 frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc22eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc2eth4thorn8cthorn2c2THORNh
2
pPRI frac14aeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22
frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2
thorn
frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc2eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN
Mh2F
where
M frac14 frac12eth1thorn cTHORNeth2thorn 3cTHORNeth2thorn 4cthorn c2THORN thorn aeth1thorn cTHORNeth2thorn 3cTHORNc2thorn aeth3thorn 4cTHORNeth2thorn 4cthorn c2THORNc a2eth1thorn 2cTHORNc3
For the equilibrium in Lemma 2 to exist the follow-ing proposition characterizes the sufficient conditionsunder which Supplier A will accept the intermedi-aryrsquos marketplace offer and Supplier B will continueto participate under the reseller mode
PROPOSITION 1 Supplier A will accept the intermediaryrsquosoffer of an online marketplace provided pPRA pRRA andSupplier B will continue to participate under the resellermode provided pPRB pO This leads to the following suffi-cient conditions for the existence of the PR mode
bull F le FPR where FPR frac14 eth4thorn 9cthorn 3c2THORN2eth1thorn cTHORNeth2thorn cTHORN2eth4thorn 8cthorn 2c2THORN2 h
2 1thorn c
2eth2thorn cTHORN2 h2 and
bull a aPR frac14 minfrac12aPRA aPRB where aPRA is theunique solution of eth1 aPRA THORNeth1thorn cTHORNfrac12eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aPRA eth1thorn 2cTHORNc22
frac124eth1thorn cTHORN2 eth1thorn aPRATHORNc22eth4thorn 8cthorn 2c2THORN2 h2 F frac14 1thorn c
2eth2thorn cTHORN2
h2 and aPRB is the unique solution offrac12eth1thorn cTHORNeth2thorn 3cTHORN aPRB eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPRBTHORNc2eth4thorn 8cthorn 2c2THORN h
2 frac14 pO
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1601
Proposition 1 characterizes the conditions underwhich Supplier A will accept the intermediaryrsquos mar-ketplace offer and Supplier B will not pursue the out-side option but continue to participate under thereseller mode The first condition relates to the order-fulfillment cost which should not exceed a certainthreshold Recall that by choosing the online market-place service the supplier gains pricing power butbears the order-fulfillment cost at the same timeThus this threshold value of the fulfillment costreflects the trade-off between gains from the pricingpower and the costs (in terms of order fulfillment) ofrealizing such gainsAssuming the first condition is satisfied the propo-
sition then specifies a second condition related to themaximum proportion a which can be charged to thesupplier by the intermediary Interesting insightsregarding this proportion a are shown in Lemma 3
LEMMA 3 (i) pPRI increases in a whereas both pPRA andpPRB decrease in a (ii) aPR decreases in F
Lemma 3(i) indicates that an increase in a willincrease the intermediaryrsquos profit pPRI and simultane-ously decrease Supplier Arsquos profit pPRA Hence ashould be small (ie a aPRA ) to ensure pPRA pRRA sothat Supplier A will accept the intermediaryrsquos market-place offer Meanwhile a counterintuitive result isthat an increase in a would also lead to a reduction inthe profit of the other supplier (B) who has not beenoffered the online marketplace service The reason forthis negative spillover effect is that when a isincreased the supplier (A) tends to increase the priceas the intermediary will keep a higher proportion ofthe revenue More specifically when the intermediaryraises the proportion a she can keep Supplier Aresponds by charging higher retail prices In responsethe intermediary sets a higher retail price for SupplierBrsquos product leading to lower demand for the sameproduct The net result is lower profit for Supplier BThus a needs to be small enough (a aPRB ) so thatSupplier B will not pursue the outside option but con-tinue to participate under the reseller mode orpPRB pO7 In equilibrium assuming the conditionrelated to the order-fulfillment cost is met (ieF le FPR) then the intermediary will set a frac14 aPR frac14minfrac12aPRA aPRB since this will allow the intermediary toextract the maximum profits from the supplier Fur-ther Lemma 3(ii) shows that this threshold value ofthe proportional fee (ie aPR) is decreasing in theorder-fulfillment cost (F)In summary from a managerial perspective our
results indicate that the intermediary requires astrategic focus in choosing the parameters for offeringthe online marketplace service This service will onlybe attractive to suppliers provided the order-
fulfillment cost is not significantly high If this is thecase then the intermediary should set the proportiona equal to its upper bound in order to realize the max-imum profits From a supplier perspective the onlinemarketplace offering is preferred over the reseller set-ting when the flexibility gains due to price setting out-weigh the explicit costs of order fulfillment
43 Mode PPIn this setting both suppliers will be offered theoption of online marketplace from the intermediaryIf both suppliers choose the online marketplace ser-vice they will take into consideration the proportiona which they need to share with the intermediary Fora given proportion a the profits for each supplier andthe intermediary under this mode (ie both supplierschoose the online marketplace service) are as follows
pAethPPTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPPTHORN frac14 eth1 aTHORNpbfrac12h pb thorn cethpa pbTHORN F
pIethPPTHORN frac14 apafrac12h pa thorn cethpb paTHORNthorn apbfrac12h pb thorn cethpa pbTHORN
We solve this game by simultaneously determiningthe retail prices pa and pb which maximize pA(PP)and pB(PP) The results are shown in the lemmabelow
LEMMA 4 Assuming both suppliers accept the propor-tion a there exists a unique equilibrium with retail pricesand demands are as follows
pPPa frac14 pPPb frac14 1
2thorn ch
dPPa frac14 dPPb frac14 1thorn c2thorn c
h
The corresponding equilibrium profits are given by
pPPA frac14 pPPB frac14 eth1 aTHORN 1thorn c
eth2thorn cTHORN2 h2 F
pPPI frac14 2a1thorn c
eth2thorn cTHORN2 h2
When is this mode a possible outcome This can beframed using the decision matrix in Table 1 where weare interested in the equilibrium when both supplierswill accept the online marketplace offer This equilib-rium will arise when both of the following conditionshold Given that Supplier B (A) adopts the resellermodel Supplier Arsquos (Brsquos) profit should be higherunder the condition of acceptance than refusal of theintermediaryrsquos marketplace offer which is pPRA pRRA(pRPB pRRB ) given that Supplier B (A) adopts theonline marketplace Supplier Arsquos (Brsquos) profit should
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1602 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
also be higher when it accepts rather than refuses theintermediaryrsquos marketplace offer which is pPPA pRPA(pPPB pPRB ) In addition to ensure this equilibrium isstable both suppliersrsquo profits under the PP modeshould not be lower than the profit under the outsideoption which is pPPA pO (pPPB pO) Based on theresults in Lemma 4 the following proposition identi-fies the conditions under which both suppliers willnot pursue the outside option but choose the onlinemarketplace service offering
PROPOSITION 2 Both suppliers will choose the onlinemarketplace service provided pPRA pRRA pPPA pRPA andpPPA pO This leads to the following sufficient condi-tions for the existence of the PP mode
bull F le FPP where FPP frac14 1thorn ceth2thorn cTHORN2 h
2 eth1thorn cTHORNeth2thorn 3cTHORN2eth2thorn cTHORNeth4thorn 8cthorn 2c2THORN
h2 and
bull a aPP frac14 minfaPRA aPPA aPPB g where aPPA is the
unique solution of eth1 aPPA THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14frac12eth1thorn cTHORNeth2thorn 3cTHORN aPPA eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPPATHORNc2eth4thorn 8cthorn 2c2THORN h
2 and aPPB is the
unique solution of eth1 aPPB THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14 pO
Similar to the hybrid mode both suppliers willaccept the intermediaryrsquos online marketplace offerwhen the order-fulfillment cost and the proportionalfee that the intermediary charges are all below a cer-tain threshold It is worth noting that the threshold inorder-fulfillment cost is more stringent than that statedin Proposition 1 since FPP lt FPR As before the inter-mediaryrsquos profit pPPI increases in a while both pPPA andpPPB decrease in a In equilibrium the intermediary isonce again incentivized to set a = aPP which decreasesin the order-fulfillment cost (F) We also note that sup-pliersrsquo profits will decrease as the upstream competi-tion intensifies (ie a higher value of c)In the next section we compare the equilibrium
solutions for the three modes with a view to provid-ing managerial insights
5 Insights
51 Pricing and DemandFrom a pricing and demand perspective the proposi-tion below provides a comparison across the threechannel modes
PROPOSITION 3 Comparing the equilibrium results ofretail prices and demand we have (i) pRRa frac14pRRb [ pPRb [ pPRa [ pPPa frac14 pPPb (ii) dPRa [ dPPa frac14dPPb [ dRRa frac14 dRRb [ dPRb
Interestingly the results of Proposition 3 are inde-pendent of the proportional fee rate a and hold regard-less of the parameter settings for competitionintensity (c) and maximum market size (h) Essen-tially market prices are highest in the pure resellermode since they are set by the intermediary whoattempts to moderate the competition between thesuppliers On the other hand when the suppliers setthe retail prices by operating in the online market-place mode they compete with each other directlyleading to lower retail prices Because of doublemarginalization in mode PR the retail price of theproduct that does not join the online marketplace ishigher than the price for which the supplier whochooses to do so Corresponding to these pricingstructures demand coverage is higher in the onlinemarketplace mode (PP) as compared to the resellermode (RR)
52 Intermediary ProfitsWe now proceed to examine how intermediary prof-itability across the three modes is affected by two keyparameters competition intensity (c) and order-fulfill-ment cost (F)8 To assess how these parameters affectprofitability of all supply chain members across thethree modes we start by making the followingassumptions First we assume that the order-fulfill-ment cost is small enough to ensure that both modePR and PP are feasible Analytically the potential val-ues of F under this assumption are that F lt FPP(ltFPR)Note that the value of F which satisfies the firstassumption also leads both aPR and aPP to be positiveOur second and final assumption is that the interme-diary chooses to set a = aPR for mode PR and a = aPP
for mode PP which represent equilibrium outcomesProposition 4 structurally characterizes which
mode choice would be preferred by the intermedi-ary as moderated by the competition intensityparameter c
PROPOSITION 4 There exist competition intensity thresh-olds c gt 0 and c(gtc) such that
bull If c lt c the optimal mode is PP
bull If c gt c the optimal mode is RR and
bull If c le c le c mode PR can be the optimal mode
Proposition 4 indicates that when competitionintensity is low the optimal mode is PP when compe-tition intensity is high the optimal mode is RR Inaddition we observe that when competition intensityis medium the optimal mode is PR The rationale
Table 1 Decision Matrix with Resulting Profits
Supplier A option
Supplier B option
Online marketplace Reseller
Online marketplace pPPA pPPB pPRA pPRB
Reseller pRPA pRPB pRRA pRRB
Note Due to symmetry we know that pPRA frac14 pRPB and pPRB frac14 pRPA
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1603
behind these results hinges on the moderating role ofthe intermediary Previous literature has shown thatthe intermediary prefers high levels of upstream com-petition as it strengthens its channel power (Wangand Shin 2015) From a mode perspective supplierscompete more fiercely in the online marketplacemode compared with the reseller mode as they set theprices directly and alternative modes allow the inter-mediary to moderate this competitive effect To elabo-rate when both suppliers are very competitive theintermediary would prefer to operate as a resellersince any other mode will lead to increased price com-petition which in turn negatively impacts the inter-mediary This result validates the conjecture of Hagiu(2007) that strong substitutability between suppliersrsquoproducts would lead to a greater preference for thepure reseller mode When the competition intensitybetween suppliers is low the intermediaryrsquos moderat-ing role is not as important Note that in mode PP theproportional fee could be considered a form of rev-enue sharing which would mitigate the impact ofdouble-marginalization (Geng et al 2018 Shy andWang 2011 Tan and Carrillo 2017) Hence the inter-mediary prefers mode PP when competition intensityis low When the competition intensity between sup-pliers is in the intermediate range the intermediaryrsquosmoderating role is important but not significantenough to alleviate double-marginalization effectsleading to a preference for the hybrid mode PRAnecdotal observations provide face validity for
our results For the household appliance market inChina the similarity between branded products leadsto strong substitutability (or high competition inten-sity) Hence most online retailers (eg JDcom) preferto act as resellers for products in this category On theother hand the media content industry (publishingvideo games) is characterized by low levels of pro-duct substitutability (or low competition intensity)This has motivated online retailers to set the terms forthe online marketplace mode in such a way that sup-pliers have chosen that mode of operation Etsycomthe largest e-commerce website in handmade or vin-tage items serves as a pure online marketplacebecause of low product substitutability FurtherAmazon serves as the reseller in certain high competi-tive product categories such as electronics but servesas a marketplace for other less competitive productcategories such as fine art (ie long-tail productcategory)Next we consider the impact of the order-fulfill-
ment cost on the choice of equilibrium mode for theintermediary Note that Proposition 4 shows thatwhen the market competition intensity is above a cer-tain threshold the intermediary will always prefer toact as a reseller To investigate the impact of theorder-fulfillment cost next we focus on the case that
the market competition intensity is not too high Theproposition below characterizes the mode preferredby the intermediary depending on the magnitude ofthe order-fulfillment cost
PROPOSITION 5 When the market competition intensityis relatively small (ie c cjFfrac140) there exists twothresholds F gt 0 and F(geF) such that
bull when F le F the preferred mode is PP and
bull when F lt F le F the preferred mode is PR and
bull when F gt F the preferred mode is RR
Proposition 5 shows that the magnitude of theorder-fulfillment cost will alter the mode choicesWhen the market competition intensity is relativelylow (c cjFfrac140) as the order-fulfillment costincreases the equilibrium mode will evolve from apure marketplace mode to a hybrid mode then to apure reseller mode The intuition of this result lies inthe fact that the higher the order-fulfillment costbecomes the more difficult it becomes for the inter-mediary to convince the suppliers to adopt the mar-ketplace mode and endure the order-fulfillment costRecall that both aPR and aPP decrease in F That is asthe order-fulfillment cost increases the intermediaryhas to set a much lower proportional fee to ensurethat the suppliers will accept the marketplace offerThis becomes increasingly expensive as the order-ful-fillment cost increases As a result the intermediaryprefers the marketplace mode for both supplierswhen the order-fulfillment cost is low and favors thereseller mode when the order-fulfillment cost issignificantFrom a practical perspective we observe that most
hotel and airline companies use the PP mode on tra-vel sites such as Expedia Travelocity Ctrip andQunar (Ctrip and Qunar are leading travel sites inChina) This is partially driven by the low (almostnegligible) order-fulfillment cost for electronic tick-etsbookings In the publishing industry a transitionto the PP mode from the RR mode is evident for e-books while the RR mode is still preferred for printbooks This could be due in part to the fact that order-fulfillment costs are greater for printed books as com-pared to e-booksHaving analytically examined the impacts of the
competition intensity and order-fulfillment cost indi-vidually we next numerically illustrate how the inter-actions between these two parameters affect the modechoice for the intermediary Figure 2 was generatedby setting h = 10 (ie the market potential equals to10) and pO = 0 (ie the outside option facing thesupplier is 0) Then we simultaneously varyF (F 2 eth0 FPPjcfrac140THORN where FPPjcfrac140 frac14 125) and c(c 2 (0 2)) Note that our results are robust to theparameter changes
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1604 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
sells goods through its own direct sale website andhas the option of selling its goods through an onlineplatform They assume that selling through the plat-form system can attract more customers to the sup-plierrsquos own website but at some expense eg a fixedparticipation fee or proportional fee charged by theonline platform Interested readers are referred toCattani et al (2004) and Tsay and Agrawal (2004b) formore comprehensive reviewsOur study on strategic mode choice between sup-
pliers and online retailers differs from the stream ofmarket channel literature in three ways First in priorresearch the online retailerintermediary is assumedto function as a reseller In our study under the onlinemarketplace mode the intermediary acts as an agentcharging a proportional fee for each sale Shy andWang (2011) suggest that such a fee could be regardedas a revenue sharing mechanism that may mitigatedouble marginalization Second the focus of ourstudy is identifying which selling mode is more bene-ficial for online intermediaries rather than a compar-ison of directindirect selling activities Third weexplicitly examine how the order-fulfillment cost andupstream competition two significant but unexploredfactors determine the mode configuration choiceOur research also complements the stream of litera-
ture on store-within-store in an offline setting (egJerath and Zhang 2010 Netemeyer et al 2012) whereretailers essentially rent out their retail space to sup-pliers and give them complete autonomy over retaildecisions like pricing and in-store services Our studyof online marketplaces differs from this literaturestream in two ways First in store-within-store con-tracts suppliers typically manage all retail decisionsand the retailers charge them a fixed periodic rent(Jerath and Zhang 2010) Our work suggests that aproportionalrevenue sharing fee could be a viablealternative in this setting Through guidelines on pro-portional fees (a) (ie the revenue sharing proportionthat the online retailers can keep) we show thatonline retailers can set the value of a to encourage ordiscourage the adoption of the online marketplacemode for the suppliers As noted by Cachon and Lari-viere (2005) revenue sharing contract is very difficultto implement in the offline setting However this isnot the case for the online marketplace setting whichis analogous to the store-within-store model Secondwe show that order-fulfillment costs and upstreamcompetition jointly determine the optimal mode con-figuration choice for online retailers To the best ofour knowledge the literature on the store-within-store setting has largely neglected the impacts of thesetwo significant but unexplored factorsIn summary the key aspects of our paper (eg the
proportional-feerevenue-sharing fee the order-ful-fillment cost and the upstream competition between
suppliers) have not been thoroughly explored in theextant store-within-store literature Thus our findingsregarding the online marketplace can also provideuseful insights into store-within-store settings In thenext section we turn to a description of our modelingframework
3 Modeling Framework
We consider a stylized supply chain consisting of twocompeting suppliers (A and B) selling two substi-tutable products (A sells product a B sells product b)through a common online intermediary (I) Given thatthe products are substitutable we follow establishednorms in the marketing and operations literature(eg Birge et al 1998 Choi 1991 Gal-Or et al 2008Garcia-Gallego and Georgantzis 2001 Li and Zhang2008 McGuire and Staelin 1983) and our demandfunction is2
da frac14 h pa thorn cethpb paTHORN eth1THORNdb frac14 h pb thorn cethpa pbTHORN eth2THORN
where di and pi refer to the realized demand andretail prices for product i(=a b) respectively and hdenotes the market potential c gt 0 is the measure ofthe intensity of price competition with higher valuesindicating a greater degree of product substitutionand greater intensity of price competition3 Ourentire analysis without loss of generality assumesthat production costs for suppliers are normalizedto be equal to zero For ease of exposition we usethe pronoun ldquoherdquo to represent the intermediary andldquosherdquo to denote the suppliers in the remainder ofthe studyThree alternative mode choices evaluated are as
follows
bull Reseller Mode RR Under mode RR the inter-mediary acts as a reseller for both suppliersSince many online intermediaries (eg Ama-zon and JDcom) were historically pure resel-lers we set RR mode to be the benchmark(current) mode
bull Hybrid Mode PR This mode represents a set-ting where the intermediary acts as a market-place for one supplier (suppose Supplier A) byspecifying a proportional fee a and acts as areseller for the other supplier (Supplier B)
bull Online Marketplace Mode PP In direct con-trast to mode RR under this mode the inter-mediary specifies a proportional fee a to bothsuppliers and acts as a marketplace for bothsuppliers
Following industry practice the intermediary actsas a Stackelberg leader in terms of whether she will
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1598 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
act as a reseller or marketplace for one or both suppli-ers The suppliers as followers accept or reject theintermediaryrsquos offer In determining the equilibriumoutcomes for each alternative mode the conditionsunder which suppliers will accept the intermediaryrsquosmode choice are explicitly incorporatedBased on this discussion it is apparent that the pos-
sible existence of each mode of operation is driven bythe intermediaryrsquos operational offering (eg the pro-portional fee a) and the suppliersrsquo willingness toaccept this offer Assuming the existence of eachmode the interaction between the intermediary andsuppliers is illustrated in Figure 1 The sequence ofevents for each mode of operation is as follows
bull Mode RR In this benchmark (current) mode theintermediary has decided not to offer the onlinemarketplace service and acts as a reseller for bothsuppliers Suppliers start by simultaneouslyoffering wholesale prices wa and wb to the inter-mediary who in turn simultaneously sets retailprices pa and pb for the consumers
bull Mode PR In this mode the intermediary hasoffered to act as a marketplace for Supplier A byspecifying a proportional fee a and still acts as areseller for Supplier B Assuming Supplier Aaccepts the online marketplace offer and Sup-plier B chooses to continue participating under
the reseller format then based on the quotedwholesale price wb from Supplier B both Sup-plier A and the intermediary simultaneously setthe retail prices pa and pb respectively
bull Mode PP In this mode the intermediary hasoffered the online marketplace service to bothsuppliers by specifying the proportional fee aAssuming both suppliers accept the offer thenboth of them determine the retail prices pa andpb simultaneously
Note that under the marketplace service the inter-mediary yields the pricing power to the supplier Theintermediary receives a commission in proportion tothe supplierrsquos revenue at a rate of a which is deter-mined by the intermediary All three modes are ana-lyzed under a complete information settingThe proportion a is specified as a referral fee by
some online intermediaries (Geng et al 2018) Thisproportional fee although different across productcategories is the same for all products within a cer-tain category Empirical data indicates that common avalues range from 6 to 25 of the sale price depend-ing on the product category on Amazon while forJDcom the fee for most product categories rangesfrom 5 to 12 In addition to this proportional feesome intermediaries also charge a fixed subscription
Figure 1 Channel Modes RR PR and PP [Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1599
fee For example Amazon Marketplace charges a$3999 monthly subscription fee for sellers who planto sell more than 40 items a month Searscom alsocharges a $3999 fee to suppliers whose sales are over$400 during one month Because this fee is relativelysmall compared with the suppliersrsquo sales volumewithout loss of generality we normalize this fixed feepart to zero Previous literature has also adopted thisassumption (Abhishek et al 2016)Order fulfillment that is delivering physical goods
to the customer is commonly cited as one of the mostexpensive and critical operations of online sellers(Agatz et al 2008) The cost of order fulfillment canrun as high as 25 of sales (Kapner 2014) To fulfill anorder firms take on costs such as warehouse build-ingrenting costs hiring staff to handle packages anddelivering the products to customers The costs of stor-age and hiring staff are quite significant and can beviewed as fixed cost Meanwhile the delivery cost foronline shopping is usually undertaken by the cus-tomers For ease of exposition we assume that order-fulfillment costs are fixed since we find that the keyinsights stemming from our analysis are similar evenif these costs are a function of the number of ordersreceived4 Essentially if the intermediary functions asa reseller he will incur a fixed cost FI gt 0 to fulfill theorder for per product category (product a or b) if theintermediary functions as an online marketplace eachsupplier must be responsible for a fixed cost FS gt 0 tofulfill orders Through analysis we find that if theintermediaryrsquos order-fulfillment cost is lower than thesupplierrsquos order-fulfillment cost (ie FI lt FS) the equi-librium mode is more likely to be the pure resellermode and vice versa To eliminate the possibility thatthe equilibrium mode is driven by the asymmetric coststructure we assume FI = FS = F in our base modelMore specifically in mode RR the intermediary bearsorder-fulfillment costs for both product categories (ie2F) in mode PR Supplier A bears the order-fulfillmentcost (F) for its product and the intermediary bears theorder-fulfillment cost (F) for reselling Supplier Brsquos pro-duct in mode PP both Supplier A and Supplier B bearthe order-fulfillment cost (F) of their own product cate-gory5 Recently some online intermediaries (egAmazon Sears JD etc) and 3rd party online order ful-fillment service providers have offered an option tosuppliers that allows them to outsource their order ful-fillment In this case the supplier will pay some feesfor order fulfillment that is even if the supplier out-sources the order fulfillment he still needs to bear therelated costs For ease of exposition and without lossof generality we focus on the case that suppliers willfulfill the order in-house in our base model6
In the next section the equilibrium outcomes ofeach channel mode are structurally characterized
4 Equilibrium Analysis
The first set of results described below is our bench-mark (current) setting where the intermediary acts asa reseller for both suppliers (mode RR) Next we con-sider the setting where one supplier switches to theonline marketplace mode while the other supplier stilladopts the reseller mode (model PR) Finally wefocus on the case in which both suppliers choose toaccept the intermediary terms and switch to theonline marketplace mode (mode PP) In derivingthese results the necessary conditions for the exis-tence of each operating mode (ie RR PR and PP)are that the equilibrium profits for each supplier areat least as large as the profits each supplier could real-ize through exercising an outside option (denoted bypO)
41 Mode RRUnder this setting the suppliers simultaneouslyquote wholesale prices (wa and wb) and then the inter-mediary (who bears the order-fulfillment costs forboth products) determines the retail prices pa and pbThe profits for the suppliers and the intermediary areas follows
pAethRRTHORN frac14 wafrac12h pa thorn cethpb paTHORNpBethRRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethRRTHORN frac14 ethpa waTHORNfrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN 2F
We solve this game by backward induction Forany given wholesale prices wa and wb we first charac-terize the equilibrium retail prices that would maxi-mize pI(RR) We then determine the wholesale pricesfor the suppliers by simultaneously maximizing theirindividual profit functions
LEMMA 1 There exists a unique equilibrium for modeRR The equilibrium prices and demands are as follows
wRRa frac14 wRR
b frac14 1
2thorn ch
pRRa frac14 pRRb frac14 3thorn c2eth2thorn cTHORN h
dRRa frac14 dRRb frac14 1thorn c2eth2thorn cTHORN h
and the corresponding optimal profits are given by
pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2
pRRI frac14 eth1thorn cTHORN22eth2thorn cTHORN2 h
2 2F
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1600 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
It is straightforward to note that an increase incompetition intensity (c) would lead to an increasein the intermediaryrsquos profit and a simultaneousdecrease in supplier profits since an increase in creduces the suppliersrsquo pricing power relative tothat of the intermediary The profit for each sup-plier under this RR benchmark mode serves as abase for evaluating whether a supplier wouldchoose the online marketplace if it was offered bythe intermediaryThe technical condition for the existence of the
RR mode and its corresponding equilibrium is thatthe supplierrsquos profit under this mode is at least aslarge as the profit that each supplier could realizeby exercising the outside option that ispO pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2 Given that the reseller
mode is adopted by a large number of suppliersselling through online intermediaries this exis-tence condition is assumed to hold for the remain-der of this study
42 Mode PRIn this setting only one supplier will be offeredthe option of the online marketplace mode fromthe intermediary Hence one supplier adopts theonline marketplace mode if accepted while theother retains the reseller mode Without loss ofgenerality we assume that the online marketplaceoption is accepted by Supplier A and the interme-diary operates as a reseller for Supplier B For agiven proportion a the suppliersrsquo and intermedi-aryrsquos profits are
pAethPRTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethPRTHORN frac14 apafrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN F
We solve this game by backward induction as fol-lows For any given wb we first characterize thesimultaneous pricing decision where Supplier A deci-des the retail price of product a (ie pa) to maximizepA(PR) and the intermediary decides the retail priceof product b (ie pb) to maximize pI(PR) Next wedetermine Supplier Brsquos wholesale price wb to maxi-mize pB(PR) Finally using this wholesale price wbthe optimal retail prices pa and pb can be set Theresults of this analysis are presented in the lemmabelow
LEMMA 2 Given a proportion a and Supplier A accept-ing the offer of an online marketplace while the intermedi-ary serves as a reseller for Supplier B there exists aunique equilibrium with prices and demands are asfollows
wPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc
2eth1thorn cTHORNeth2thorn 4cthorn c2THORN h
pPRa frac14 eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORNh
pPRb frac14 eth3thorn 6cthorn 2c2THORNeth2thorn 3cTHORNthorn aeth1thorn c c2THORNcfrac124eth1thorn cTHORN2eth1thorn aTHORNc2eth2thorn 4cthorn c2THORN h
dPRa frac14 eth4thorn 9cthorn 3c2THORNeth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn cTHORNeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORN h
dPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2 h
and the corresponding equilibrium profits are
pPRA frac14
eth1aTHORNeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2F
pPRB frac14 frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc22eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc2eth4thorn8cthorn2c2THORNh
2
pPRI frac14aeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22
frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2
thorn
frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc2eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN
Mh2F
where
M frac14 frac12eth1thorn cTHORNeth2thorn 3cTHORNeth2thorn 4cthorn c2THORN thorn aeth1thorn cTHORNeth2thorn 3cTHORNc2thorn aeth3thorn 4cTHORNeth2thorn 4cthorn c2THORNc a2eth1thorn 2cTHORNc3
For the equilibrium in Lemma 2 to exist the follow-ing proposition characterizes the sufficient conditionsunder which Supplier A will accept the intermedi-aryrsquos marketplace offer and Supplier B will continueto participate under the reseller mode
PROPOSITION 1 Supplier A will accept the intermediaryrsquosoffer of an online marketplace provided pPRA pRRA andSupplier B will continue to participate under the resellermode provided pPRB pO This leads to the following suffi-cient conditions for the existence of the PR mode
bull F le FPR where FPR frac14 eth4thorn 9cthorn 3c2THORN2eth1thorn cTHORNeth2thorn cTHORN2eth4thorn 8cthorn 2c2THORN2 h
2 1thorn c
2eth2thorn cTHORN2 h2 and
bull a aPR frac14 minfrac12aPRA aPRB where aPRA is theunique solution of eth1 aPRA THORNeth1thorn cTHORNfrac12eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aPRA eth1thorn 2cTHORNc22
frac124eth1thorn cTHORN2 eth1thorn aPRATHORNc22eth4thorn 8cthorn 2c2THORN2 h2 F frac14 1thorn c
2eth2thorn cTHORN2
h2 and aPRB is the unique solution offrac12eth1thorn cTHORNeth2thorn 3cTHORN aPRB eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPRBTHORNc2eth4thorn 8cthorn 2c2THORN h
2 frac14 pO
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1601
Proposition 1 characterizes the conditions underwhich Supplier A will accept the intermediaryrsquos mar-ketplace offer and Supplier B will not pursue the out-side option but continue to participate under thereseller mode The first condition relates to the order-fulfillment cost which should not exceed a certainthreshold Recall that by choosing the online market-place service the supplier gains pricing power butbears the order-fulfillment cost at the same timeThus this threshold value of the fulfillment costreflects the trade-off between gains from the pricingpower and the costs (in terms of order fulfillment) ofrealizing such gainsAssuming the first condition is satisfied the propo-
sition then specifies a second condition related to themaximum proportion a which can be charged to thesupplier by the intermediary Interesting insightsregarding this proportion a are shown in Lemma 3
LEMMA 3 (i) pPRI increases in a whereas both pPRA andpPRB decrease in a (ii) aPR decreases in F
Lemma 3(i) indicates that an increase in a willincrease the intermediaryrsquos profit pPRI and simultane-ously decrease Supplier Arsquos profit pPRA Hence ashould be small (ie a aPRA ) to ensure pPRA pRRA sothat Supplier A will accept the intermediaryrsquos market-place offer Meanwhile a counterintuitive result isthat an increase in a would also lead to a reduction inthe profit of the other supplier (B) who has not beenoffered the online marketplace service The reason forthis negative spillover effect is that when a isincreased the supplier (A) tends to increase the priceas the intermediary will keep a higher proportion ofthe revenue More specifically when the intermediaryraises the proportion a she can keep Supplier Aresponds by charging higher retail prices In responsethe intermediary sets a higher retail price for SupplierBrsquos product leading to lower demand for the sameproduct The net result is lower profit for Supplier BThus a needs to be small enough (a aPRB ) so thatSupplier B will not pursue the outside option but con-tinue to participate under the reseller mode orpPRB pO7 In equilibrium assuming the conditionrelated to the order-fulfillment cost is met (ieF le FPR) then the intermediary will set a frac14 aPR frac14minfrac12aPRA aPRB since this will allow the intermediary toextract the maximum profits from the supplier Fur-ther Lemma 3(ii) shows that this threshold value ofthe proportional fee (ie aPR) is decreasing in theorder-fulfillment cost (F)In summary from a managerial perspective our
results indicate that the intermediary requires astrategic focus in choosing the parameters for offeringthe online marketplace service This service will onlybe attractive to suppliers provided the order-
fulfillment cost is not significantly high If this is thecase then the intermediary should set the proportiona equal to its upper bound in order to realize the max-imum profits From a supplier perspective the onlinemarketplace offering is preferred over the reseller set-ting when the flexibility gains due to price setting out-weigh the explicit costs of order fulfillment
43 Mode PPIn this setting both suppliers will be offered theoption of online marketplace from the intermediaryIf both suppliers choose the online marketplace ser-vice they will take into consideration the proportiona which they need to share with the intermediary Fora given proportion a the profits for each supplier andthe intermediary under this mode (ie both supplierschoose the online marketplace service) are as follows
pAethPPTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPPTHORN frac14 eth1 aTHORNpbfrac12h pb thorn cethpa pbTHORN F
pIethPPTHORN frac14 apafrac12h pa thorn cethpb paTHORNthorn apbfrac12h pb thorn cethpa pbTHORN
We solve this game by simultaneously determiningthe retail prices pa and pb which maximize pA(PP)and pB(PP) The results are shown in the lemmabelow
LEMMA 4 Assuming both suppliers accept the propor-tion a there exists a unique equilibrium with retail pricesand demands are as follows
pPPa frac14 pPPb frac14 1
2thorn ch
dPPa frac14 dPPb frac14 1thorn c2thorn c
h
The corresponding equilibrium profits are given by
pPPA frac14 pPPB frac14 eth1 aTHORN 1thorn c
eth2thorn cTHORN2 h2 F
pPPI frac14 2a1thorn c
eth2thorn cTHORN2 h2
When is this mode a possible outcome This can beframed using the decision matrix in Table 1 where weare interested in the equilibrium when both supplierswill accept the online marketplace offer This equilib-rium will arise when both of the following conditionshold Given that Supplier B (A) adopts the resellermodel Supplier Arsquos (Brsquos) profit should be higherunder the condition of acceptance than refusal of theintermediaryrsquos marketplace offer which is pPRA pRRA(pRPB pRRB ) given that Supplier B (A) adopts theonline marketplace Supplier Arsquos (Brsquos) profit should
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1602 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
also be higher when it accepts rather than refuses theintermediaryrsquos marketplace offer which is pPPA pRPA(pPPB pPRB ) In addition to ensure this equilibrium isstable both suppliersrsquo profits under the PP modeshould not be lower than the profit under the outsideoption which is pPPA pO (pPPB pO) Based on theresults in Lemma 4 the following proposition identi-fies the conditions under which both suppliers willnot pursue the outside option but choose the onlinemarketplace service offering
PROPOSITION 2 Both suppliers will choose the onlinemarketplace service provided pPRA pRRA pPPA pRPA andpPPA pO This leads to the following sufficient condi-tions for the existence of the PP mode
bull F le FPP where FPP frac14 1thorn ceth2thorn cTHORN2 h
2 eth1thorn cTHORNeth2thorn 3cTHORN2eth2thorn cTHORNeth4thorn 8cthorn 2c2THORN
h2 and
bull a aPP frac14 minfaPRA aPPA aPPB g where aPPA is the
unique solution of eth1 aPPA THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14frac12eth1thorn cTHORNeth2thorn 3cTHORN aPPA eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPPATHORNc2eth4thorn 8cthorn 2c2THORN h
2 and aPPB is the
unique solution of eth1 aPPB THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14 pO
Similar to the hybrid mode both suppliers willaccept the intermediaryrsquos online marketplace offerwhen the order-fulfillment cost and the proportionalfee that the intermediary charges are all below a cer-tain threshold It is worth noting that the threshold inorder-fulfillment cost is more stringent than that statedin Proposition 1 since FPP lt FPR As before the inter-mediaryrsquos profit pPPI increases in a while both pPPA andpPPB decrease in a In equilibrium the intermediary isonce again incentivized to set a = aPP which decreasesin the order-fulfillment cost (F) We also note that sup-pliersrsquo profits will decrease as the upstream competi-tion intensifies (ie a higher value of c)In the next section we compare the equilibrium
solutions for the three modes with a view to provid-ing managerial insights
5 Insights
51 Pricing and DemandFrom a pricing and demand perspective the proposi-tion below provides a comparison across the threechannel modes
PROPOSITION 3 Comparing the equilibrium results ofretail prices and demand we have (i) pRRa frac14pRRb [ pPRb [ pPRa [ pPPa frac14 pPPb (ii) dPRa [ dPPa frac14dPPb [ dRRa frac14 dRRb [ dPRb
Interestingly the results of Proposition 3 are inde-pendent of the proportional fee rate a and hold regard-less of the parameter settings for competitionintensity (c) and maximum market size (h) Essen-tially market prices are highest in the pure resellermode since they are set by the intermediary whoattempts to moderate the competition between thesuppliers On the other hand when the suppliers setthe retail prices by operating in the online market-place mode they compete with each other directlyleading to lower retail prices Because of doublemarginalization in mode PR the retail price of theproduct that does not join the online marketplace ishigher than the price for which the supplier whochooses to do so Corresponding to these pricingstructures demand coverage is higher in the onlinemarketplace mode (PP) as compared to the resellermode (RR)
52 Intermediary ProfitsWe now proceed to examine how intermediary prof-itability across the three modes is affected by two keyparameters competition intensity (c) and order-fulfill-ment cost (F)8 To assess how these parameters affectprofitability of all supply chain members across thethree modes we start by making the followingassumptions First we assume that the order-fulfill-ment cost is small enough to ensure that both modePR and PP are feasible Analytically the potential val-ues of F under this assumption are that F lt FPP(ltFPR)Note that the value of F which satisfies the firstassumption also leads both aPR and aPP to be positiveOur second and final assumption is that the interme-diary chooses to set a = aPR for mode PR and a = aPP
for mode PP which represent equilibrium outcomesProposition 4 structurally characterizes which
mode choice would be preferred by the intermedi-ary as moderated by the competition intensityparameter c
PROPOSITION 4 There exist competition intensity thresh-olds c gt 0 and c(gtc) such that
bull If c lt c the optimal mode is PP
bull If c gt c the optimal mode is RR and
bull If c le c le c mode PR can be the optimal mode
Proposition 4 indicates that when competitionintensity is low the optimal mode is PP when compe-tition intensity is high the optimal mode is RR Inaddition we observe that when competition intensityis medium the optimal mode is PR The rationale
Table 1 Decision Matrix with Resulting Profits
Supplier A option
Supplier B option
Online marketplace Reseller
Online marketplace pPPA pPPB pPRA pPRB
Reseller pRPA pRPB pRRA pRRB
Note Due to symmetry we know that pPRA frac14 pRPB and pPRB frac14 pRPA
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1603
behind these results hinges on the moderating role ofthe intermediary Previous literature has shown thatthe intermediary prefers high levels of upstream com-petition as it strengthens its channel power (Wangand Shin 2015) From a mode perspective supplierscompete more fiercely in the online marketplacemode compared with the reseller mode as they set theprices directly and alternative modes allow the inter-mediary to moderate this competitive effect To elabo-rate when both suppliers are very competitive theintermediary would prefer to operate as a resellersince any other mode will lead to increased price com-petition which in turn negatively impacts the inter-mediary This result validates the conjecture of Hagiu(2007) that strong substitutability between suppliersrsquoproducts would lead to a greater preference for thepure reseller mode When the competition intensitybetween suppliers is low the intermediaryrsquos moderat-ing role is not as important Note that in mode PP theproportional fee could be considered a form of rev-enue sharing which would mitigate the impact ofdouble-marginalization (Geng et al 2018 Shy andWang 2011 Tan and Carrillo 2017) Hence the inter-mediary prefers mode PP when competition intensityis low When the competition intensity between sup-pliers is in the intermediate range the intermediaryrsquosmoderating role is important but not significantenough to alleviate double-marginalization effectsleading to a preference for the hybrid mode PRAnecdotal observations provide face validity for
our results For the household appliance market inChina the similarity between branded products leadsto strong substitutability (or high competition inten-sity) Hence most online retailers (eg JDcom) preferto act as resellers for products in this category On theother hand the media content industry (publishingvideo games) is characterized by low levels of pro-duct substitutability (or low competition intensity)This has motivated online retailers to set the terms forthe online marketplace mode in such a way that sup-pliers have chosen that mode of operation Etsycomthe largest e-commerce website in handmade or vin-tage items serves as a pure online marketplacebecause of low product substitutability FurtherAmazon serves as the reseller in certain high competi-tive product categories such as electronics but servesas a marketplace for other less competitive productcategories such as fine art (ie long-tail productcategory)Next we consider the impact of the order-fulfill-
ment cost on the choice of equilibrium mode for theintermediary Note that Proposition 4 shows thatwhen the market competition intensity is above a cer-tain threshold the intermediary will always prefer toact as a reseller To investigate the impact of theorder-fulfillment cost next we focus on the case that
the market competition intensity is not too high Theproposition below characterizes the mode preferredby the intermediary depending on the magnitude ofthe order-fulfillment cost
PROPOSITION 5 When the market competition intensityis relatively small (ie c cjFfrac140) there exists twothresholds F gt 0 and F(geF) such that
bull when F le F the preferred mode is PP and
bull when F lt F le F the preferred mode is PR and
bull when F gt F the preferred mode is RR
Proposition 5 shows that the magnitude of theorder-fulfillment cost will alter the mode choicesWhen the market competition intensity is relativelylow (c cjFfrac140) as the order-fulfillment costincreases the equilibrium mode will evolve from apure marketplace mode to a hybrid mode then to apure reseller mode The intuition of this result lies inthe fact that the higher the order-fulfillment costbecomes the more difficult it becomes for the inter-mediary to convince the suppliers to adopt the mar-ketplace mode and endure the order-fulfillment costRecall that both aPR and aPP decrease in F That is asthe order-fulfillment cost increases the intermediaryhas to set a much lower proportional fee to ensurethat the suppliers will accept the marketplace offerThis becomes increasingly expensive as the order-ful-fillment cost increases As a result the intermediaryprefers the marketplace mode for both supplierswhen the order-fulfillment cost is low and favors thereseller mode when the order-fulfillment cost issignificantFrom a practical perspective we observe that most
hotel and airline companies use the PP mode on tra-vel sites such as Expedia Travelocity Ctrip andQunar (Ctrip and Qunar are leading travel sites inChina) This is partially driven by the low (almostnegligible) order-fulfillment cost for electronic tick-etsbookings In the publishing industry a transitionto the PP mode from the RR mode is evident for e-books while the RR mode is still preferred for printbooks This could be due in part to the fact that order-fulfillment costs are greater for printed books as com-pared to e-booksHaving analytically examined the impacts of the
competition intensity and order-fulfillment cost indi-vidually we next numerically illustrate how the inter-actions between these two parameters affect the modechoice for the intermediary Figure 2 was generatedby setting h = 10 (ie the market potential equals to10) and pO = 0 (ie the outside option facing thesupplier is 0) Then we simultaneously varyF (F 2 eth0 FPPjcfrac140THORN where FPPjcfrac140 frac14 125) and c(c 2 (0 2)) Note that our results are robust to theparameter changes
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1604 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
act as a reseller or marketplace for one or both suppli-ers The suppliers as followers accept or reject theintermediaryrsquos offer In determining the equilibriumoutcomes for each alternative mode the conditionsunder which suppliers will accept the intermediaryrsquosmode choice are explicitly incorporatedBased on this discussion it is apparent that the pos-
sible existence of each mode of operation is driven bythe intermediaryrsquos operational offering (eg the pro-portional fee a) and the suppliersrsquo willingness toaccept this offer Assuming the existence of eachmode the interaction between the intermediary andsuppliers is illustrated in Figure 1 The sequence ofevents for each mode of operation is as follows
bull Mode RR In this benchmark (current) mode theintermediary has decided not to offer the onlinemarketplace service and acts as a reseller for bothsuppliers Suppliers start by simultaneouslyoffering wholesale prices wa and wb to the inter-mediary who in turn simultaneously sets retailprices pa and pb for the consumers
bull Mode PR In this mode the intermediary hasoffered to act as a marketplace for Supplier A byspecifying a proportional fee a and still acts as areseller for Supplier B Assuming Supplier Aaccepts the online marketplace offer and Sup-plier B chooses to continue participating under
the reseller format then based on the quotedwholesale price wb from Supplier B both Sup-plier A and the intermediary simultaneously setthe retail prices pa and pb respectively
bull Mode PP In this mode the intermediary hasoffered the online marketplace service to bothsuppliers by specifying the proportional fee aAssuming both suppliers accept the offer thenboth of them determine the retail prices pa andpb simultaneously
Note that under the marketplace service the inter-mediary yields the pricing power to the supplier Theintermediary receives a commission in proportion tothe supplierrsquos revenue at a rate of a which is deter-mined by the intermediary All three modes are ana-lyzed under a complete information settingThe proportion a is specified as a referral fee by
some online intermediaries (Geng et al 2018) Thisproportional fee although different across productcategories is the same for all products within a cer-tain category Empirical data indicates that common avalues range from 6 to 25 of the sale price depend-ing on the product category on Amazon while forJDcom the fee for most product categories rangesfrom 5 to 12 In addition to this proportional feesome intermediaries also charge a fixed subscription
Figure 1 Channel Modes RR PR and PP [Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1599
fee For example Amazon Marketplace charges a$3999 monthly subscription fee for sellers who planto sell more than 40 items a month Searscom alsocharges a $3999 fee to suppliers whose sales are over$400 during one month Because this fee is relativelysmall compared with the suppliersrsquo sales volumewithout loss of generality we normalize this fixed feepart to zero Previous literature has also adopted thisassumption (Abhishek et al 2016)Order fulfillment that is delivering physical goods
to the customer is commonly cited as one of the mostexpensive and critical operations of online sellers(Agatz et al 2008) The cost of order fulfillment canrun as high as 25 of sales (Kapner 2014) To fulfill anorder firms take on costs such as warehouse build-ingrenting costs hiring staff to handle packages anddelivering the products to customers The costs of stor-age and hiring staff are quite significant and can beviewed as fixed cost Meanwhile the delivery cost foronline shopping is usually undertaken by the cus-tomers For ease of exposition we assume that order-fulfillment costs are fixed since we find that the keyinsights stemming from our analysis are similar evenif these costs are a function of the number of ordersreceived4 Essentially if the intermediary functions asa reseller he will incur a fixed cost FI gt 0 to fulfill theorder for per product category (product a or b) if theintermediary functions as an online marketplace eachsupplier must be responsible for a fixed cost FS gt 0 tofulfill orders Through analysis we find that if theintermediaryrsquos order-fulfillment cost is lower than thesupplierrsquos order-fulfillment cost (ie FI lt FS) the equi-librium mode is more likely to be the pure resellermode and vice versa To eliminate the possibility thatthe equilibrium mode is driven by the asymmetric coststructure we assume FI = FS = F in our base modelMore specifically in mode RR the intermediary bearsorder-fulfillment costs for both product categories (ie2F) in mode PR Supplier A bears the order-fulfillmentcost (F) for its product and the intermediary bears theorder-fulfillment cost (F) for reselling Supplier Brsquos pro-duct in mode PP both Supplier A and Supplier B bearthe order-fulfillment cost (F) of their own product cate-gory5 Recently some online intermediaries (egAmazon Sears JD etc) and 3rd party online order ful-fillment service providers have offered an option tosuppliers that allows them to outsource their order ful-fillment In this case the supplier will pay some feesfor order fulfillment that is even if the supplier out-sources the order fulfillment he still needs to bear therelated costs For ease of exposition and without lossof generality we focus on the case that suppliers willfulfill the order in-house in our base model6
In the next section the equilibrium outcomes ofeach channel mode are structurally characterized
4 Equilibrium Analysis
The first set of results described below is our bench-mark (current) setting where the intermediary acts asa reseller for both suppliers (mode RR) Next we con-sider the setting where one supplier switches to theonline marketplace mode while the other supplier stilladopts the reseller mode (model PR) Finally wefocus on the case in which both suppliers choose toaccept the intermediary terms and switch to theonline marketplace mode (mode PP) In derivingthese results the necessary conditions for the exis-tence of each operating mode (ie RR PR and PP)are that the equilibrium profits for each supplier areat least as large as the profits each supplier could real-ize through exercising an outside option (denoted bypO)
41 Mode RRUnder this setting the suppliers simultaneouslyquote wholesale prices (wa and wb) and then the inter-mediary (who bears the order-fulfillment costs forboth products) determines the retail prices pa and pbThe profits for the suppliers and the intermediary areas follows
pAethRRTHORN frac14 wafrac12h pa thorn cethpb paTHORNpBethRRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethRRTHORN frac14 ethpa waTHORNfrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN 2F
We solve this game by backward induction Forany given wholesale prices wa and wb we first charac-terize the equilibrium retail prices that would maxi-mize pI(RR) We then determine the wholesale pricesfor the suppliers by simultaneously maximizing theirindividual profit functions
LEMMA 1 There exists a unique equilibrium for modeRR The equilibrium prices and demands are as follows
wRRa frac14 wRR
b frac14 1
2thorn ch
pRRa frac14 pRRb frac14 3thorn c2eth2thorn cTHORN h
dRRa frac14 dRRb frac14 1thorn c2eth2thorn cTHORN h
and the corresponding optimal profits are given by
pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2
pRRI frac14 eth1thorn cTHORN22eth2thorn cTHORN2 h
2 2F
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1600 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
It is straightforward to note that an increase incompetition intensity (c) would lead to an increasein the intermediaryrsquos profit and a simultaneousdecrease in supplier profits since an increase in creduces the suppliersrsquo pricing power relative tothat of the intermediary The profit for each sup-plier under this RR benchmark mode serves as abase for evaluating whether a supplier wouldchoose the online marketplace if it was offered bythe intermediaryThe technical condition for the existence of the
RR mode and its corresponding equilibrium is thatthe supplierrsquos profit under this mode is at least aslarge as the profit that each supplier could realizeby exercising the outside option that ispO pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2 Given that the reseller
mode is adopted by a large number of suppliersselling through online intermediaries this exis-tence condition is assumed to hold for the remain-der of this study
42 Mode PRIn this setting only one supplier will be offeredthe option of the online marketplace mode fromthe intermediary Hence one supplier adopts theonline marketplace mode if accepted while theother retains the reseller mode Without loss ofgenerality we assume that the online marketplaceoption is accepted by Supplier A and the interme-diary operates as a reseller for Supplier B For agiven proportion a the suppliersrsquo and intermedi-aryrsquos profits are
pAethPRTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethPRTHORN frac14 apafrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN F
We solve this game by backward induction as fol-lows For any given wb we first characterize thesimultaneous pricing decision where Supplier A deci-des the retail price of product a (ie pa) to maximizepA(PR) and the intermediary decides the retail priceof product b (ie pb) to maximize pI(PR) Next wedetermine Supplier Brsquos wholesale price wb to maxi-mize pB(PR) Finally using this wholesale price wbthe optimal retail prices pa and pb can be set Theresults of this analysis are presented in the lemmabelow
LEMMA 2 Given a proportion a and Supplier A accept-ing the offer of an online marketplace while the intermedi-ary serves as a reseller for Supplier B there exists aunique equilibrium with prices and demands are asfollows
wPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc
2eth1thorn cTHORNeth2thorn 4cthorn c2THORN h
pPRa frac14 eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORNh
pPRb frac14 eth3thorn 6cthorn 2c2THORNeth2thorn 3cTHORNthorn aeth1thorn c c2THORNcfrac124eth1thorn cTHORN2eth1thorn aTHORNc2eth2thorn 4cthorn c2THORN h
dPRa frac14 eth4thorn 9cthorn 3c2THORNeth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn cTHORNeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORN h
dPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2 h
and the corresponding equilibrium profits are
pPRA frac14
eth1aTHORNeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2F
pPRB frac14 frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc22eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc2eth4thorn8cthorn2c2THORNh
2
pPRI frac14aeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22
frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2
thorn
frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc2eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN
Mh2F
where
M frac14 frac12eth1thorn cTHORNeth2thorn 3cTHORNeth2thorn 4cthorn c2THORN thorn aeth1thorn cTHORNeth2thorn 3cTHORNc2thorn aeth3thorn 4cTHORNeth2thorn 4cthorn c2THORNc a2eth1thorn 2cTHORNc3
For the equilibrium in Lemma 2 to exist the follow-ing proposition characterizes the sufficient conditionsunder which Supplier A will accept the intermedi-aryrsquos marketplace offer and Supplier B will continueto participate under the reseller mode
PROPOSITION 1 Supplier A will accept the intermediaryrsquosoffer of an online marketplace provided pPRA pRRA andSupplier B will continue to participate under the resellermode provided pPRB pO This leads to the following suffi-cient conditions for the existence of the PR mode
bull F le FPR where FPR frac14 eth4thorn 9cthorn 3c2THORN2eth1thorn cTHORNeth2thorn cTHORN2eth4thorn 8cthorn 2c2THORN2 h
2 1thorn c
2eth2thorn cTHORN2 h2 and
bull a aPR frac14 minfrac12aPRA aPRB where aPRA is theunique solution of eth1 aPRA THORNeth1thorn cTHORNfrac12eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aPRA eth1thorn 2cTHORNc22
frac124eth1thorn cTHORN2 eth1thorn aPRATHORNc22eth4thorn 8cthorn 2c2THORN2 h2 F frac14 1thorn c
2eth2thorn cTHORN2
h2 and aPRB is the unique solution offrac12eth1thorn cTHORNeth2thorn 3cTHORN aPRB eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPRBTHORNc2eth4thorn 8cthorn 2c2THORN h
2 frac14 pO
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1601
Proposition 1 characterizes the conditions underwhich Supplier A will accept the intermediaryrsquos mar-ketplace offer and Supplier B will not pursue the out-side option but continue to participate under thereseller mode The first condition relates to the order-fulfillment cost which should not exceed a certainthreshold Recall that by choosing the online market-place service the supplier gains pricing power butbears the order-fulfillment cost at the same timeThus this threshold value of the fulfillment costreflects the trade-off between gains from the pricingpower and the costs (in terms of order fulfillment) ofrealizing such gainsAssuming the first condition is satisfied the propo-
sition then specifies a second condition related to themaximum proportion a which can be charged to thesupplier by the intermediary Interesting insightsregarding this proportion a are shown in Lemma 3
LEMMA 3 (i) pPRI increases in a whereas both pPRA andpPRB decrease in a (ii) aPR decreases in F
Lemma 3(i) indicates that an increase in a willincrease the intermediaryrsquos profit pPRI and simultane-ously decrease Supplier Arsquos profit pPRA Hence ashould be small (ie a aPRA ) to ensure pPRA pRRA sothat Supplier A will accept the intermediaryrsquos market-place offer Meanwhile a counterintuitive result isthat an increase in a would also lead to a reduction inthe profit of the other supplier (B) who has not beenoffered the online marketplace service The reason forthis negative spillover effect is that when a isincreased the supplier (A) tends to increase the priceas the intermediary will keep a higher proportion ofthe revenue More specifically when the intermediaryraises the proportion a she can keep Supplier Aresponds by charging higher retail prices In responsethe intermediary sets a higher retail price for SupplierBrsquos product leading to lower demand for the sameproduct The net result is lower profit for Supplier BThus a needs to be small enough (a aPRB ) so thatSupplier B will not pursue the outside option but con-tinue to participate under the reseller mode orpPRB pO7 In equilibrium assuming the conditionrelated to the order-fulfillment cost is met (ieF le FPR) then the intermediary will set a frac14 aPR frac14minfrac12aPRA aPRB since this will allow the intermediary toextract the maximum profits from the supplier Fur-ther Lemma 3(ii) shows that this threshold value ofthe proportional fee (ie aPR) is decreasing in theorder-fulfillment cost (F)In summary from a managerial perspective our
results indicate that the intermediary requires astrategic focus in choosing the parameters for offeringthe online marketplace service This service will onlybe attractive to suppliers provided the order-
fulfillment cost is not significantly high If this is thecase then the intermediary should set the proportiona equal to its upper bound in order to realize the max-imum profits From a supplier perspective the onlinemarketplace offering is preferred over the reseller set-ting when the flexibility gains due to price setting out-weigh the explicit costs of order fulfillment
43 Mode PPIn this setting both suppliers will be offered theoption of online marketplace from the intermediaryIf both suppliers choose the online marketplace ser-vice they will take into consideration the proportiona which they need to share with the intermediary Fora given proportion a the profits for each supplier andthe intermediary under this mode (ie both supplierschoose the online marketplace service) are as follows
pAethPPTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPPTHORN frac14 eth1 aTHORNpbfrac12h pb thorn cethpa pbTHORN F
pIethPPTHORN frac14 apafrac12h pa thorn cethpb paTHORNthorn apbfrac12h pb thorn cethpa pbTHORN
We solve this game by simultaneously determiningthe retail prices pa and pb which maximize pA(PP)and pB(PP) The results are shown in the lemmabelow
LEMMA 4 Assuming both suppliers accept the propor-tion a there exists a unique equilibrium with retail pricesand demands are as follows
pPPa frac14 pPPb frac14 1
2thorn ch
dPPa frac14 dPPb frac14 1thorn c2thorn c
h
The corresponding equilibrium profits are given by
pPPA frac14 pPPB frac14 eth1 aTHORN 1thorn c
eth2thorn cTHORN2 h2 F
pPPI frac14 2a1thorn c
eth2thorn cTHORN2 h2
When is this mode a possible outcome This can beframed using the decision matrix in Table 1 where weare interested in the equilibrium when both supplierswill accept the online marketplace offer This equilib-rium will arise when both of the following conditionshold Given that Supplier B (A) adopts the resellermodel Supplier Arsquos (Brsquos) profit should be higherunder the condition of acceptance than refusal of theintermediaryrsquos marketplace offer which is pPRA pRRA(pRPB pRRB ) given that Supplier B (A) adopts theonline marketplace Supplier Arsquos (Brsquos) profit should
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1602 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
also be higher when it accepts rather than refuses theintermediaryrsquos marketplace offer which is pPPA pRPA(pPPB pPRB ) In addition to ensure this equilibrium isstable both suppliersrsquo profits under the PP modeshould not be lower than the profit under the outsideoption which is pPPA pO (pPPB pO) Based on theresults in Lemma 4 the following proposition identi-fies the conditions under which both suppliers willnot pursue the outside option but choose the onlinemarketplace service offering
PROPOSITION 2 Both suppliers will choose the onlinemarketplace service provided pPRA pRRA pPPA pRPA andpPPA pO This leads to the following sufficient condi-tions for the existence of the PP mode
bull F le FPP where FPP frac14 1thorn ceth2thorn cTHORN2 h
2 eth1thorn cTHORNeth2thorn 3cTHORN2eth2thorn cTHORNeth4thorn 8cthorn 2c2THORN
h2 and
bull a aPP frac14 minfaPRA aPPA aPPB g where aPPA is the
unique solution of eth1 aPPA THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14frac12eth1thorn cTHORNeth2thorn 3cTHORN aPPA eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPPATHORNc2eth4thorn 8cthorn 2c2THORN h
2 and aPPB is the
unique solution of eth1 aPPB THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14 pO
Similar to the hybrid mode both suppliers willaccept the intermediaryrsquos online marketplace offerwhen the order-fulfillment cost and the proportionalfee that the intermediary charges are all below a cer-tain threshold It is worth noting that the threshold inorder-fulfillment cost is more stringent than that statedin Proposition 1 since FPP lt FPR As before the inter-mediaryrsquos profit pPPI increases in a while both pPPA andpPPB decrease in a In equilibrium the intermediary isonce again incentivized to set a = aPP which decreasesin the order-fulfillment cost (F) We also note that sup-pliersrsquo profits will decrease as the upstream competi-tion intensifies (ie a higher value of c)In the next section we compare the equilibrium
solutions for the three modes with a view to provid-ing managerial insights
5 Insights
51 Pricing and DemandFrom a pricing and demand perspective the proposi-tion below provides a comparison across the threechannel modes
PROPOSITION 3 Comparing the equilibrium results ofretail prices and demand we have (i) pRRa frac14pRRb [ pPRb [ pPRa [ pPPa frac14 pPPb (ii) dPRa [ dPPa frac14dPPb [ dRRa frac14 dRRb [ dPRb
Interestingly the results of Proposition 3 are inde-pendent of the proportional fee rate a and hold regard-less of the parameter settings for competitionintensity (c) and maximum market size (h) Essen-tially market prices are highest in the pure resellermode since they are set by the intermediary whoattempts to moderate the competition between thesuppliers On the other hand when the suppliers setthe retail prices by operating in the online market-place mode they compete with each other directlyleading to lower retail prices Because of doublemarginalization in mode PR the retail price of theproduct that does not join the online marketplace ishigher than the price for which the supplier whochooses to do so Corresponding to these pricingstructures demand coverage is higher in the onlinemarketplace mode (PP) as compared to the resellermode (RR)
52 Intermediary ProfitsWe now proceed to examine how intermediary prof-itability across the three modes is affected by two keyparameters competition intensity (c) and order-fulfill-ment cost (F)8 To assess how these parameters affectprofitability of all supply chain members across thethree modes we start by making the followingassumptions First we assume that the order-fulfill-ment cost is small enough to ensure that both modePR and PP are feasible Analytically the potential val-ues of F under this assumption are that F lt FPP(ltFPR)Note that the value of F which satisfies the firstassumption also leads both aPR and aPP to be positiveOur second and final assumption is that the interme-diary chooses to set a = aPR for mode PR and a = aPP
for mode PP which represent equilibrium outcomesProposition 4 structurally characterizes which
mode choice would be preferred by the intermedi-ary as moderated by the competition intensityparameter c
PROPOSITION 4 There exist competition intensity thresh-olds c gt 0 and c(gtc) such that
bull If c lt c the optimal mode is PP
bull If c gt c the optimal mode is RR and
bull If c le c le c mode PR can be the optimal mode
Proposition 4 indicates that when competitionintensity is low the optimal mode is PP when compe-tition intensity is high the optimal mode is RR Inaddition we observe that when competition intensityis medium the optimal mode is PR The rationale
Table 1 Decision Matrix with Resulting Profits
Supplier A option
Supplier B option
Online marketplace Reseller
Online marketplace pPPA pPPB pPRA pPRB
Reseller pRPA pRPB pRRA pRRB
Note Due to symmetry we know that pPRA frac14 pRPB and pPRB frac14 pRPA
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1603
behind these results hinges on the moderating role ofthe intermediary Previous literature has shown thatthe intermediary prefers high levels of upstream com-petition as it strengthens its channel power (Wangand Shin 2015) From a mode perspective supplierscompete more fiercely in the online marketplacemode compared with the reseller mode as they set theprices directly and alternative modes allow the inter-mediary to moderate this competitive effect To elabo-rate when both suppliers are very competitive theintermediary would prefer to operate as a resellersince any other mode will lead to increased price com-petition which in turn negatively impacts the inter-mediary This result validates the conjecture of Hagiu(2007) that strong substitutability between suppliersrsquoproducts would lead to a greater preference for thepure reseller mode When the competition intensitybetween suppliers is low the intermediaryrsquos moderat-ing role is not as important Note that in mode PP theproportional fee could be considered a form of rev-enue sharing which would mitigate the impact ofdouble-marginalization (Geng et al 2018 Shy andWang 2011 Tan and Carrillo 2017) Hence the inter-mediary prefers mode PP when competition intensityis low When the competition intensity between sup-pliers is in the intermediate range the intermediaryrsquosmoderating role is important but not significantenough to alleviate double-marginalization effectsleading to a preference for the hybrid mode PRAnecdotal observations provide face validity for
our results For the household appliance market inChina the similarity between branded products leadsto strong substitutability (or high competition inten-sity) Hence most online retailers (eg JDcom) preferto act as resellers for products in this category On theother hand the media content industry (publishingvideo games) is characterized by low levels of pro-duct substitutability (or low competition intensity)This has motivated online retailers to set the terms forthe online marketplace mode in such a way that sup-pliers have chosen that mode of operation Etsycomthe largest e-commerce website in handmade or vin-tage items serves as a pure online marketplacebecause of low product substitutability FurtherAmazon serves as the reseller in certain high competi-tive product categories such as electronics but servesas a marketplace for other less competitive productcategories such as fine art (ie long-tail productcategory)Next we consider the impact of the order-fulfill-
ment cost on the choice of equilibrium mode for theintermediary Note that Proposition 4 shows thatwhen the market competition intensity is above a cer-tain threshold the intermediary will always prefer toact as a reseller To investigate the impact of theorder-fulfillment cost next we focus on the case that
the market competition intensity is not too high Theproposition below characterizes the mode preferredby the intermediary depending on the magnitude ofthe order-fulfillment cost
PROPOSITION 5 When the market competition intensityis relatively small (ie c cjFfrac140) there exists twothresholds F gt 0 and F(geF) such that
bull when F le F the preferred mode is PP and
bull when F lt F le F the preferred mode is PR and
bull when F gt F the preferred mode is RR
Proposition 5 shows that the magnitude of theorder-fulfillment cost will alter the mode choicesWhen the market competition intensity is relativelylow (c cjFfrac140) as the order-fulfillment costincreases the equilibrium mode will evolve from apure marketplace mode to a hybrid mode then to apure reseller mode The intuition of this result lies inthe fact that the higher the order-fulfillment costbecomes the more difficult it becomes for the inter-mediary to convince the suppliers to adopt the mar-ketplace mode and endure the order-fulfillment costRecall that both aPR and aPP decrease in F That is asthe order-fulfillment cost increases the intermediaryhas to set a much lower proportional fee to ensurethat the suppliers will accept the marketplace offerThis becomes increasingly expensive as the order-ful-fillment cost increases As a result the intermediaryprefers the marketplace mode for both supplierswhen the order-fulfillment cost is low and favors thereseller mode when the order-fulfillment cost issignificantFrom a practical perspective we observe that most
hotel and airline companies use the PP mode on tra-vel sites such as Expedia Travelocity Ctrip andQunar (Ctrip and Qunar are leading travel sites inChina) This is partially driven by the low (almostnegligible) order-fulfillment cost for electronic tick-etsbookings In the publishing industry a transitionto the PP mode from the RR mode is evident for e-books while the RR mode is still preferred for printbooks This could be due in part to the fact that order-fulfillment costs are greater for printed books as com-pared to e-booksHaving analytically examined the impacts of the
competition intensity and order-fulfillment cost indi-vidually we next numerically illustrate how the inter-actions between these two parameters affect the modechoice for the intermediary Figure 2 was generatedby setting h = 10 (ie the market potential equals to10) and pO = 0 (ie the outside option facing thesupplier is 0) Then we simultaneously varyF (F 2 eth0 FPPjcfrac140THORN where FPPjcfrac140 frac14 125) and c(c 2 (0 2)) Note that our results are robust to theparameter changes
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1604 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
fee For example Amazon Marketplace charges a$3999 monthly subscription fee for sellers who planto sell more than 40 items a month Searscom alsocharges a $3999 fee to suppliers whose sales are over$400 during one month Because this fee is relativelysmall compared with the suppliersrsquo sales volumewithout loss of generality we normalize this fixed feepart to zero Previous literature has also adopted thisassumption (Abhishek et al 2016)Order fulfillment that is delivering physical goods
to the customer is commonly cited as one of the mostexpensive and critical operations of online sellers(Agatz et al 2008) The cost of order fulfillment canrun as high as 25 of sales (Kapner 2014) To fulfill anorder firms take on costs such as warehouse build-ingrenting costs hiring staff to handle packages anddelivering the products to customers The costs of stor-age and hiring staff are quite significant and can beviewed as fixed cost Meanwhile the delivery cost foronline shopping is usually undertaken by the cus-tomers For ease of exposition we assume that order-fulfillment costs are fixed since we find that the keyinsights stemming from our analysis are similar evenif these costs are a function of the number of ordersreceived4 Essentially if the intermediary functions asa reseller he will incur a fixed cost FI gt 0 to fulfill theorder for per product category (product a or b) if theintermediary functions as an online marketplace eachsupplier must be responsible for a fixed cost FS gt 0 tofulfill orders Through analysis we find that if theintermediaryrsquos order-fulfillment cost is lower than thesupplierrsquos order-fulfillment cost (ie FI lt FS) the equi-librium mode is more likely to be the pure resellermode and vice versa To eliminate the possibility thatthe equilibrium mode is driven by the asymmetric coststructure we assume FI = FS = F in our base modelMore specifically in mode RR the intermediary bearsorder-fulfillment costs for both product categories (ie2F) in mode PR Supplier A bears the order-fulfillmentcost (F) for its product and the intermediary bears theorder-fulfillment cost (F) for reselling Supplier Brsquos pro-duct in mode PP both Supplier A and Supplier B bearthe order-fulfillment cost (F) of their own product cate-gory5 Recently some online intermediaries (egAmazon Sears JD etc) and 3rd party online order ful-fillment service providers have offered an option tosuppliers that allows them to outsource their order ful-fillment In this case the supplier will pay some feesfor order fulfillment that is even if the supplier out-sources the order fulfillment he still needs to bear therelated costs For ease of exposition and without lossof generality we focus on the case that suppliers willfulfill the order in-house in our base model6
In the next section the equilibrium outcomes ofeach channel mode are structurally characterized
4 Equilibrium Analysis
The first set of results described below is our bench-mark (current) setting where the intermediary acts asa reseller for both suppliers (mode RR) Next we con-sider the setting where one supplier switches to theonline marketplace mode while the other supplier stilladopts the reseller mode (model PR) Finally wefocus on the case in which both suppliers choose toaccept the intermediary terms and switch to theonline marketplace mode (mode PP) In derivingthese results the necessary conditions for the exis-tence of each operating mode (ie RR PR and PP)are that the equilibrium profits for each supplier areat least as large as the profits each supplier could real-ize through exercising an outside option (denoted bypO)
41 Mode RRUnder this setting the suppliers simultaneouslyquote wholesale prices (wa and wb) and then the inter-mediary (who bears the order-fulfillment costs forboth products) determines the retail prices pa and pbThe profits for the suppliers and the intermediary areas follows
pAethRRTHORN frac14 wafrac12h pa thorn cethpb paTHORNpBethRRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethRRTHORN frac14 ethpa waTHORNfrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN 2F
We solve this game by backward induction Forany given wholesale prices wa and wb we first charac-terize the equilibrium retail prices that would maxi-mize pI(RR) We then determine the wholesale pricesfor the suppliers by simultaneously maximizing theirindividual profit functions
LEMMA 1 There exists a unique equilibrium for modeRR The equilibrium prices and demands are as follows
wRRa frac14 wRR
b frac14 1
2thorn ch
pRRa frac14 pRRb frac14 3thorn c2eth2thorn cTHORN h
dRRa frac14 dRRb frac14 1thorn c2eth2thorn cTHORN h
and the corresponding optimal profits are given by
pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2
pRRI frac14 eth1thorn cTHORN22eth2thorn cTHORN2 h
2 2F
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1600 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
It is straightforward to note that an increase incompetition intensity (c) would lead to an increasein the intermediaryrsquos profit and a simultaneousdecrease in supplier profits since an increase in creduces the suppliersrsquo pricing power relative tothat of the intermediary The profit for each sup-plier under this RR benchmark mode serves as abase for evaluating whether a supplier wouldchoose the online marketplace if it was offered bythe intermediaryThe technical condition for the existence of the
RR mode and its corresponding equilibrium is thatthe supplierrsquos profit under this mode is at least aslarge as the profit that each supplier could realizeby exercising the outside option that ispO pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2 Given that the reseller
mode is adopted by a large number of suppliersselling through online intermediaries this exis-tence condition is assumed to hold for the remain-der of this study
42 Mode PRIn this setting only one supplier will be offeredthe option of the online marketplace mode fromthe intermediary Hence one supplier adopts theonline marketplace mode if accepted while theother retains the reseller mode Without loss ofgenerality we assume that the online marketplaceoption is accepted by Supplier A and the interme-diary operates as a reseller for Supplier B For agiven proportion a the suppliersrsquo and intermedi-aryrsquos profits are
pAethPRTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethPRTHORN frac14 apafrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN F
We solve this game by backward induction as fol-lows For any given wb we first characterize thesimultaneous pricing decision where Supplier A deci-des the retail price of product a (ie pa) to maximizepA(PR) and the intermediary decides the retail priceof product b (ie pb) to maximize pI(PR) Next wedetermine Supplier Brsquos wholesale price wb to maxi-mize pB(PR) Finally using this wholesale price wbthe optimal retail prices pa and pb can be set Theresults of this analysis are presented in the lemmabelow
LEMMA 2 Given a proportion a and Supplier A accept-ing the offer of an online marketplace while the intermedi-ary serves as a reseller for Supplier B there exists aunique equilibrium with prices and demands are asfollows
wPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc
2eth1thorn cTHORNeth2thorn 4cthorn c2THORN h
pPRa frac14 eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORNh
pPRb frac14 eth3thorn 6cthorn 2c2THORNeth2thorn 3cTHORNthorn aeth1thorn c c2THORNcfrac124eth1thorn cTHORN2eth1thorn aTHORNc2eth2thorn 4cthorn c2THORN h
dPRa frac14 eth4thorn 9cthorn 3c2THORNeth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn cTHORNeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORN h
dPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2 h
and the corresponding equilibrium profits are
pPRA frac14
eth1aTHORNeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2F
pPRB frac14 frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc22eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc2eth4thorn8cthorn2c2THORNh
2
pPRI frac14aeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22
frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2
thorn
frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc2eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN
Mh2F
where
M frac14 frac12eth1thorn cTHORNeth2thorn 3cTHORNeth2thorn 4cthorn c2THORN thorn aeth1thorn cTHORNeth2thorn 3cTHORNc2thorn aeth3thorn 4cTHORNeth2thorn 4cthorn c2THORNc a2eth1thorn 2cTHORNc3
For the equilibrium in Lemma 2 to exist the follow-ing proposition characterizes the sufficient conditionsunder which Supplier A will accept the intermedi-aryrsquos marketplace offer and Supplier B will continueto participate under the reseller mode
PROPOSITION 1 Supplier A will accept the intermediaryrsquosoffer of an online marketplace provided pPRA pRRA andSupplier B will continue to participate under the resellermode provided pPRB pO This leads to the following suffi-cient conditions for the existence of the PR mode
bull F le FPR where FPR frac14 eth4thorn 9cthorn 3c2THORN2eth1thorn cTHORNeth2thorn cTHORN2eth4thorn 8cthorn 2c2THORN2 h
2 1thorn c
2eth2thorn cTHORN2 h2 and
bull a aPR frac14 minfrac12aPRA aPRB where aPRA is theunique solution of eth1 aPRA THORNeth1thorn cTHORNfrac12eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aPRA eth1thorn 2cTHORNc22
frac124eth1thorn cTHORN2 eth1thorn aPRATHORNc22eth4thorn 8cthorn 2c2THORN2 h2 F frac14 1thorn c
2eth2thorn cTHORN2
h2 and aPRB is the unique solution offrac12eth1thorn cTHORNeth2thorn 3cTHORN aPRB eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPRBTHORNc2eth4thorn 8cthorn 2c2THORN h
2 frac14 pO
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1601
Proposition 1 characterizes the conditions underwhich Supplier A will accept the intermediaryrsquos mar-ketplace offer and Supplier B will not pursue the out-side option but continue to participate under thereseller mode The first condition relates to the order-fulfillment cost which should not exceed a certainthreshold Recall that by choosing the online market-place service the supplier gains pricing power butbears the order-fulfillment cost at the same timeThus this threshold value of the fulfillment costreflects the trade-off between gains from the pricingpower and the costs (in terms of order fulfillment) ofrealizing such gainsAssuming the first condition is satisfied the propo-
sition then specifies a second condition related to themaximum proportion a which can be charged to thesupplier by the intermediary Interesting insightsregarding this proportion a are shown in Lemma 3
LEMMA 3 (i) pPRI increases in a whereas both pPRA andpPRB decrease in a (ii) aPR decreases in F
Lemma 3(i) indicates that an increase in a willincrease the intermediaryrsquos profit pPRI and simultane-ously decrease Supplier Arsquos profit pPRA Hence ashould be small (ie a aPRA ) to ensure pPRA pRRA sothat Supplier A will accept the intermediaryrsquos market-place offer Meanwhile a counterintuitive result isthat an increase in a would also lead to a reduction inthe profit of the other supplier (B) who has not beenoffered the online marketplace service The reason forthis negative spillover effect is that when a isincreased the supplier (A) tends to increase the priceas the intermediary will keep a higher proportion ofthe revenue More specifically when the intermediaryraises the proportion a she can keep Supplier Aresponds by charging higher retail prices In responsethe intermediary sets a higher retail price for SupplierBrsquos product leading to lower demand for the sameproduct The net result is lower profit for Supplier BThus a needs to be small enough (a aPRB ) so thatSupplier B will not pursue the outside option but con-tinue to participate under the reseller mode orpPRB pO7 In equilibrium assuming the conditionrelated to the order-fulfillment cost is met (ieF le FPR) then the intermediary will set a frac14 aPR frac14minfrac12aPRA aPRB since this will allow the intermediary toextract the maximum profits from the supplier Fur-ther Lemma 3(ii) shows that this threshold value ofthe proportional fee (ie aPR) is decreasing in theorder-fulfillment cost (F)In summary from a managerial perspective our
results indicate that the intermediary requires astrategic focus in choosing the parameters for offeringthe online marketplace service This service will onlybe attractive to suppliers provided the order-
fulfillment cost is not significantly high If this is thecase then the intermediary should set the proportiona equal to its upper bound in order to realize the max-imum profits From a supplier perspective the onlinemarketplace offering is preferred over the reseller set-ting when the flexibility gains due to price setting out-weigh the explicit costs of order fulfillment
43 Mode PPIn this setting both suppliers will be offered theoption of online marketplace from the intermediaryIf both suppliers choose the online marketplace ser-vice they will take into consideration the proportiona which they need to share with the intermediary Fora given proportion a the profits for each supplier andthe intermediary under this mode (ie both supplierschoose the online marketplace service) are as follows
pAethPPTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPPTHORN frac14 eth1 aTHORNpbfrac12h pb thorn cethpa pbTHORN F
pIethPPTHORN frac14 apafrac12h pa thorn cethpb paTHORNthorn apbfrac12h pb thorn cethpa pbTHORN
We solve this game by simultaneously determiningthe retail prices pa and pb which maximize pA(PP)and pB(PP) The results are shown in the lemmabelow
LEMMA 4 Assuming both suppliers accept the propor-tion a there exists a unique equilibrium with retail pricesand demands are as follows
pPPa frac14 pPPb frac14 1
2thorn ch
dPPa frac14 dPPb frac14 1thorn c2thorn c
h
The corresponding equilibrium profits are given by
pPPA frac14 pPPB frac14 eth1 aTHORN 1thorn c
eth2thorn cTHORN2 h2 F
pPPI frac14 2a1thorn c
eth2thorn cTHORN2 h2
When is this mode a possible outcome This can beframed using the decision matrix in Table 1 where weare interested in the equilibrium when both supplierswill accept the online marketplace offer This equilib-rium will arise when both of the following conditionshold Given that Supplier B (A) adopts the resellermodel Supplier Arsquos (Brsquos) profit should be higherunder the condition of acceptance than refusal of theintermediaryrsquos marketplace offer which is pPRA pRRA(pRPB pRRB ) given that Supplier B (A) adopts theonline marketplace Supplier Arsquos (Brsquos) profit should
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1602 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
also be higher when it accepts rather than refuses theintermediaryrsquos marketplace offer which is pPPA pRPA(pPPB pPRB ) In addition to ensure this equilibrium isstable both suppliersrsquo profits under the PP modeshould not be lower than the profit under the outsideoption which is pPPA pO (pPPB pO) Based on theresults in Lemma 4 the following proposition identi-fies the conditions under which both suppliers willnot pursue the outside option but choose the onlinemarketplace service offering
PROPOSITION 2 Both suppliers will choose the onlinemarketplace service provided pPRA pRRA pPPA pRPA andpPPA pO This leads to the following sufficient condi-tions for the existence of the PP mode
bull F le FPP where FPP frac14 1thorn ceth2thorn cTHORN2 h
2 eth1thorn cTHORNeth2thorn 3cTHORN2eth2thorn cTHORNeth4thorn 8cthorn 2c2THORN
h2 and
bull a aPP frac14 minfaPRA aPPA aPPB g where aPPA is the
unique solution of eth1 aPPA THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14frac12eth1thorn cTHORNeth2thorn 3cTHORN aPPA eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPPATHORNc2eth4thorn 8cthorn 2c2THORN h
2 and aPPB is the
unique solution of eth1 aPPB THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14 pO
Similar to the hybrid mode both suppliers willaccept the intermediaryrsquos online marketplace offerwhen the order-fulfillment cost and the proportionalfee that the intermediary charges are all below a cer-tain threshold It is worth noting that the threshold inorder-fulfillment cost is more stringent than that statedin Proposition 1 since FPP lt FPR As before the inter-mediaryrsquos profit pPPI increases in a while both pPPA andpPPB decrease in a In equilibrium the intermediary isonce again incentivized to set a = aPP which decreasesin the order-fulfillment cost (F) We also note that sup-pliersrsquo profits will decrease as the upstream competi-tion intensifies (ie a higher value of c)In the next section we compare the equilibrium
solutions for the three modes with a view to provid-ing managerial insights
5 Insights
51 Pricing and DemandFrom a pricing and demand perspective the proposi-tion below provides a comparison across the threechannel modes
PROPOSITION 3 Comparing the equilibrium results ofretail prices and demand we have (i) pRRa frac14pRRb [ pPRb [ pPRa [ pPPa frac14 pPPb (ii) dPRa [ dPPa frac14dPPb [ dRRa frac14 dRRb [ dPRb
Interestingly the results of Proposition 3 are inde-pendent of the proportional fee rate a and hold regard-less of the parameter settings for competitionintensity (c) and maximum market size (h) Essen-tially market prices are highest in the pure resellermode since they are set by the intermediary whoattempts to moderate the competition between thesuppliers On the other hand when the suppliers setthe retail prices by operating in the online market-place mode they compete with each other directlyleading to lower retail prices Because of doublemarginalization in mode PR the retail price of theproduct that does not join the online marketplace ishigher than the price for which the supplier whochooses to do so Corresponding to these pricingstructures demand coverage is higher in the onlinemarketplace mode (PP) as compared to the resellermode (RR)
52 Intermediary ProfitsWe now proceed to examine how intermediary prof-itability across the three modes is affected by two keyparameters competition intensity (c) and order-fulfill-ment cost (F)8 To assess how these parameters affectprofitability of all supply chain members across thethree modes we start by making the followingassumptions First we assume that the order-fulfill-ment cost is small enough to ensure that both modePR and PP are feasible Analytically the potential val-ues of F under this assumption are that F lt FPP(ltFPR)Note that the value of F which satisfies the firstassumption also leads both aPR and aPP to be positiveOur second and final assumption is that the interme-diary chooses to set a = aPR for mode PR and a = aPP
for mode PP which represent equilibrium outcomesProposition 4 structurally characterizes which
mode choice would be preferred by the intermedi-ary as moderated by the competition intensityparameter c
PROPOSITION 4 There exist competition intensity thresh-olds c gt 0 and c(gtc) such that
bull If c lt c the optimal mode is PP
bull If c gt c the optimal mode is RR and
bull If c le c le c mode PR can be the optimal mode
Proposition 4 indicates that when competitionintensity is low the optimal mode is PP when compe-tition intensity is high the optimal mode is RR Inaddition we observe that when competition intensityis medium the optimal mode is PR The rationale
Table 1 Decision Matrix with Resulting Profits
Supplier A option
Supplier B option
Online marketplace Reseller
Online marketplace pPPA pPPB pPRA pPRB
Reseller pRPA pRPB pRRA pRRB
Note Due to symmetry we know that pPRA frac14 pRPB and pPRB frac14 pRPA
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1603
behind these results hinges on the moderating role ofthe intermediary Previous literature has shown thatthe intermediary prefers high levels of upstream com-petition as it strengthens its channel power (Wangand Shin 2015) From a mode perspective supplierscompete more fiercely in the online marketplacemode compared with the reseller mode as they set theprices directly and alternative modes allow the inter-mediary to moderate this competitive effect To elabo-rate when both suppliers are very competitive theintermediary would prefer to operate as a resellersince any other mode will lead to increased price com-petition which in turn negatively impacts the inter-mediary This result validates the conjecture of Hagiu(2007) that strong substitutability between suppliersrsquoproducts would lead to a greater preference for thepure reseller mode When the competition intensitybetween suppliers is low the intermediaryrsquos moderat-ing role is not as important Note that in mode PP theproportional fee could be considered a form of rev-enue sharing which would mitigate the impact ofdouble-marginalization (Geng et al 2018 Shy andWang 2011 Tan and Carrillo 2017) Hence the inter-mediary prefers mode PP when competition intensityis low When the competition intensity between sup-pliers is in the intermediate range the intermediaryrsquosmoderating role is important but not significantenough to alleviate double-marginalization effectsleading to a preference for the hybrid mode PRAnecdotal observations provide face validity for
our results For the household appliance market inChina the similarity between branded products leadsto strong substitutability (or high competition inten-sity) Hence most online retailers (eg JDcom) preferto act as resellers for products in this category On theother hand the media content industry (publishingvideo games) is characterized by low levels of pro-duct substitutability (or low competition intensity)This has motivated online retailers to set the terms forthe online marketplace mode in such a way that sup-pliers have chosen that mode of operation Etsycomthe largest e-commerce website in handmade or vin-tage items serves as a pure online marketplacebecause of low product substitutability FurtherAmazon serves as the reseller in certain high competi-tive product categories such as electronics but servesas a marketplace for other less competitive productcategories such as fine art (ie long-tail productcategory)Next we consider the impact of the order-fulfill-
ment cost on the choice of equilibrium mode for theintermediary Note that Proposition 4 shows thatwhen the market competition intensity is above a cer-tain threshold the intermediary will always prefer toact as a reseller To investigate the impact of theorder-fulfillment cost next we focus on the case that
the market competition intensity is not too high Theproposition below characterizes the mode preferredby the intermediary depending on the magnitude ofthe order-fulfillment cost
PROPOSITION 5 When the market competition intensityis relatively small (ie c cjFfrac140) there exists twothresholds F gt 0 and F(geF) such that
bull when F le F the preferred mode is PP and
bull when F lt F le F the preferred mode is PR and
bull when F gt F the preferred mode is RR
Proposition 5 shows that the magnitude of theorder-fulfillment cost will alter the mode choicesWhen the market competition intensity is relativelylow (c cjFfrac140) as the order-fulfillment costincreases the equilibrium mode will evolve from apure marketplace mode to a hybrid mode then to apure reseller mode The intuition of this result lies inthe fact that the higher the order-fulfillment costbecomes the more difficult it becomes for the inter-mediary to convince the suppliers to adopt the mar-ketplace mode and endure the order-fulfillment costRecall that both aPR and aPP decrease in F That is asthe order-fulfillment cost increases the intermediaryhas to set a much lower proportional fee to ensurethat the suppliers will accept the marketplace offerThis becomes increasingly expensive as the order-ful-fillment cost increases As a result the intermediaryprefers the marketplace mode for both supplierswhen the order-fulfillment cost is low and favors thereseller mode when the order-fulfillment cost issignificantFrom a practical perspective we observe that most
hotel and airline companies use the PP mode on tra-vel sites such as Expedia Travelocity Ctrip andQunar (Ctrip and Qunar are leading travel sites inChina) This is partially driven by the low (almostnegligible) order-fulfillment cost for electronic tick-etsbookings In the publishing industry a transitionto the PP mode from the RR mode is evident for e-books while the RR mode is still preferred for printbooks This could be due in part to the fact that order-fulfillment costs are greater for printed books as com-pared to e-booksHaving analytically examined the impacts of the
competition intensity and order-fulfillment cost indi-vidually we next numerically illustrate how the inter-actions between these two parameters affect the modechoice for the intermediary Figure 2 was generatedby setting h = 10 (ie the market potential equals to10) and pO = 0 (ie the outside option facing thesupplier is 0) Then we simultaneously varyF (F 2 eth0 FPPjcfrac140THORN where FPPjcfrac140 frac14 125) and c(c 2 (0 2)) Note that our results are robust to theparameter changes
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1604 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
It is straightforward to note that an increase incompetition intensity (c) would lead to an increasein the intermediaryrsquos profit and a simultaneousdecrease in supplier profits since an increase in creduces the suppliersrsquo pricing power relative tothat of the intermediary The profit for each sup-plier under this RR benchmark mode serves as abase for evaluating whether a supplier wouldchoose the online marketplace if it was offered bythe intermediaryThe technical condition for the existence of the
RR mode and its corresponding equilibrium is thatthe supplierrsquos profit under this mode is at least aslarge as the profit that each supplier could realizeby exercising the outside option that ispO pRRA frac14 pRRB frac14 1thorn c
2eth2thorn cTHORN2 h2 Given that the reseller
mode is adopted by a large number of suppliersselling through online intermediaries this exis-tence condition is assumed to hold for the remain-der of this study
42 Mode PRIn this setting only one supplier will be offeredthe option of the online marketplace mode fromthe intermediary Hence one supplier adopts theonline marketplace mode if accepted while theother retains the reseller mode Without loss ofgenerality we assume that the online marketplaceoption is accepted by Supplier A and the interme-diary operates as a reseller for Supplier B For agiven proportion a the suppliersrsquo and intermedi-aryrsquos profits are
pAethPRTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPRTHORN frac14 wbfrac12h pb thorn cethpa pbTHORNpIethPRTHORN frac14 apafrac12h pa thorn cethpb paTHORN
thorn ethpb wbTHORNfrac12h pb thorn cethpa pbTHORN F
We solve this game by backward induction as fol-lows For any given wb we first characterize thesimultaneous pricing decision where Supplier A deci-des the retail price of product a (ie pa) to maximizepA(PR) and the intermediary decides the retail priceof product b (ie pb) to maximize pI(PR) Next wedetermine Supplier Brsquos wholesale price wb to maxi-mize pB(PR) Finally using this wholesale price wbthe optimal retail prices pa and pb can be set Theresults of this analysis are presented in the lemmabelow
LEMMA 2 Given a proportion a and Supplier A accept-ing the offer of an online marketplace while the intermedi-ary serves as a reseller for Supplier B there exists aunique equilibrium with prices and demands are asfollows
wPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc
2eth1thorn cTHORNeth2thorn 4cthorn c2THORN h
pPRa frac14 eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORNh
pPRb frac14 eth3thorn 6cthorn 2c2THORNeth2thorn 3cTHORNthorn aeth1thorn c c2THORNcfrac124eth1thorn cTHORN2eth1thorn aTHORNc2eth2thorn 4cthorn c2THORN h
dPRa frac14 eth4thorn 9cthorn 3c2THORNeth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn cTHORNeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2eth4thorn 8cthorn 2c2THORN h
dPRb frac14 eth1thorn cTHORNeth2thorn 3cTHORN aeth1thorn 2cTHORNc2frac124eth1thorn cTHORN2eth1thorn aTHORNc2 h
and the corresponding equilibrium profits are
pPRA frac14
eth1aTHORNeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2F
pPRB frac14 frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc22eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc2eth4thorn8cthorn2c2THORNh
2
pPRI frac14aeth1thorncTHORNfrac12eth4thorn9cthorn3c2THORNeth2thorn3cTHORNaeth1thorn2cTHORNc22
frac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN2 h2
thorn
frac12eth1thorncTHORNeth2thorn3cTHORNaeth1thorn2cTHORNc2eth1thorncTHORNfrac124eth1thorncTHORN2eth1thornaTHORNc22eth4thorn8cthorn2c2THORN
Mh2F
where
M frac14 frac12eth1thorn cTHORNeth2thorn 3cTHORNeth2thorn 4cthorn c2THORN thorn aeth1thorn cTHORNeth2thorn 3cTHORNc2thorn aeth3thorn 4cTHORNeth2thorn 4cthorn c2THORNc a2eth1thorn 2cTHORNc3
For the equilibrium in Lemma 2 to exist the follow-ing proposition characterizes the sufficient conditionsunder which Supplier A will accept the intermedi-aryrsquos marketplace offer and Supplier B will continueto participate under the reseller mode
PROPOSITION 1 Supplier A will accept the intermediaryrsquosoffer of an online marketplace provided pPRA pRRA andSupplier B will continue to participate under the resellermode provided pPRB pO This leads to the following suffi-cient conditions for the existence of the PR mode
bull F le FPR where FPR frac14 eth4thorn 9cthorn 3c2THORN2eth1thorn cTHORNeth2thorn cTHORN2eth4thorn 8cthorn 2c2THORN2 h
2 1thorn c
2eth2thorn cTHORN2 h2 and
bull a aPR frac14 minfrac12aPRA aPRB where aPRA is theunique solution of eth1 aPRA THORNeth1thorn cTHORNfrac12eth4thorn 9cthorn 3c2THORNeth2thorn 3cTHORN aPRA eth1thorn 2cTHORNc22
frac124eth1thorn cTHORN2 eth1thorn aPRATHORNc22eth4thorn 8cthorn 2c2THORN2 h2 F frac14 1thorn c
2eth2thorn cTHORN2
h2 and aPRB is the unique solution offrac12eth1thorn cTHORNeth2thorn 3cTHORN aPRB eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPRBTHORNc2eth4thorn 8cthorn 2c2THORN h
2 frac14 pO
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1601
Proposition 1 characterizes the conditions underwhich Supplier A will accept the intermediaryrsquos mar-ketplace offer and Supplier B will not pursue the out-side option but continue to participate under thereseller mode The first condition relates to the order-fulfillment cost which should not exceed a certainthreshold Recall that by choosing the online market-place service the supplier gains pricing power butbears the order-fulfillment cost at the same timeThus this threshold value of the fulfillment costreflects the trade-off between gains from the pricingpower and the costs (in terms of order fulfillment) ofrealizing such gainsAssuming the first condition is satisfied the propo-
sition then specifies a second condition related to themaximum proportion a which can be charged to thesupplier by the intermediary Interesting insightsregarding this proportion a are shown in Lemma 3
LEMMA 3 (i) pPRI increases in a whereas both pPRA andpPRB decrease in a (ii) aPR decreases in F
Lemma 3(i) indicates that an increase in a willincrease the intermediaryrsquos profit pPRI and simultane-ously decrease Supplier Arsquos profit pPRA Hence ashould be small (ie a aPRA ) to ensure pPRA pRRA sothat Supplier A will accept the intermediaryrsquos market-place offer Meanwhile a counterintuitive result isthat an increase in a would also lead to a reduction inthe profit of the other supplier (B) who has not beenoffered the online marketplace service The reason forthis negative spillover effect is that when a isincreased the supplier (A) tends to increase the priceas the intermediary will keep a higher proportion ofthe revenue More specifically when the intermediaryraises the proportion a she can keep Supplier Aresponds by charging higher retail prices In responsethe intermediary sets a higher retail price for SupplierBrsquos product leading to lower demand for the sameproduct The net result is lower profit for Supplier BThus a needs to be small enough (a aPRB ) so thatSupplier B will not pursue the outside option but con-tinue to participate under the reseller mode orpPRB pO7 In equilibrium assuming the conditionrelated to the order-fulfillment cost is met (ieF le FPR) then the intermediary will set a frac14 aPR frac14minfrac12aPRA aPRB since this will allow the intermediary toextract the maximum profits from the supplier Fur-ther Lemma 3(ii) shows that this threshold value ofthe proportional fee (ie aPR) is decreasing in theorder-fulfillment cost (F)In summary from a managerial perspective our
results indicate that the intermediary requires astrategic focus in choosing the parameters for offeringthe online marketplace service This service will onlybe attractive to suppliers provided the order-
fulfillment cost is not significantly high If this is thecase then the intermediary should set the proportiona equal to its upper bound in order to realize the max-imum profits From a supplier perspective the onlinemarketplace offering is preferred over the reseller set-ting when the flexibility gains due to price setting out-weigh the explicit costs of order fulfillment
43 Mode PPIn this setting both suppliers will be offered theoption of online marketplace from the intermediaryIf both suppliers choose the online marketplace ser-vice they will take into consideration the proportiona which they need to share with the intermediary Fora given proportion a the profits for each supplier andthe intermediary under this mode (ie both supplierschoose the online marketplace service) are as follows
pAethPPTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPPTHORN frac14 eth1 aTHORNpbfrac12h pb thorn cethpa pbTHORN F
pIethPPTHORN frac14 apafrac12h pa thorn cethpb paTHORNthorn apbfrac12h pb thorn cethpa pbTHORN
We solve this game by simultaneously determiningthe retail prices pa and pb which maximize pA(PP)and pB(PP) The results are shown in the lemmabelow
LEMMA 4 Assuming both suppliers accept the propor-tion a there exists a unique equilibrium with retail pricesand demands are as follows
pPPa frac14 pPPb frac14 1
2thorn ch
dPPa frac14 dPPb frac14 1thorn c2thorn c
h
The corresponding equilibrium profits are given by
pPPA frac14 pPPB frac14 eth1 aTHORN 1thorn c
eth2thorn cTHORN2 h2 F
pPPI frac14 2a1thorn c
eth2thorn cTHORN2 h2
When is this mode a possible outcome This can beframed using the decision matrix in Table 1 where weare interested in the equilibrium when both supplierswill accept the online marketplace offer This equilib-rium will arise when both of the following conditionshold Given that Supplier B (A) adopts the resellermodel Supplier Arsquos (Brsquos) profit should be higherunder the condition of acceptance than refusal of theintermediaryrsquos marketplace offer which is pPRA pRRA(pRPB pRRB ) given that Supplier B (A) adopts theonline marketplace Supplier Arsquos (Brsquos) profit should
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1602 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
also be higher when it accepts rather than refuses theintermediaryrsquos marketplace offer which is pPPA pRPA(pPPB pPRB ) In addition to ensure this equilibrium isstable both suppliersrsquo profits under the PP modeshould not be lower than the profit under the outsideoption which is pPPA pO (pPPB pO) Based on theresults in Lemma 4 the following proposition identi-fies the conditions under which both suppliers willnot pursue the outside option but choose the onlinemarketplace service offering
PROPOSITION 2 Both suppliers will choose the onlinemarketplace service provided pPRA pRRA pPPA pRPA andpPPA pO This leads to the following sufficient condi-tions for the existence of the PP mode
bull F le FPP where FPP frac14 1thorn ceth2thorn cTHORN2 h
2 eth1thorn cTHORNeth2thorn 3cTHORN2eth2thorn cTHORNeth4thorn 8cthorn 2c2THORN
h2 and
bull a aPP frac14 minfaPRA aPPA aPPB g where aPPA is the
unique solution of eth1 aPPA THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14frac12eth1thorn cTHORNeth2thorn 3cTHORN aPPA eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPPATHORNc2eth4thorn 8cthorn 2c2THORN h
2 and aPPB is the
unique solution of eth1 aPPB THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14 pO
Similar to the hybrid mode both suppliers willaccept the intermediaryrsquos online marketplace offerwhen the order-fulfillment cost and the proportionalfee that the intermediary charges are all below a cer-tain threshold It is worth noting that the threshold inorder-fulfillment cost is more stringent than that statedin Proposition 1 since FPP lt FPR As before the inter-mediaryrsquos profit pPPI increases in a while both pPPA andpPPB decrease in a In equilibrium the intermediary isonce again incentivized to set a = aPP which decreasesin the order-fulfillment cost (F) We also note that sup-pliersrsquo profits will decrease as the upstream competi-tion intensifies (ie a higher value of c)In the next section we compare the equilibrium
solutions for the three modes with a view to provid-ing managerial insights
5 Insights
51 Pricing and DemandFrom a pricing and demand perspective the proposi-tion below provides a comparison across the threechannel modes
PROPOSITION 3 Comparing the equilibrium results ofretail prices and demand we have (i) pRRa frac14pRRb [ pPRb [ pPRa [ pPPa frac14 pPPb (ii) dPRa [ dPPa frac14dPPb [ dRRa frac14 dRRb [ dPRb
Interestingly the results of Proposition 3 are inde-pendent of the proportional fee rate a and hold regard-less of the parameter settings for competitionintensity (c) and maximum market size (h) Essen-tially market prices are highest in the pure resellermode since they are set by the intermediary whoattempts to moderate the competition between thesuppliers On the other hand when the suppliers setthe retail prices by operating in the online market-place mode they compete with each other directlyleading to lower retail prices Because of doublemarginalization in mode PR the retail price of theproduct that does not join the online marketplace ishigher than the price for which the supplier whochooses to do so Corresponding to these pricingstructures demand coverage is higher in the onlinemarketplace mode (PP) as compared to the resellermode (RR)
52 Intermediary ProfitsWe now proceed to examine how intermediary prof-itability across the three modes is affected by two keyparameters competition intensity (c) and order-fulfill-ment cost (F)8 To assess how these parameters affectprofitability of all supply chain members across thethree modes we start by making the followingassumptions First we assume that the order-fulfill-ment cost is small enough to ensure that both modePR and PP are feasible Analytically the potential val-ues of F under this assumption are that F lt FPP(ltFPR)Note that the value of F which satisfies the firstassumption also leads both aPR and aPP to be positiveOur second and final assumption is that the interme-diary chooses to set a = aPR for mode PR and a = aPP
for mode PP which represent equilibrium outcomesProposition 4 structurally characterizes which
mode choice would be preferred by the intermedi-ary as moderated by the competition intensityparameter c
PROPOSITION 4 There exist competition intensity thresh-olds c gt 0 and c(gtc) such that
bull If c lt c the optimal mode is PP
bull If c gt c the optimal mode is RR and
bull If c le c le c mode PR can be the optimal mode
Proposition 4 indicates that when competitionintensity is low the optimal mode is PP when compe-tition intensity is high the optimal mode is RR Inaddition we observe that when competition intensityis medium the optimal mode is PR The rationale
Table 1 Decision Matrix with Resulting Profits
Supplier A option
Supplier B option
Online marketplace Reseller
Online marketplace pPPA pPPB pPRA pPRB
Reseller pRPA pRPB pRRA pRRB
Note Due to symmetry we know that pPRA frac14 pRPB and pPRB frac14 pRPA
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1603
behind these results hinges on the moderating role ofthe intermediary Previous literature has shown thatthe intermediary prefers high levels of upstream com-petition as it strengthens its channel power (Wangand Shin 2015) From a mode perspective supplierscompete more fiercely in the online marketplacemode compared with the reseller mode as they set theprices directly and alternative modes allow the inter-mediary to moderate this competitive effect To elabo-rate when both suppliers are very competitive theintermediary would prefer to operate as a resellersince any other mode will lead to increased price com-petition which in turn negatively impacts the inter-mediary This result validates the conjecture of Hagiu(2007) that strong substitutability between suppliersrsquoproducts would lead to a greater preference for thepure reseller mode When the competition intensitybetween suppliers is low the intermediaryrsquos moderat-ing role is not as important Note that in mode PP theproportional fee could be considered a form of rev-enue sharing which would mitigate the impact ofdouble-marginalization (Geng et al 2018 Shy andWang 2011 Tan and Carrillo 2017) Hence the inter-mediary prefers mode PP when competition intensityis low When the competition intensity between sup-pliers is in the intermediate range the intermediaryrsquosmoderating role is important but not significantenough to alleviate double-marginalization effectsleading to a preference for the hybrid mode PRAnecdotal observations provide face validity for
our results For the household appliance market inChina the similarity between branded products leadsto strong substitutability (or high competition inten-sity) Hence most online retailers (eg JDcom) preferto act as resellers for products in this category On theother hand the media content industry (publishingvideo games) is characterized by low levels of pro-duct substitutability (or low competition intensity)This has motivated online retailers to set the terms forthe online marketplace mode in such a way that sup-pliers have chosen that mode of operation Etsycomthe largest e-commerce website in handmade or vin-tage items serves as a pure online marketplacebecause of low product substitutability FurtherAmazon serves as the reseller in certain high competi-tive product categories such as electronics but servesas a marketplace for other less competitive productcategories such as fine art (ie long-tail productcategory)Next we consider the impact of the order-fulfill-
ment cost on the choice of equilibrium mode for theintermediary Note that Proposition 4 shows thatwhen the market competition intensity is above a cer-tain threshold the intermediary will always prefer toact as a reseller To investigate the impact of theorder-fulfillment cost next we focus on the case that
the market competition intensity is not too high Theproposition below characterizes the mode preferredby the intermediary depending on the magnitude ofthe order-fulfillment cost
PROPOSITION 5 When the market competition intensityis relatively small (ie c cjFfrac140) there exists twothresholds F gt 0 and F(geF) such that
bull when F le F the preferred mode is PP and
bull when F lt F le F the preferred mode is PR and
bull when F gt F the preferred mode is RR
Proposition 5 shows that the magnitude of theorder-fulfillment cost will alter the mode choicesWhen the market competition intensity is relativelylow (c cjFfrac140) as the order-fulfillment costincreases the equilibrium mode will evolve from apure marketplace mode to a hybrid mode then to apure reseller mode The intuition of this result lies inthe fact that the higher the order-fulfillment costbecomes the more difficult it becomes for the inter-mediary to convince the suppliers to adopt the mar-ketplace mode and endure the order-fulfillment costRecall that both aPR and aPP decrease in F That is asthe order-fulfillment cost increases the intermediaryhas to set a much lower proportional fee to ensurethat the suppliers will accept the marketplace offerThis becomes increasingly expensive as the order-ful-fillment cost increases As a result the intermediaryprefers the marketplace mode for both supplierswhen the order-fulfillment cost is low and favors thereseller mode when the order-fulfillment cost issignificantFrom a practical perspective we observe that most
hotel and airline companies use the PP mode on tra-vel sites such as Expedia Travelocity Ctrip andQunar (Ctrip and Qunar are leading travel sites inChina) This is partially driven by the low (almostnegligible) order-fulfillment cost for electronic tick-etsbookings In the publishing industry a transitionto the PP mode from the RR mode is evident for e-books while the RR mode is still preferred for printbooks This could be due in part to the fact that order-fulfillment costs are greater for printed books as com-pared to e-booksHaving analytically examined the impacts of the
competition intensity and order-fulfillment cost indi-vidually we next numerically illustrate how the inter-actions between these two parameters affect the modechoice for the intermediary Figure 2 was generatedby setting h = 10 (ie the market potential equals to10) and pO = 0 (ie the outside option facing thesupplier is 0) Then we simultaneously varyF (F 2 eth0 FPPjcfrac140THORN where FPPjcfrac140 frac14 125) and c(c 2 (0 2)) Note that our results are robust to theparameter changes
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1604 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Proposition 1 characterizes the conditions underwhich Supplier A will accept the intermediaryrsquos mar-ketplace offer and Supplier B will not pursue the out-side option but continue to participate under thereseller mode The first condition relates to the order-fulfillment cost which should not exceed a certainthreshold Recall that by choosing the online market-place service the supplier gains pricing power butbears the order-fulfillment cost at the same timeThus this threshold value of the fulfillment costreflects the trade-off between gains from the pricingpower and the costs (in terms of order fulfillment) ofrealizing such gainsAssuming the first condition is satisfied the propo-
sition then specifies a second condition related to themaximum proportion a which can be charged to thesupplier by the intermediary Interesting insightsregarding this proportion a are shown in Lemma 3
LEMMA 3 (i) pPRI increases in a whereas both pPRA andpPRB decrease in a (ii) aPR decreases in F
Lemma 3(i) indicates that an increase in a willincrease the intermediaryrsquos profit pPRI and simultane-ously decrease Supplier Arsquos profit pPRA Hence ashould be small (ie a aPRA ) to ensure pPRA pRRA sothat Supplier A will accept the intermediaryrsquos market-place offer Meanwhile a counterintuitive result isthat an increase in a would also lead to a reduction inthe profit of the other supplier (B) who has not beenoffered the online marketplace service The reason forthis negative spillover effect is that when a isincreased the supplier (A) tends to increase the priceas the intermediary will keep a higher proportion ofthe revenue More specifically when the intermediaryraises the proportion a she can keep Supplier Aresponds by charging higher retail prices In responsethe intermediary sets a higher retail price for SupplierBrsquos product leading to lower demand for the sameproduct The net result is lower profit for Supplier BThus a needs to be small enough (a aPRB ) so thatSupplier B will not pursue the outside option but con-tinue to participate under the reseller mode orpPRB pO7 In equilibrium assuming the conditionrelated to the order-fulfillment cost is met (ieF le FPR) then the intermediary will set a frac14 aPR frac14minfrac12aPRA aPRB since this will allow the intermediary toextract the maximum profits from the supplier Fur-ther Lemma 3(ii) shows that this threshold value ofthe proportional fee (ie aPR) is decreasing in theorder-fulfillment cost (F)In summary from a managerial perspective our
results indicate that the intermediary requires astrategic focus in choosing the parameters for offeringthe online marketplace service This service will onlybe attractive to suppliers provided the order-
fulfillment cost is not significantly high If this is thecase then the intermediary should set the proportiona equal to its upper bound in order to realize the max-imum profits From a supplier perspective the onlinemarketplace offering is preferred over the reseller set-ting when the flexibility gains due to price setting out-weigh the explicit costs of order fulfillment
43 Mode PPIn this setting both suppliers will be offered theoption of online marketplace from the intermediaryIf both suppliers choose the online marketplace ser-vice they will take into consideration the proportiona which they need to share with the intermediary Fora given proportion a the profits for each supplier andthe intermediary under this mode (ie both supplierschoose the online marketplace service) are as follows
pAethPPTHORN frac14 eth1 aTHORNpafrac12h pa thorn cethpb paTHORN F
pBethPPTHORN frac14 eth1 aTHORNpbfrac12h pb thorn cethpa pbTHORN F
pIethPPTHORN frac14 apafrac12h pa thorn cethpb paTHORNthorn apbfrac12h pb thorn cethpa pbTHORN
We solve this game by simultaneously determiningthe retail prices pa and pb which maximize pA(PP)and pB(PP) The results are shown in the lemmabelow
LEMMA 4 Assuming both suppliers accept the propor-tion a there exists a unique equilibrium with retail pricesand demands are as follows
pPPa frac14 pPPb frac14 1
2thorn ch
dPPa frac14 dPPb frac14 1thorn c2thorn c
h
The corresponding equilibrium profits are given by
pPPA frac14 pPPB frac14 eth1 aTHORN 1thorn c
eth2thorn cTHORN2 h2 F
pPPI frac14 2a1thorn c
eth2thorn cTHORN2 h2
When is this mode a possible outcome This can beframed using the decision matrix in Table 1 where weare interested in the equilibrium when both supplierswill accept the online marketplace offer This equilib-rium will arise when both of the following conditionshold Given that Supplier B (A) adopts the resellermodel Supplier Arsquos (Brsquos) profit should be higherunder the condition of acceptance than refusal of theintermediaryrsquos marketplace offer which is pPRA pRRA(pRPB pRRB ) given that Supplier B (A) adopts theonline marketplace Supplier Arsquos (Brsquos) profit should
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1602 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
also be higher when it accepts rather than refuses theintermediaryrsquos marketplace offer which is pPPA pRPA(pPPB pPRB ) In addition to ensure this equilibrium isstable both suppliersrsquo profits under the PP modeshould not be lower than the profit under the outsideoption which is pPPA pO (pPPB pO) Based on theresults in Lemma 4 the following proposition identi-fies the conditions under which both suppliers willnot pursue the outside option but choose the onlinemarketplace service offering
PROPOSITION 2 Both suppliers will choose the onlinemarketplace service provided pPRA pRRA pPPA pRPA andpPPA pO This leads to the following sufficient condi-tions for the existence of the PP mode
bull F le FPP where FPP frac14 1thorn ceth2thorn cTHORN2 h
2 eth1thorn cTHORNeth2thorn 3cTHORN2eth2thorn cTHORNeth4thorn 8cthorn 2c2THORN
h2 and
bull a aPP frac14 minfaPRA aPPA aPPB g where aPPA is the
unique solution of eth1 aPPA THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14frac12eth1thorn cTHORNeth2thorn 3cTHORN aPPA eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPPATHORNc2eth4thorn 8cthorn 2c2THORN h
2 and aPPB is the
unique solution of eth1 aPPB THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14 pO
Similar to the hybrid mode both suppliers willaccept the intermediaryrsquos online marketplace offerwhen the order-fulfillment cost and the proportionalfee that the intermediary charges are all below a cer-tain threshold It is worth noting that the threshold inorder-fulfillment cost is more stringent than that statedin Proposition 1 since FPP lt FPR As before the inter-mediaryrsquos profit pPPI increases in a while both pPPA andpPPB decrease in a In equilibrium the intermediary isonce again incentivized to set a = aPP which decreasesin the order-fulfillment cost (F) We also note that sup-pliersrsquo profits will decrease as the upstream competi-tion intensifies (ie a higher value of c)In the next section we compare the equilibrium
solutions for the three modes with a view to provid-ing managerial insights
5 Insights
51 Pricing and DemandFrom a pricing and demand perspective the proposi-tion below provides a comparison across the threechannel modes
PROPOSITION 3 Comparing the equilibrium results ofretail prices and demand we have (i) pRRa frac14pRRb [ pPRb [ pPRa [ pPPa frac14 pPPb (ii) dPRa [ dPPa frac14dPPb [ dRRa frac14 dRRb [ dPRb
Interestingly the results of Proposition 3 are inde-pendent of the proportional fee rate a and hold regard-less of the parameter settings for competitionintensity (c) and maximum market size (h) Essen-tially market prices are highest in the pure resellermode since they are set by the intermediary whoattempts to moderate the competition between thesuppliers On the other hand when the suppliers setthe retail prices by operating in the online market-place mode they compete with each other directlyleading to lower retail prices Because of doublemarginalization in mode PR the retail price of theproduct that does not join the online marketplace ishigher than the price for which the supplier whochooses to do so Corresponding to these pricingstructures demand coverage is higher in the onlinemarketplace mode (PP) as compared to the resellermode (RR)
52 Intermediary ProfitsWe now proceed to examine how intermediary prof-itability across the three modes is affected by two keyparameters competition intensity (c) and order-fulfill-ment cost (F)8 To assess how these parameters affectprofitability of all supply chain members across thethree modes we start by making the followingassumptions First we assume that the order-fulfill-ment cost is small enough to ensure that both modePR and PP are feasible Analytically the potential val-ues of F under this assumption are that F lt FPP(ltFPR)Note that the value of F which satisfies the firstassumption also leads both aPR and aPP to be positiveOur second and final assumption is that the interme-diary chooses to set a = aPR for mode PR and a = aPP
for mode PP which represent equilibrium outcomesProposition 4 structurally characterizes which
mode choice would be preferred by the intermedi-ary as moderated by the competition intensityparameter c
PROPOSITION 4 There exist competition intensity thresh-olds c gt 0 and c(gtc) such that
bull If c lt c the optimal mode is PP
bull If c gt c the optimal mode is RR and
bull If c le c le c mode PR can be the optimal mode
Proposition 4 indicates that when competitionintensity is low the optimal mode is PP when compe-tition intensity is high the optimal mode is RR Inaddition we observe that when competition intensityis medium the optimal mode is PR The rationale
Table 1 Decision Matrix with Resulting Profits
Supplier A option
Supplier B option
Online marketplace Reseller
Online marketplace pPPA pPPB pPRA pPRB
Reseller pRPA pRPB pRRA pRRB
Note Due to symmetry we know that pPRA frac14 pRPB and pPRB frac14 pRPA
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1603
behind these results hinges on the moderating role ofthe intermediary Previous literature has shown thatthe intermediary prefers high levels of upstream com-petition as it strengthens its channel power (Wangand Shin 2015) From a mode perspective supplierscompete more fiercely in the online marketplacemode compared with the reseller mode as they set theprices directly and alternative modes allow the inter-mediary to moderate this competitive effect To elabo-rate when both suppliers are very competitive theintermediary would prefer to operate as a resellersince any other mode will lead to increased price com-petition which in turn negatively impacts the inter-mediary This result validates the conjecture of Hagiu(2007) that strong substitutability between suppliersrsquoproducts would lead to a greater preference for thepure reseller mode When the competition intensitybetween suppliers is low the intermediaryrsquos moderat-ing role is not as important Note that in mode PP theproportional fee could be considered a form of rev-enue sharing which would mitigate the impact ofdouble-marginalization (Geng et al 2018 Shy andWang 2011 Tan and Carrillo 2017) Hence the inter-mediary prefers mode PP when competition intensityis low When the competition intensity between sup-pliers is in the intermediate range the intermediaryrsquosmoderating role is important but not significantenough to alleviate double-marginalization effectsleading to a preference for the hybrid mode PRAnecdotal observations provide face validity for
our results For the household appliance market inChina the similarity between branded products leadsto strong substitutability (or high competition inten-sity) Hence most online retailers (eg JDcom) preferto act as resellers for products in this category On theother hand the media content industry (publishingvideo games) is characterized by low levels of pro-duct substitutability (or low competition intensity)This has motivated online retailers to set the terms forthe online marketplace mode in such a way that sup-pliers have chosen that mode of operation Etsycomthe largest e-commerce website in handmade or vin-tage items serves as a pure online marketplacebecause of low product substitutability FurtherAmazon serves as the reseller in certain high competi-tive product categories such as electronics but servesas a marketplace for other less competitive productcategories such as fine art (ie long-tail productcategory)Next we consider the impact of the order-fulfill-
ment cost on the choice of equilibrium mode for theintermediary Note that Proposition 4 shows thatwhen the market competition intensity is above a cer-tain threshold the intermediary will always prefer toact as a reseller To investigate the impact of theorder-fulfillment cost next we focus on the case that
the market competition intensity is not too high Theproposition below characterizes the mode preferredby the intermediary depending on the magnitude ofthe order-fulfillment cost
PROPOSITION 5 When the market competition intensityis relatively small (ie c cjFfrac140) there exists twothresholds F gt 0 and F(geF) such that
bull when F le F the preferred mode is PP and
bull when F lt F le F the preferred mode is PR and
bull when F gt F the preferred mode is RR
Proposition 5 shows that the magnitude of theorder-fulfillment cost will alter the mode choicesWhen the market competition intensity is relativelylow (c cjFfrac140) as the order-fulfillment costincreases the equilibrium mode will evolve from apure marketplace mode to a hybrid mode then to apure reseller mode The intuition of this result lies inthe fact that the higher the order-fulfillment costbecomes the more difficult it becomes for the inter-mediary to convince the suppliers to adopt the mar-ketplace mode and endure the order-fulfillment costRecall that both aPR and aPP decrease in F That is asthe order-fulfillment cost increases the intermediaryhas to set a much lower proportional fee to ensurethat the suppliers will accept the marketplace offerThis becomes increasingly expensive as the order-ful-fillment cost increases As a result the intermediaryprefers the marketplace mode for both supplierswhen the order-fulfillment cost is low and favors thereseller mode when the order-fulfillment cost issignificantFrom a practical perspective we observe that most
hotel and airline companies use the PP mode on tra-vel sites such as Expedia Travelocity Ctrip andQunar (Ctrip and Qunar are leading travel sites inChina) This is partially driven by the low (almostnegligible) order-fulfillment cost for electronic tick-etsbookings In the publishing industry a transitionto the PP mode from the RR mode is evident for e-books while the RR mode is still preferred for printbooks This could be due in part to the fact that order-fulfillment costs are greater for printed books as com-pared to e-booksHaving analytically examined the impacts of the
competition intensity and order-fulfillment cost indi-vidually we next numerically illustrate how the inter-actions between these two parameters affect the modechoice for the intermediary Figure 2 was generatedby setting h = 10 (ie the market potential equals to10) and pO = 0 (ie the outside option facing thesupplier is 0) Then we simultaneously varyF (F 2 eth0 FPPjcfrac140THORN where FPPjcfrac140 frac14 125) and c(c 2 (0 2)) Note that our results are robust to theparameter changes
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1604 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
also be higher when it accepts rather than refuses theintermediaryrsquos marketplace offer which is pPPA pRPA(pPPB pPRB ) In addition to ensure this equilibrium isstable both suppliersrsquo profits under the PP modeshould not be lower than the profit under the outsideoption which is pPPA pO (pPPB pO) Based on theresults in Lemma 4 the following proposition identi-fies the conditions under which both suppliers willnot pursue the outside option but choose the onlinemarketplace service offering
PROPOSITION 2 Both suppliers will choose the onlinemarketplace service provided pPRA pRRA pPPA pRPA andpPPA pO This leads to the following sufficient condi-tions for the existence of the PP mode
bull F le FPP where FPP frac14 1thorn ceth2thorn cTHORN2 h
2 eth1thorn cTHORNeth2thorn 3cTHORN2eth2thorn cTHORNeth4thorn 8cthorn 2c2THORN
h2 and
bull a aPP frac14 minfaPRA aPPA aPPB g where aPPA is the
unique solution of eth1 aPPA THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14frac12eth1thorn cTHORNeth2thorn 3cTHORN aPPA eth1thorn 2cTHORNc2
2eth1thorn cTHORNfrac124eth1thorn cTHORN2 eth1thorn aPPATHORNc2eth4thorn 8cthorn 2c2THORN h
2 and aPPB is the
unique solution of eth1 aPPB THORN 1thorn ceth2thorn cTHORN2 h
2 F frac14 pO
Similar to the hybrid mode both suppliers willaccept the intermediaryrsquos online marketplace offerwhen the order-fulfillment cost and the proportionalfee that the intermediary charges are all below a cer-tain threshold It is worth noting that the threshold inorder-fulfillment cost is more stringent than that statedin Proposition 1 since FPP lt FPR As before the inter-mediaryrsquos profit pPPI increases in a while both pPPA andpPPB decrease in a In equilibrium the intermediary isonce again incentivized to set a = aPP which decreasesin the order-fulfillment cost (F) We also note that sup-pliersrsquo profits will decrease as the upstream competi-tion intensifies (ie a higher value of c)In the next section we compare the equilibrium
solutions for the three modes with a view to provid-ing managerial insights
5 Insights
51 Pricing and DemandFrom a pricing and demand perspective the proposi-tion below provides a comparison across the threechannel modes
PROPOSITION 3 Comparing the equilibrium results ofretail prices and demand we have (i) pRRa frac14pRRb [ pPRb [ pPRa [ pPPa frac14 pPPb (ii) dPRa [ dPPa frac14dPPb [ dRRa frac14 dRRb [ dPRb
Interestingly the results of Proposition 3 are inde-pendent of the proportional fee rate a and hold regard-less of the parameter settings for competitionintensity (c) and maximum market size (h) Essen-tially market prices are highest in the pure resellermode since they are set by the intermediary whoattempts to moderate the competition between thesuppliers On the other hand when the suppliers setthe retail prices by operating in the online market-place mode they compete with each other directlyleading to lower retail prices Because of doublemarginalization in mode PR the retail price of theproduct that does not join the online marketplace ishigher than the price for which the supplier whochooses to do so Corresponding to these pricingstructures demand coverage is higher in the onlinemarketplace mode (PP) as compared to the resellermode (RR)
52 Intermediary ProfitsWe now proceed to examine how intermediary prof-itability across the three modes is affected by two keyparameters competition intensity (c) and order-fulfill-ment cost (F)8 To assess how these parameters affectprofitability of all supply chain members across thethree modes we start by making the followingassumptions First we assume that the order-fulfill-ment cost is small enough to ensure that both modePR and PP are feasible Analytically the potential val-ues of F under this assumption are that F lt FPP(ltFPR)Note that the value of F which satisfies the firstassumption also leads both aPR and aPP to be positiveOur second and final assumption is that the interme-diary chooses to set a = aPR for mode PR and a = aPP
for mode PP which represent equilibrium outcomesProposition 4 structurally characterizes which
mode choice would be preferred by the intermedi-ary as moderated by the competition intensityparameter c
PROPOSITION 4 There exist competition intensity thresh-olds c gt 0 and c(gtc) such that
bull If c lt c the optimal mode is PP
bull If c gt c the optimal mode is RR and
bull If c le c le c mode PR can be the optimal mode
Proposition 4 indicates that when competitionintensity is low the optimal mode is PP when compe-tition intensity is high the optimal mode is RR Inaddition we observe that when competition intensityis medium the optimal mode is PR The rationale
Table 1 Decision Matrix with Resulting Profits
Supplier A option
Supplier B option
Online marketplace Reseller
Online marketplace pPPA pPPB pPRA pPRB
Reseller pRPA pRPB pRRA pRRB
Note Due to symmetry we know that pPRA frac14 pRPB and pPRB frac14 pRPA
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1603
behind these results hinges on the moderating role ofthe intermediary Previous literature has shown thatthe intermediary prefers high levels of upstream com-petition as it strengthens its channel power (Wangand Shin 2015) From a mode perspective supplierscompete more fiercely in the online marketplacemode compared with the reseller mode as they set theprices directly and alternative modes allow the inter-mediary to moderate this competitive effect To elabo-rate when both suppliers are very competitive theintermediary would prefer to operate as a resellersince any other mode will lead to increased price com-petition which in turn negatively impacts the inter-mediary This result validates the conjecture of Hagiu(2007) that strong substitutability between suppliersrsquoproducts would lead to a greater preference for thepure reseller mode When the competition intensitybetween suppliers is low the intermediaryrsquos moderat-ing role is not as important Note that in mode PP theproportional fee could be considered a form of rev-enue sharing which would mitigate the impact ofdouble-marginalization (Geng et al 2018 Shy andWang 2011 Tan and Carrillo 2017) Hence the inter-mediary prefers mode PP when competition intensityis low When the competition intensity between sup-pliers is in the intermediate range the intermediaryrsquosmoderating role is important but not significantenough to alleviate double-marginalization effectsleading to a preference for the hybrid mode PRAnecdotal observations provide face validity for
our results For the household appliance market inChina the similarity between branded products leadsto strong substitutability (or high competition inten-sity) Hence most online retailers (eg JDcom) preferto act as resellers for products in this category On theother hand the media content industry (publishingvideo games) is characterized by low levels of pro-duct substitutability (or low competition intensity)This has motivated online retailers to set the terms forthe online marketplace mode in such a way that sup-pliers have chosen that mode of operation Etsycomthe largest e-commerce website in handmade or vin-tage items serves as a pure online marketplacebecause of low product substitutability FurtherAmazon serves as the reseller in certain high competi-tive product categories such as electronics but servesas a marketplace for other less competitive productcategories such as fine art (ie long-tail productcategory)Next we consider the impact of the order-fulfill-
ment cost on the choice of equilibrium mode for theintermediary Note that Proposition 4 shows thatwhen the market competition intensity is above a cer-tain threshold the intermediary will always prefer toact as a reseller To investigate the impact of theorder-fulfillment cost next we focus on the case that
the market competition intensity is not too high Theproposition below characterizes the mode preferredby the intermediary depending on the magnitude ofthe order-fulfillment cost
PROPOSITION 5 When the market competition intensityis relatively small (ie c cjFfrac140) there exists twothresholds F gt 0 and F(geF) such that
bull when F le F the preferred mode is PP and
bull when F lt F le F the preferred mode is PR and
bull when F gt F the preferred mode is RR
Proposition 5 shows that the magnitude of theorder-fulfillment cost will alter the mode choicesWhen the market competition intensity is relativelylow (c cjFfrac140) as the order-fulfillment costincreases the equilibrium mode will evolve from apure marketplace mode to a hybrid mode then to apure reseller mode The intuition of this result lies inthe fact that the higher the order-fulfillment costbecomes the more difficult it becomes for the inter-mediary to convince the suppliers to adopt the mar-ketplace mode and endure the order-fulfillment costRecall that both aPR and aPP decrease in F That is asthe order-fulfillment cost increases the intermediaryhas to set a much lower proportional fee to ensurethat the suppliers will accept the marketplace offerThis becomes increasingly expensive as the order-ful-fillment cost increases As a result the intermediaryprefers the marketplace mode for both supplierswhen the order-fulfillment cost is low and favors thereseller mode when the order-fulfillment cost issignificantFrom a practical perspective we observe that most
hotel and airline companies use the PP mode on tra-vel sites such as Expedia Travelocity Ctrip andQunar (Ctrip and Qunar are leading travel sites inChina) This is partially driven by the low (almostnegligible) order-fulfillment cost for electronic tick-etsbookings In the publishing industry a transitionto the PP mode from the RR mode is evident for e-books while the RR mode is still preferred for printbooks This could be due in part to the fact that order-fulfillment costs are greater for printed books as com-pared to e-booksHaving analytically examined the impacts of the
competition intensity and order-fulfillment cost indi-vidually we next numerically illustrate how the inter-actions between these two parameters affect the modechoice for the intermediary Figure 2 was generatedby setting h = 10 (ie the market potential equals to10) and pO = 0 (ie the outside option facing thesupplier is 0) Then we simultaneously varyF (F 2 eth0 FPPjcfrac140THORN where FPPjcfrac140 frac14 125) and c(c 2 (0 2)) Note that our results are robust to theparameter changes
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1604 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
behind these results hinges on the moderating role ofthe intermediary Previous literature has shown thatthe intermediary prefers high levels of upstream com-petition as it strengthens its channel power (Wangand Shin 2015) From a mode perspective supplierscompete more fiercely in the online marketplacemode compared with the reseller mode as they set theprices directly and alternative modes allow the inter-mediary to moderate this competitive effect To elabo-rate when both suppliers are very competitive theintermediary would prefer to operate as a resellersince any other mode will lead to increased price com-petition which in turn negatively impacts the inter-mediary This result validates the conjecture of Hagiu(2007) that strong substitutability between suppliersrsquoproducts would lead to a greater preference for thepure reseller mode When the competition intensitybetween suppliers is low the intermediaryrsquos moderat-ing role is not as important Note that in mode PP theproportional fee could be considered a form of rev-enue sharing which would mitigate the impact ofdouble-marginalization (Geng et al 2018 Shy andWang 2011 Tan and Carrillo 2017) Hence the inter-mediary prefers mode PP when competition intensityis low When the competition intensity between sup-pliers is in the intermediate range the intermediaryrsquosmoderating role is important but not significantenough to alleviate double-marginalization effectsleading to a preference for the hybrid mode PRAnecdotal observations provide face validity for
our results For the household appliance market inChina the similarity between branded products leadsto strong substitutability (or high competition inten-sity) Hence most online retailers (eg JDcom) preferto act as resellers for products in this category On theother hand the media content industry (publishingvideo games) is characterized by low levels of pro-duct substitutability (or low competition intensity)This has motivated online retailers to set the terms forthe online marketplace mode in such a way that sup-pliers have chosen that mode of operation Etsycomthe largest e-commerce website in handmade or vin-tage items serves as a pure online marketplacebecause of low product substitutability FurtherAmazon serves as the reseller in certain high competi-tive product categories such as electronics but servesas a marketplace for other less competitive productcategories such as fine art (ie long-tail productcategory)Next we consider the impact of the order-fulfill-
ment cost on the choice of equilibrium mode for theintermediary Note that Proposition 4 shows thatwhen the market competition intensity is above a cer-tain threshold the intermediary will always prefer toact as a reseller To investigate the impact of theorder-fulfillment cost next we focus on the case that
the market competition intensity is not too high Theproposition below characterizes the mode preferredby the intermediary depending on the magnitude ofthe order-fulfillment cost
PROPOSITION 5 When the market competition intensityis relatively small (ie c cjFfrac140) there exists twothresholds F gt 0 and F(geF) such that
bull when F le F the preferred mode is PP and
bull when F lt F le F the preferred mode is PR and
bull when F gt F the preferred mode is RR
Proposition 5 shows that the magnitude of theorder-fulfillment cost will alter the mode choicesWhen the market competition intensity is relativelylow (c cjFfrac140) as the order-fulfillment costincreases the equilibrium mode will evolve from apure marketplace mode to a hybrid mode then to apure reseller mode The intuition of this result lies inthe fact that the higher the order-fulfillment costbecomes the more difficult it becomes for the inter-mediary to convince the suppliers to adopt the mar-ketplace mode and endure the order-fulfillment costRecall that both aPR and aPP decrease in F That is asthe order-fulfillment cost increases the intermediaryhas to set a much lower proportional fee to ensurethat the suppliers will accept the marketplace offerThis becomes increasingly expensive as the order-ful-fillment cost increases As a result the intermediaryprefers the marketplace mode for both supplierswhen the order-fulfillment cost is low and favors thereseller mode when the order-fulfillment cost issignificantFrom a practical perspective we observe that most
hotel and airline companies use the PP mode on tra-vel sites such as Expedia Travelocity Ctrip andQunar (Ctrip and Qunar are leading travel sites inChina) This is partially driven by the low (almostnegligible) order-fulfillment cost for electronic tick-etsbookings In the publishing industry a transitionto the PP mode from the RR mode is evident for e-books while the RR mode is still preferred for printbooks This could be due in part to the fact that order-fulfillment costs are greater for printed books as com-pared to e-booksHaving analytically examined the impacts of the
competition intensity and order-fulfillment cost indi-vidually we next numerically illustrate how the inter-actions between these two parameters affect the modechoice for the intermediary Figure 2 was generatedby setting h = 10 (ie the market potential equals to10) and pO = 0 (ie the outside option facing thesupplier is 0) Then we simultaneously varyF (F 2 eth0 FPPjcfrac140THORN where FPPjcfrac140 frac14 125) and c(c 2 (0 2)) Note that our results are robust to theparameter changes
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1604 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
From Figure 2 we first observe that for any givenorder-fulfillment cost increasing levels of competi-tion intensity (ie c) results in a transition frommode PP to mode PR and then to mode RR as inter-mediaryrsquos preferred mode choice Second when themarket competition intensity is relatively low (iec cjFfrac140) as the order-fulfillment cost increases theequilibrium mode will evolve from mode PP tomode PR and then mode RR Third when the mar-ket competition intensity is in a medium range (iecjFfrac140 c cjFfrac140) the preferred transition isbetween modes PR and RR as the order-fulfillmentcost increases In addition when the market competi-tion intensity is relatively high (ie c cjFfrac140) thenthe reseller mode RR will dominate as the preferredmode choice The intuition behind these resultshinges on the trade-off between the transfer of con-trol rights for product pricing and the responsibilityfor order fulfillment The results here provide exe-cutable managerial insights to the online platformsSpecifically if competition intensity is low and theorder-fulfillment cost is not too high then the inter-mediary should specify a relatively low proportionalfee a to induce suppliers to switch to mode PP Asboth competition intensity and order-fulfillmentcosts increase the intermediary should increase theproportional fee a to motivate some supplier switch-ing Finally when both competition intensity andorder-fulfillment cost are high then the proportionalfee a should be set to a large value to discourage thesuppliers from switching to the online marketplacemode
53 Supplier ProfitabilityIt is straightforward to show that if the intermediarysets the referral proportion a such that a 1
2 then forboth suppliers mode RR always optimizes their ownprofitability It is only when a 1
2 that supplierchoices are moderated by the order-fulfillment cost F
the referral proportion a and the level of competitionintensity cMore specifically assume that the supplier terms
are such that F le FPP and a = aPP so that mode PP isnow a feasible alternative In this case the choice ofboth suppliers to move to the online marketplacemode PP from the reseller mode RR is dependentupon the level of competition intensity such that (a)under low levels of competition intensity both sup-pliersrsquo profits are larger under mode RR and (b)under high levels of competition intensity the reverseis true These results indicate that regardless of theterms offered by the intermediary through the referralproportion a the exogenous market parameter c willaffect whether suppliers are better or worse off undermode PP or mode RR Hence from a managerial per-spective it is relevant to not only evaluate the inter-mediaryrsquos terms for the online marketplace mode butalso to examine the intensity of competition within aproduct categoryAssume that F le FPR and a = aPR so that mode PR
is a feasible alternative The supplier choosing theonline marketplace service would realize higher prof-its under mode PR as compared to mode RR whilewe observe that the supplier operating under the resel-ler mode (ie the supplier who is not offered theoption of online marketplace by the intermediary)would realize lower profits in mode PR as comparedto mode RR Without the alleviation of the double-marginalization effect Supplier A (who chooses theonline marketplace mode) will decrease the retailprice of its product to attract consumers which leadsto a reduction in the total demand for the other sup-plier (ie B) Although Supplier B attempts to allevi-ate this demand reduction with a relative pricedecrease to enhance profitability the net impact doesnot result in a larger profit for this supplier as com-pared to when both suppliers operate as resellers (iemode RR) Managerially this reflects a ldquofirst moverrdquoadvantage for suppliers As soon as the intermediaryoffers a reduction in the referral proportion a and setsa = aPR the first supplier to choose to switch to theonline marketplace service would be the one whoreaps the gains from this mode of operation Note thatthe competition intensity parameter c does not mod-erate this finding
54 Supply Chain ProfitsOur next focus is on total supply chain profits (ie thesum of both the suppliersrsquo profits and the intermedi-aryrsquos profit) across the three modes The fulfillmentcost parameter is irrelevant in terms of this compar-ison and hence we set F = 0 Figure 3 illustrates thetotal supply chain profits as a function of the competi-tion intensity parameter c (the maximum market sizeh = 10 and the outside option profit pO = 0) When
0 1 2 3 4 5 6 7 8 9 10 110
02
04
06
08
1
12
14
16
18
2
PP
PR
RR
F
γFigure 2 Preferred Mode Interaction of F and c (h = 10 pO = 0)
[Color figure can be viewed at wileyonlinelibrarycom]
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1605
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
competition intensity is low total supply chain profitsare at a maximum under mode PP at the otherextreme when competition intensity is high totalsupply chain profits are at a maximum under modeRR It is only when competition intensity is the mid-dle range that mode PR leads to maximum supplychain profits The driver of these results is the trade-off between the alleviation of the double-marginaliza-tion effect and price competition
55 Anecdotal EvidenceSince fulfillment costs (F) are related to inventorystorage and transportation it is reasonable to assumethat ldquolargerrdquo products (in terms of weight andorsize) would lead to higher fulfillment costs Given ourresults that higher fulfillment costs would lead to astronger preference for the reseller mode (RR) weprovide limited anecdotal evidence that these resultshold in practiceJDcom is one of the largest B2C online retailers in
the world with annual sales around $23 billion Wefocus our attention on a few product categories (egcleaning products household appliances food anddigital products) sold on JDcom during March 2015For each product category the percentage of suppli-ers operating in a reseller mode (RR) is shown inTable 2 This Table shows that for each individualproduct as product sizeweight increases the per-centage of the reseller mode being adopted increasesGiven that fulfillment costs would be higher for prod-ucts with higher weight andor larger size this dataprovides validity to our results on the impact oforder-fulfillment costIt is also interesting to note that there are significant
differences in the percentage of suppliers who choosethe reseller mode (RR) within and across product cat-egories For example comparing the foodoil cate-gory to the household appliance category we see that
a significantly larger percentage of suppliers choosethe reseller mode in the latter category as comparedto the former One explanation for this is that con-sumers might consider products within the householdappliances category as being more homogeneous com-pared to products within the foodoil category Sincethis would imply higher competition intensity forproducts in the household appliances category as com-pared to products in the foodoil category it is prefer-able for suppliers to choose the reseller mode in thelatter category and the online marketplace mode in theformer categoryFinally there are also differences for individual
products within a product category As product sizeandor weight increases for both student tablets andlaptop computers a greater percentage of supplierschoose to operate under the reseller mode These per-centages are significantly different across the individ-ual products As we explored earlier these differencescould stem from differences in product differentiation(or competition intensity) for individual productswithin a product category that is lower fulfillmentcosts coupled with lower product differentiation(or higher competition intensity) leads to a higher per-centage of suppliers choosing the reseller mode for stu-dent tablets as compared to lower fulfillment costs
0 05 1 15 236
38
40
42
44
46
48
50
The competition intensity γ
The
sup
ply
chai
n pr
ofit
RRPRPP
Figure 3 The Supply Chain Profit under Different Modes (h = 10pO = 0 F = 0) [Color figure can be viewed at wileyonlinelibrarycom]
Table 2 The Percentage of the Reseller Mode Used on JDCOM
Household applianceElectric kettle 11 L and below 12ndash15 15 abovendash 177 426 481Multi-role pot 2ndash4 L 4ndash6 6 abovendash 135 226 471Water heater 60 L and below 60ndash79 80 and
abovendash 104 250 313Egg boilers 5 below 5ndash10 10 abovendash 190 284 353Yogurt maker 500 ml 800ndash1000 1200 and
abovendash 118 145 563
Cleaning productsCloth cleaning 0ndash1 kg 1ndash3 3 abovendash 292 324 412Home cleaning 500 ml below 500ndash1000 1000 abovendash 35 43 69Skin care 100 ml below 100ndash199 200ndash399 400 and
abovendash 44 43 67 78
FoodRice grains 1 kg and below 1ndash5 5 abovendash 58 99 127
Digital productsStudent tablet 8 inch below 8ndash10 10 abovendash 136 418 448Laptopcomputer
13 inch below 13ndash15 15 above
ndash 27 105 128
Source Data collected from JDcom during March 2015
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1606 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
coupled with higher product differentiation (or lowercompetition intensity) for laptop computers
56 Variable vs Fixed Order-Fulfillment CostSince customers are often responsible for deliverycosts in online shopping we normalize the variablecost of the order fulfillment to zero and focus on thefixed cost in our base model It could be argued thatvariable order-fulfillment costs are proportional tomarket demand and hence instead of fixed fulfill-ment costs we should incorporate them into our anal-ysis When we do so the results do not lead to anychanges in our key insights The primary reason forthis is that variable order-fulfillment costs not onlyinfluence pricing decisions but also affect propor-tional fees in the equilibrium Thus they affect theresults in a similar qualitative manner as fixed order-fulfillment costs In particular a higher (lower) vari-able order-fulfillment cost results in higher (lower)market prices With a higher market price the sup-plier is less likely to accept the intermediaryrsquos market-place offer with the same proportional fee ratemaking the pure reseller mode more likely to emergeas an equilibrium choice Since the focus of our studyis on the strategic interactions between the onlineintermediary and the suppliers in mode choice ratherthan on pricing strategy a positive variable order-ful-fillment cost only leads to more analytical complexitywithout yielding additional insights into our researchquestions This is the rationale for choosing to incor-porate fixed rather than variable fulfillment costs inour model
57 Asymmetric Order-Fulfillment Costs orMarket PotentialsOur analysis so far has assumed symmetric suppliersin terms of order fulfillments costs and market poten-tials For the case of asymmetries in these parameterswe find that our key results and qualitative insightsstill hold There are some interesting new insights thatemerge from this investigation which are both relatedto the hybrid mode PR (details are provided in theOnline Appendix)First the intermediary should offer the online mar-
ketplace option to the supplier with the lower relativeorder-fulfillment cost This result is driven by the factthat the sufficient condition stated in Proposition 1 forthe order-fulfillment cost is easier to achieve when theonline marketplace offer is made to the supplier withthe lower cost Second if the two suppliers have dif-ferent market potentials it is always optimal for theintermediary to provide and convince the supplierwith higher market potential to accept the online mar-ketplace offer The proportional fee is a form ofrevenue sharing and hence can mitigate double-marginalization effects (Geng et al 2018 Tan and
Carrillo 2017) A higher market potential induces alarger loss due to double-marginalization and thusby offering the online marketplace option to the sup-plier with a higher market potential this loss can bereduced This finding is consistent with industrypractice For example Walmart Marketplace is invita-tion-only They claim that ldquoWe are looking for rela-tionships with reputable retailers and brands thatprovide first-class customer service a compelling pro-duct assortment competitive pricing and fast reliablefulfillmentrdquo9
6 Conclusions and Discussions
A relatively new mode of operation for online interme-diaries is the ldquoonline marketplacerdquo Under this modethe online intermediary offers an option to suppliersThe intermediary can act as a reseller of the suppliersrsquoproducts or let the suppliers to operate as indepen-dent suppliers on the intermediary platform that isthe intermediary simply acts as an ldquoonline market-placerdquo In this study we have provided insights intothe key aspects that moderate the decision of an inter-mediary to operate as a reseller operate as an onlinemarketplace or adopt a hybrid mode of operationPrevious research has indicated that the online mar-
ketplace is always the preferred option for supplychain members since it reduces the double marginal-ization impact and thus increases profits for all mem-bers We show that this is not necessarily the caseInstead we note that the degree of competition inten-sity and magnitude of order-fulfillment cost moderatethe preference regions within which distinct modesare preferred by the online intermediary For givenfulfillment costs (a) high levels of competition inten-sity lead to the intermediary preferring the resellermode (b) moderate levels of competition intensitylead to a preference for the hybrid mode and (c) atlow levels of competition intensity the preferredmode is the online marketplace The key intuitionbehind this finding is the moderating role of the inter-mediary Similarly for a given level of competitionintensity we show that (a) when order-fulfillmentcosts are low the marketplace mode is the preferredchoice (b) moderate levels of order-fulfillment costslead to the choice of the hybrid mode and (c) thereseller mode is preferred when order-fulfillmentcosts are highWe also find that a non-trivial interaction effect
between the degree of competition intensity and themagnitude of fulfillment costs affects optimal modechoice for the intermediary For example the thresh-old order-fulfillment costs for the intermediary toswitch from a marketplace to a hybrid mode or froma hybrid to a reseller mode decreases as competitionintensity increases Similarly the threshold level of
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1607
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
competition intensity for the intermediary to switchfrom a marketplace to a hybrid mode or from a hybridto a reseller mode decreases as fulfillment costsincrease Through data collected from one onlineintermediary we also provide some face validity forour findings Finally these key insights hold regard-less of whether the order fulfillments costs are fixed(as our analysis assumes) or are variable and demanddependentFrom a practical perspective online retailers should
consider providing the online marketplace option forproduct categories in which the competition amongdifferent suppliers is not intense while adopting thetraditional reseller mode for highly competitive prod-ucts An alternative is to discriminate between sellersacross different product categories by using distinctrevenue sharing proportions For highly competitiveproduct categories the online intermediary candemand a higher proportional fee while charging alower fee for the sellers in the long-tail product cate-gories Also if the related order-fulfillment cost isvery high even when the competition is not intensethe online intermediary can still leverage the resellermode to improve profitabilityWe conclude the paper by pointing out a few
caveats about our model and some directions forfuture research First we have assumed that the sup-plierrsquos profit stemming from the outside option is notvery high If a supplierrsquos profit under the outsideoption is very high the supplier will choose to leavethe intermediary and pursue the outside option Con-sequently the original operating modes will nolonger exist and we will have a monopoly case Exist-ing literature (Hao and Fan 2014 Tan and Carrillo2017) have studied the monopoly case and showedthat the marketplace mode always dominates thereseller mode due to the revenue sharing structure inthe marketplace mode The focus of our paper is toillustrate how upstream competition and order-fulfill-ment cost can alter the existing results and provideunique managerial insights into the online intermedi-ary managers For this reason the monopoly case isnot as pertinent as the competitive case In summaryto focus on the more relevant and realistic competitivecase we assume that the outside option is not the pre-ferred choice for the suppliersSecond in practice many major suppliers and
brand owners introduce their own direct-sale web-sites in addition to selling through popular onlineretailing platforms In such situations the supplierrsquosdecision within its separate online store will interactwith the selling mode (ie reseller or marketplace)and also the price in the online retailing platformThat is there will be cross-channel effects How willthe two channels interact with each other How willtheir interactions influence the suppliersrsquo and the
intermediaryrsquos preferences on selling modes and thecorresponding pricing decisions Ryan et al (2012)and Abhishek et al (2016) have provided someinsights but further studies are required We leave itfor future research to explicitly examine the effect ofcross-channel effects on the mode choice for onlineintermediaries Additionally note that there are also afew large-scale suppliersbrand owners who do notsell through online retailing platforms For exampleApple sells computers and accessories on Amazonbut it does not sell iPhones on Amazon We find someanecdotal evidence regarding why certain companies(eg Apple) choose not to sell on the third-partyonline marketplaces10 Essentially it is due to compe-tition as Amazon itself also provides many compet-ing products at the same time Although our modelcannot directly explain this rationale we think thisis also a promising research question for futurescholarsThird our analysis and expositions have focused
on the emerging online marketplace In reality thereexists offline store-within-store models The most typ-ical and representative example is the consignmentstore where the seller displays products in the storeand shares the revenue with the product ownersThough our research resonates with such settingsthere are some subtle but important differences withonline marketplaces To begin with typically it is theconsignment shop not the product owner who setsprices Second unlike the online marketplace con-signment shops typically sell second-hand and vin-tage products Third the revenue sharing contract isdifficult to implement in the offline setting as pointedout by Cachon and Lariviere (2005) The marketplaceidea is applicable to an offline setting but it is not aspopular as the online marketplace Finally a consign-ment storersquos reach is limited by consumer proximityAll these reasons might explain why consignmentstores do not proliferate in practiceExtensions of our work that may be of interest
would be to capture other factors such as complemen-tary markets asymmetric information and economiesof scale of order-fulfillment cost in addition to compe-tition intensity and order-fulfillment cost Anotherpossible avenue would be to investigate the optimalmode choice not only in the case of supplier compe-tition but also integrating downstream competitionwhere buyers might be in direct competition withone another Finally although we have confirmedthat our findings are consistent with practical obser-vations our research poses some testable empiricalquestions Our hope is that as this body of litera-ture expands in scope many other related researchissues will be investigated with a view to furtheringour understanding of the efficacy of onlinemarketplaces
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1608 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Acknowledgments
The authors contributed equally to this work The authorsthank the department editor the senior editor and twoanonymous reviewers for their valuable suggestions thathave significantly improved this study They also thankFuqiang Zhang Baojun Jiang and Qingning Cao for com-ments on earlier versions of the study This research is sup-ported in part by the National Natural Science Foundationof China (grants 71702093 71531005 and 71431004)
Notes
1httpwwwbusinesswire comnewshome20160614006063enProducts-Amazon-Carry-Categories (accessed date May10 2016)2Note that our linear demand functions can be derivedfrom the maximization problem of a representative con-sumer with a quadratic and strictly concave utility func-tion which is defined in Singh and Vives (1984)
Uethda dbTHORN frac14 hda thorn hdb ethd2a thorn 2wdadb thornd2bTHORN2
2 where frac14 1thorn c1thorn 2c
w frac14 c1thorn 2c and di is the amount of product i In specific
the representative consumer maximizes U(da db) pada pbdb where pi is the price of product iPlease refer to Singh and Vives (1984) for detailed clarifi-cations and analysis3It is trivial to show that these demand functions satisfythe regularity properties for product substitutes The ratioof price elasticity of demand for a supplierrsquos own productoffering is greater than the price elasticity of demand forthe substitute product offering since 1 + c gt c Note c hasan alternative interpretation which is product differentia-tion If consumers perceive that products are highly differ-entiated (ie low competition intensity) then the value ofc is low and vice versa4In our extension we have illustrated that the impact of avariable cost for order fulfillment is very similar to theimpact of a fixed order-fulfillment cost When the mar-ginal order-fulfillment cost is relatively small or moderateour main results remain the same when the marginalorder-fulfillment cost becomes much higher the resultsbecome similar to the case in which there is a very largefixed order-fulfillment cost5For mode RR (ie the intermediary choosing to act as areseller for both suppliers) our entire analysis is based onassuming that there are no economies of scale in order ful-fillment costs realized by the intermediary If such econo-mies do exist then these can be integrated by assumingtotal fulfillment costs under mode RR equal bF with(1 le b lt 2) Since order fulfillment costs have no impacton the equilibrium wholesale and market prices (andhence realized demands) for both suppliers then the onlyimpact would be on the corresponding optimal profits ofthe intermediary under mode RR which would increaseby (2 b)F Hence integrating economies of scale inorder fulfillment costs would not change our insights sub-stantively except that there would be changes in the pref-erence regions for the three strategy choices with thepreference regions for mode RR increasing and the prefer-ence regions for PR and PP would decline
6Note that we have implicitly assumed that the twosuppliers are symmetric for both order-fulfillment cost(eg FS) and the market potential(eg h) The analysisand results of asymmetric cases are discussed in section577Although one may argue that the condition pPRB pO
seems restrictive it not only reflects real-world practice(ie for certain products some suppliers choose to oper-ate under the reseller mode while others choose the onlinemarketplace option) but also market conditions By sellingthrough an intermediary suppliers can ensure access to alarger customer base and at the same time offer their cus-tomers the benefit of an integrated shopping experienceusing the intermediary platform Thus in a competitivesetting a supplier could choose to operate under the resel-ler mode even though its competitor chooses an alterna-tive mode8The reason to focus on the intermediaryrsquos profit ratherthan suppliersrsquo profits is that the intermediary can directlycontrol the channel mode by structuring the terms of theonline marketplace service Suppliers on the other handare assumed to make rational choices based on theseterms and hence play a passive role in the process9httpsmarketplacewalmartcomFAQ (accessed dateMay 10 2017)10httpswwwcnetcomnewswhy-doesnt-amazon-sell-iphones (accessed date October 10 2017)
ReferencesAbhishek V K Jerath Z J Zhang 2016 To platform-sell or
resell Channel structures in electronic retailing ManagementSci 62(8) 2259ndash2280
Agatz N A H M Fleischmann J van Nunen 2008 E-fulfill-ment and multi-channel distribution A review Eur J OperRes 187(2) 339ndash356
Birge J R J Drogosz I Duenyas 1998 Setting single-periodoptimal capacity levels and prices for substitutable productsInt J Flex Manuf Syst 10(4) 407ndash430
Brynjolfsson E J Hu M S Rahman 2009 Battle of the retailchannels How product selection and geography drive cross-channel competition Management Sci 55(11) 1755ndash1765
Cachon G M Lariviere 2005 Supply chain coordination withrevenue-sharing contracts Strengths and limitations Manage-ment Sci 51(1) 30ndash44
Cai G 2010 Channel selection and coordination in dual-channelsupply chains J Retail 86(1) 22ndash36
Cai G Y Dai S X Zhou 2012 Exclusive channels and revenuesharing in a complementary goods market Market Sci 31(1)172ndash187
Cattani K D W G Gilland J Swaminathan 2004 Coordinatingtraditional and internet supply chains International Series inOperations Research and Management Science D Simchi-Levi D Wu M Shen eds Modeling in the EBusiness EraKluwer Academic Publishers Boston MA 643ndash677
Chiang W K D Chhajed J D Hess 2003 Direct marketingindirect profits A strategic analysis of dual-channel supply-chain design Management Sci 49(1) 1ndash20
Choi S C 1991 Price competition in a channel structure with acommon retailer Market Sci 10(4) 271ndash296
Coughlan A T 1985 Competition and cooperation in marketingchannel choice Theory and application Market Sci 4(22)110ndash129
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce ModelProduction and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society 1609
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society
Gal-Or E T Geylani A Dukes 2008 Information sharing in achannel with partially informed retailers Market Sci 27(4)642ndash658
Garcia-Gallego A N Georgantzis 2001 Multiproduct activity inan experimental differentiated oligopoly Int J Ind Organ19(3ndash4) 493ndash518
Geng X Y Tan L Wei 2018 How add-on pricing interacts withdistribution contracts Prod Oper Manag 27(4) 605ndash623
Hagiu A 2007 Merchant or two-sided platform Rev Netw Econ6(2) 115ndash133
Hagiu A J Wright 2015 Marketplace or reseller ManagementSci 61(1) 184ndash203
Hao L M Fan 2014 An analysis of pricing models in the elec-tronic book market MIS Q 38(4) 1017ndash1032
Jerath K Z J Zhang 2010 Store within a store J Mark Res47(4) 748ndash763
Jiang B K Jerath K Srinivasan 2011 Firm strategies in the ldquoMidTailrdquo of platform-based retailingMarket Sci 30(5) 757ndash775
Kapner S 2014 How the web drags on some retailers The WallStreet Journal Available at httpwwwwsjcomarticleshow-the-web-drags-on-some-retailers-1417477790 (accessed dateNovember 30 2015)
Li L H Zhang 2008 Confidentiality and information sharingin supply chain coordination Management Sci 54(8)1467ndash1481
McGuire T W R Staelin 1983 An industry equilibrium analysisof downstream vertical integration Market Sci 2(2) 161ndash191
Netemeyer R G C M Heilman J G Maxham III 2012 Theimpact of a new retail brand in-store boutique and its per-ceived fit with the parent retail brand on store performanceand customer spending J Retail 88(4) 462ndash475
Ryan J K D Sun X Zhao 2012 Competition and coordinationin online marketplaces Prod Oper Manag 21(6)997ndash1014
Shy O Z Wang 2011 Why do payment card networks chargeproportional fees Am Econ Rev 101(4) 1575ndash1590
Singh N X Vives 1984 Price and quantity competition in a dif-ferentiated duopoly RAND J Econ 15(4) 546ndash554
Tan Y J Carrillo 2017 Strategic analysis of the agency modelfor digital goods Prod Oper Manag 26(4) 724ndash741
Tan Y J Carrillo H Cheng 2016 The agency model for digitalgoods Decis Sci 47(4) 628ndash660
Trivedi M 1998 Distribution channels An extension of exclusiveretailership Management Sci 44(7) 896ndash909
Tsay A A N Agrawal 2004a Channel conflict and coordinationin the e-commerce age Prod Oper Manag 13(1) 93ndash110
Tsay A A N Agrawal 2004b Modeling conflict and coordina-tion in multi-channel distribution systems A review Interna-tional Series in Operations Research and ManagementScience D Simchi-Levi D Wu M Shen eds Modeling inthe EBusiness Era Kluwer Academic Publishers Boston MA557ndash606
Wang J H Shin 2015 The impact of contracts and competitionon upstream innovation in a supply chain Prod Oper Manag24(1) 134ndash146
Young K J Chen S Raghunathan 2014 Online productreviews Implications for retailers and competing manufactur-ers Inf Syst Res 25(1) 93ndash110
Supporting InformationAdditional supporting information may be found onlinein the Supporting Information section at the end of thearticle
Appendix S1 Proofs of Lemmas and PropositionsAppendix S2 Analysis for Asymmetric Cases
Tian Vakharia Tan and Xu Strategic Analysis of an E-Commerce Model1610 Production and Operations Management 27(8) pp 1595ndash1610 copy 2018 Production and Operations Management Society