doctoral doctoral thesis thesis - diva...

DOCTORA L T H E S I S

Luleå University of TechnologyDepartment of Business Administration and Social Sciences

Industrial Marketing and e-Commerce Research Group

2007:70|: 02-5|: - -- 07⁄70 --

2007:70

Multiple Marketing Channel Conflict with a Focus on the Internet

A Dual Perspective

Lena Goldkuhl

Multiple Marketing Channel Conflict with a Focus on the Internet:

A Dual Perspective

LENA GOLDKUHL

Department of Business Administration and Social Science Industrial Marketing and e-Commerce Research Group

End Customers

Internet

ResellerCompany 2

Physical Store

Internet Store

Manufacturer

Physical Store

Internet Store

Catalogue

Internet Reseller

ResellerCompany 1

DOCTORAL THESIS

Multiple Marketing Channel Conflict with a Focus on the Internet:

A Dual Perspective

LENA GOLDKUHL

Luleå University of Technology Department of Business Administration and Social Science Industrial Marketing and e-Commerce Research Group

2007

With love and affection toPeter, Alexandra and Rickard

Mom and Dad

_______________________________________________________________________________ ABSTRACT ______

i

ABSTRACTThe emergence of electronic commerce has led many companies to complement their reseller networks with direct-to-the-customer Internet sales. This inevitably creates conflict. The way companies view and manage channel conflict is consid-ered to be an important success factor. Accordingly, the purpose of this research was to explore and describe managers’ perspectives on conflict in multiple market-ing channels, with a focus on the Internet.

To ground the research, personal, open-ended interviews were conducted with managers at Ducati and SAS. The results of these interviews, along with an ex-tended literature review, formed the foundation for a survey among resellers of those companies.

The results of this research question some of the premises in the channel conflict literature. Specifically, goal incompatibility and inadequate communication were found to have a direct and negative effect on resellers’ performance. In contrast, channel conflict per se was not detrimental to firms. It is postulated that it is the causes of conflict, and not necessarily conflict itself that are harmful. Clearly, goal incompatibility and inadequate communication are damaging to performance of firms.

If companies do not attend to goal compatibility and communication among chan-nel members, resellers are likely to end the relationship or otherwise behave in their own interests. For example, resellers that represent multiple firms are likely to discriminate against a company with whom they have poor communication and incompatible goals. Consequently, producers are encouraged to promote open communication with their channel members and ensure that their goals and those of their resellers are in alignment.

A tangible example of the importance of goal compatibility is found in the fact that a firm that involves its resellers in its Internet sales achieves positive outcomes. To put it simply, channel integration yields lower conflict and higher performance.

___________________________________________________________________ ACKNOWLEDGEMENTS ______

iii

ACKNOWLEDGEMENTSSeveral people have guided, supported, and inspired me during my work on this thesis. First of all, I would like to express my deepest thanks to Professor Esmail Salehi-Sangari. Your commitment and support is beyond what words can describe! My innermost gratitude and thanks also go to my supervisor, Professor Pierre Ber-thon, for your brilliant insights, invaluable comments and tremendous support. I would also like to thank Professor Leyland Pitt and my colleague Magnus Hultman for your valuable comments in the final seminar, which further improved this the-sis. I would like to thank Professor Tawfik Jelassi for your support during the first part of this research.

I also would like to thank all my wonderful colleagues (past and present) at Indus-trial Marketing & e-Commerce, for all your support during this process. I cannot fully express how fortunate I feel having such wonderful colleagues! A special ac-knowledgement is directed to my dear friend and colleague Maria Styvén. Thank you for countless discussions concerning all aspects of life inside and outside the world of research. Particular thanks are also directed to Lars-Ole Forsberg, since you are the one responsible for recruiting me to the research world.

This thesis, however, would never have been accomplished without respondents. Therefore, I would like to gratefully acknowledge the kind cooperation and sup-port of management at both Ducati and Scandinavian Airlines. I am also grateful to those resellers of Ducati and SAS who contributed information necessary for com-pleting this thesis.

I would like to express my gratitude to a number of organisations that financially supported this research, i.e. Handelns Utvecklingsråd, Längmanska Företagarfon-den, Nordbankens Norrlandsstiftelse, Luleå University of Technology, Norrbot-tens Forskningsråd, Mål 1 Norra Norrland, Innovationsbron Luleå AB, and Spar-banksstiftelsen Norrbotten.

Last, but certainly not least, my warmest thanks are directed to my family and friends. Your love and support have made these years of study a very enjoyable and unforgettable experience. Peter, you and our children Alexandra and Rickard, more than anyone, deserve special, heartfelt thanks for your support and patience during my work with this thesis. Thank you for filling my life with happiness and joy!

Luleå, November 2007

Lena Goldkuhl

____________________________________________________________________ TABLE OF CONTENTS ______

v

TABLE OF CONTENTS

1 INTRODUCTION ......................................................................... 11.1 Introduction ......................................................................................... 11.2 The Internet as a Marketing Channel ................................................... 1

1.2.1 Definition of Marketing Channels...............................................................21.2.2 Definition of Multiple Marketing Channels ................................................3

1.3 Advantages with Multiple Marketing Channels..................................... 51.3.1 Illustrative Example of a Multiple Marketing Channel System .....................6

1.4 Challenges with Multiple Marketing Channels ..................................... 61.4.1 Definition of Channel Conflict ...................................................................8

1.5 Research Purpose ................................................................................. 81.6 Disposition of the Thesis ...................................................................... 9

2 STUDY I: PRODUCERS’ PERSPECTIVE ........................................ 112.1 Managing Multiple Marketing Channels..............................................11

2.1.1 Causes of Channel Conflict ......................................................................112.1.2 Channel Conflict Issues When Adding the Internet...................................172.1.3 Assessing the Seriousness of Channel Conflict ...........................................182.1.4 Minimising Channel Conflict ...................................................................22

2.2 Research Questions & Frame of Reference .........................................272.2.1 Research Questions ..................................................................................272.2.2 Developing a Conceptual Frame of Reference ..........................................302.2.3 Conceptual Frame of Reference ...............................................................36

2.3 Methodology.......................................................................................372.3.1 Research Purpose .....................................................................................382.3.2 Research Approach ..................................................................................382.3.3 Research Strategy .....................................................................................392.3.4 Case Study ...............................................................................................392.3.5 The Empirical Investigation......................................................................412.3.6 Issues of Validity and Reliability ...............................................................43

2.4 Empirical Evidence..............................................................................452.4.1 Case One: Ducati Motor Holding SpA .....................................................452.4.2 Case Two: Scandinavian Airlines ..............................................................57

2.5 Analysis ...............................................................................................682.5.1 Within-Case Analysis – Ducati .................................................................682.5.2 Within-Case Analysis - SAS......................................................................782.5.3 Cross-Case Analysis ..................................................................................86

2.6 Conclusions.........................................................................................962.6.1 Findings and Conclusions .........................................................................962.6.2 Theoretical Contributions....................................................................... 1032.6.3 Managerial Implications .......................................................................... 1042.6.4 Suggestions for Future Research ............................................................. 105

______ TABLE OF CONTENTS ____________________________________________________________________

vi

3 STUDY II: RESELLERS’ PERSPECTIVE........................................ 1073.1 Research Purpose and Research Questions ........................................107

3.1.1 Research Question One ......................................................................... 1073.1.2 Research Question Two......................................................................... 1083.1.3 Research Question Three....................................................................... 108

3.2 Delimitations.....................................................................................1083.2.1 Commitment ......................................................................................... 1093.2.2 The Extranet .......................................................................................... 1093.2.3 Functionality of Conflict and Impact on Brand ....................................... 1093.2.4 Power and Interdependence ................................................................... 110

4 DEVELOPING A RESEARCH MODEL ......................................... 1124.1 Causes of Channel Conflict ...............................................................112

4.1.1 Goal Incompatibility............................................................................... 1134.1.2 Domain Dissensus................................................................................... 1154.1.3 Inadequate Communication.................................................................... 117

4.2 Channel Conflict...............................................................................1194.2.1 Level of Channel Conflict....................................................................... 121

4.3 Outcomes .........................................................................................1224.3.1 Performance ........................................................................................... 1224.3.2 Satisfaction ............................................................................................. 124

4.4 Research Model & Hypothesis Overview..........................................1265 METHODOLOGY ..................................................................... 128

5.1 Research Purpose ..............................................................................1285.2 Research Approach ...........................................................................1285.3 Research Strategy ..............................................................................1295.4 The Empirical Investigation...............................................................129

5.4.1 Sample ................................................................................................... 1305.4.2 Questionnaire Development ................................................................... 1305.4.3 Data Collection and Response Rate........................................................ 1345.4.4 Data Analysis .......................................................................................... 1365.4.5 Issues of Validity and Reliability ............................................................. 140

6 RESULTS & ANALYSIS.............................................................. 1446.1 Descriptive Statistics ..........................................................................1446.2 Normality Analyses............................................................................148

6.2.1 Normality Analysis - Ducati.................................................................... 1486.2.2 Normality Analysis - SAS ....................................................................... 148

6.3 Missing Data......................................................................................1496.3.1 Missing Data Analysis - Ducati................................................................ 1496.3.2 Missing Data Analysis - SAS ................................................................... 149

6.4 Measure Validation............................................................................1496.4.1 Exploratory Factor Analysis of the Causes of Channel Conflict................ 1506.4.2 Exploratory Factor Analysis of the Level of Channel Conflict .................. 1516.4.3 Exploratory Factor Analysis of Outcomes................................................ 1526.4.4 Confirmatory Factor Analysis.................................................................. 153


vii

6.4.5 Reliability Analyses of Resulting Scales ................................................... 1546.4.6 Validity & Common Method Variance ................................................... 1576.4.7 Summated Averaged Scales ..................................................................... 160

6.5 Hypothesis Testing ............................................................................1616.5.1 Structural Equation Modelling - SAS ...................................................... 1616.5.2 Hypothesis Testing – SAS....................................................................... 1676.5.3 Hypothesis Testing - Ducati ................................................................... 168

7 DISCUSSION, CONCLUSIONS & CONTRIBUTIONS .................... 1737.1 Discussion .........................................................................................173

7.1.1 Causes of Channel Conflict .................................................................... 1747.1.2 Level of Channel Conflict....................................................................... 1787.1.3 Relationship between Level of Channel Conflict and Outcomes ............. 179

7.2 Limitations, Contributions and Conclusions ......................................1797.2.1 Limitations of the Study.......................................................................... 1807.2.2 Theoretical Contributions....................................................................... 1807.2.3 Managerial Implications .......................................................................... 187

7.3 Suggestions for Future Research ........................................................189REFERENCES .......................................................................................

APPENDICES

Appendix 1 Examples of Cause-Related Issues from the Research Instrument used in the Three Channel Dyads .......................................................................................................Appendix 2 Empirical Studies of Channel Conflict in Marketing Channels ........................Appendix 3 Interview Guide .............................................................................................Appendix 4 Questionnaire.................................................................................................Appendix 5 Sample Cover Letter.......................................................................................

FIGURES

Figure 1.1 Use of Physical and Virtual Channels in an Integrated Click-and-Mortar Business...........................................................................................................................5Figure 1.2 Example of a Multiple Marketing Channel System ..........................................6Figure 1.3 Disposition of Thesis.......................................................................................9Figure 2.1 Dual Distribution with On-line Selling: Channel Structure Options ..............14Figure 2.2 The Intrachannel Conflict Process.................................................................18Figure 2.3 Decision-Making Framework .......................................................................20Figure 2.4 Conceptual Frame of Reference....................................................................37Figure 2.5 Ducati’s Marketing Channel System..............................................................47Figure 2.6 The MH900e ...............................................................................................48Figure 2.7 SAS’s Marketing Channel System..................................................................58Figure 3.1 The Intrachannel Conflict Process............................................................... 107Figure 4.1 Research Model ......................................................................................... 126Figure 6.1 Performance ............................................................................................... 153Figure 6.2 Level of Channel Conflict........................................................................... 154

______ TABLE OF CONTENTS ____________________________________________________________________

viii

Figure 6.3 Outcomes Part of Research Model ............................................................. 162Figure 6.4 Research Model ......................................................................................... 163Figure 6.5 Re-Specified Research Model .................................................................... 165Figure 7.1 Empirically Derived Model......................................................................... 183

TABLES

Table 2.1 Measures of the Causes of Channel Conflict ...................................................32Table 2.2 Measures of the Seriousness of Channel Conflict ............................................33Table 2.3 Measures of the Conflict Reducing Approaches .............................................34Table 2.4 Conflict Reducing Pricing Approaches when Adding the Internet..................35Table 2.5 Case Study Tactics for Four Design Tests .......................................................43Table 2.6 Within-Case Analysis of the Causes of Channel Conflict – Ducati ..................69Table 2.7 Within-Case Analysis of the Seriousness of Channel Conflict – Ducati ...........71Table 2.8 Within-Case Analysis of Conflict Reducing Approaches – Ducati ..................76Table 2.9 Within-Case Analysis of the Causes of Channel Conflict – SAS......................78Table 2.10 Within-Case Analysis of the Seriousness of Channel Conflict – SAS .............81Table 2.11 Within-Case Analysis of Conflict Reducing Approaches – SAS ....................84Table 2.12 Cross-Case Analysis of the Causes of Channel Conflict.................................87Table 2.13 Cross-Case Analysis of the Seriousness of Channel Conflict ..........................89Table 2.14 Cross-Case Analysis of Conflict Reducing Approaches .................................94Table 2.15 Conflict-Reducing Pricing Approaches When Adding the Internet ............ 104Table 4.1 Previous Measures of Marketing Channel Conflict ....................................... 121Table 4.2 Overview of Hypotheses.............................................................................. 127Table 5.1 Conceptual and Operational Definitions of the Constructs ........................... 131Table 5.2 Strength of Association between Variables .................................................... 139Table 5.3 Reference Values to Test Model Fit ............................................................. 140Table 6.1 Descriptive Statistics – Ducati and SAS......................................................... 145Table 6.2 Factor Analysis of the Causes of Channel Conflict ........................................ 150Table 6.3 Factor Analysis of the Level of Channel Conflict .......................................... 151Table 6.4 Rotated Component Matrix ........................................................................ 152Table 6.5 Fit Indices for Scales..................................................................................... 154Table 6.6 Reliability Analysis of Scales......................................................................... 155Table 6.7 Factor Analysis of Retained Indicators - SAS ................................................ 158Table 6.8 Discriminant and Convergent Validity - Bivariate Correlations..................... 159Table 6.9 Summated Averaged Scales - SAS ................................................................ 161Table 6.10 Summated Averaged Scales - Ducati........................................................... 161Table 6.11 R2-Correlations for Outcomes Part of Research Model .............................. 161Table 6.12 Maximum Likelihood Estimates - Outcomes Part of Research Model......... 162Table 6.13 Modification Indices and Parameter Change Statistics ................................. 163Table 6.14 R2 Correlations for the Re-Specified Research Model................................ 165Table 6.15 Maximum Likelihood Estimates - Re-Specified Model............................... 166Table 6.16 Fit Indices for the Tested Models ............................................................... 166Table 6.17 Hypothesis Testing – SAS (AMOS Estimates)............................................. 167Table 6.18 Multiple Regression on Level of Channel Conflict ..................................... 169Table 6.19 Simple Regression of Goal Incompatibility on Domain Dissensus ............... 170Table 6.20 Simple Regression of Inadequate Communication on Goal Incomp............ 170


ix

Table 6.21 Simple Regression of Inadequate Communication on Domain Diss. ........... 170Table 6.22 Multiple Regression on Performance.......................................................... 171Table 6.23 Multiple Regression on Satisfaction............................................................ 172Table 6.24 Hypothesis Testing – Ducati (Regression Analysis) ..................................... 172Table 7.1 Hypothesis Tests – Comparison of Ducati and SAS ...................................... 173Table 7.2 Overview of Propositions ............................................................................ 183

“Companies are recognizing that success in the new economy will go to those who can execute clicks-and-mortar strategies

that bridge the physical and the virtual worlds”

- Gulati and Garino (2000, p. 107)

_________________________________________________________________________ INTRODUCTION ______

1

1 INTRODUCTIONThis chapter will serve as an introduction to the purpose of this thesis, which is to explore and describe managers’ perspectives on conflict in multiple marketing channels, with a focus on the Internet. Definitions of some key issues will be provided. The chapter ends with a specification of the research purpose and an overview of the disposition of this thesis.

1.1 Introductionhe emergence of e-commerce1 has created a new business paradigm, one that presents marketers with striking opportunities, including reduced costs, access to new market segments and the ability to provide informa-

tion worldwide on a continuous basis (Webb, 2002). Even so, e-commerce does not come without problems. That is, the addition of the Internet has made channel conflict the top issue for many businesses today (Gilbert & Bacheldor, 2000). The way in which businesses manage this channel conflict is an important factor in their success (Webb, 2002). Consequently, a particularly important aspect of e-commerce is its impact on marketing channels2 (ibid.).

1.2 The Internet as a Marketing Channel lthough the Internet is an entirely new marketing channel3 it shares many of its characteristics with other conventional marketing channels (Peterson, Balasubramanian & Bronnenberg, 1997). However, there are some excep-

tions. Some of the unique characteristics of the Internet marketing channel, as compared to traditional marketing channels, include the ability to inexpensively store vast amounts of information, the availability of inexpensive means of search-ing, the organising and disseminating of such information, interactivity and the ability to supply information on demand, the ability to provide perceptual experi-ences far superior to a printed catalogue, the ability to mediate transactions, and the relatively low entry and setting up costs for sellers (Peterson et al., 1997; Kiang, Raghu & Shang, 2000).

Pitt, Berthon and Berthon (1999) discuss the irrelevance of location, the homogenization of time and the death of distance. The irrelevance of location means that any screen-based activity can be operated from anywhere in the world. The homogenization of timemeans that a web site is always open. Sellers do not have to be awake to serve cus-tomers, and in the Christmas rush, for instance, customers do not have to queue up

1 e-commerce, i.e. electronic commerce, refers to the process of buying, selling, or exchanging goods, services and information via computer networks, including the Internet (Turban, King, Lee, Warkentin & Chung, 2002). 2 Authors writing about channels of distribution from a marketing perspective use the terms distri-bution channels and marketing channels interchangeably. 3 Marketing channels will be defined in Section 1.2.1

T

A

______ CHAPTER 1 ______________________________________________________________________________

2

in order to purchase gifts. Time can thus be made uniformly consistent, or ho-mogenized, for all buyers and sellers. The death of distance refers to the ability of the Internet to serve as a physical distribution channel for digital products, such as pic-tures, videos, sound and words. In such cases, the costs of distribution is not af-fected by distance (ibid.).

One of the main assumptions of the Internet is that in many cases it eliminates in-termediaries (cf. Pitt et al., 1999). Intermediaries, however, perform a number of activities that are not easily duplicated by manufacturers. These include making in-ventory immediately available when customers want it, and providing services that allow one-stop shopping, such as personal assistance, dressing rooms, repairs and returns (Levy & Weitz, 2004). Except for one-stop shopping, these activities can-not be fully provided by the Internet marketing channel alone. Therefore, apart from the distribution of digital products, the Internet marketing channel probably will not eliminate or serve as a substitute for conventional retail channels (Peterson et al., 1997). Thus, in most cases, companies will add the Internet to existing mar-keting channels, thereby employing multiple marketing channels. Before a discus-sion on the use of more than one marketing channel, definitions of marketing channels and multiple marketing channels will be provided.

1.2.1 Definition of Marketing Channels Authors writing about channels of distribution from a marketing perspective use the terms distribution channels and marketing channels interchangeably. Stern and El-Ansary (1977), for instance, describe a “marketing or distribution channel” as “an inter-organization system comprised of a set of interdependent institutions and agencies involved with the task of moving anything of value from its point of con-ception, extraction, or production to points of consumption” (p. 4). Coughlan, Anderson, Stern and El-Ansary (2001), on the other hand, employ only the term marketing channel: “A marketing channel is a set of interdependent organizations involved in the process of making a product or service available for use or con-sumption” (p. 3). O’Connor and Frew (2002), when writing about the Internet, use the term distribution channel, and define it as a mechanism that provides “suf-ficient information to the right people at the right time and in the right place to al-low a purchase decision to be made, and also allows the consumer to make a reser-vation and pay for the required product” (p. 33).

The activities or functions of a marketing channel that provide the services de-manded by end users are labelled marketing flows (Vaile, Grether & Cox, 1952). Eight generic marketing flows exist, namely, physical possession, ownership, promotion, negotiation, financing, risking, ordering and payment (ibid.). Physical possession refers to all storage activities, including transportation between two channel members (Coughlan et al., 2001). Ownership refers to the transfer of ownership from one or-ganisation or person to another. Promotion includes personal selling, media advertis-

_________________________________________________________________________ INTRODUCTION ______

3

ing, sales promotions, publicity, etc. Negotiation refers to the effort to reach a final agreement on prices and other terms, so that the delivery can occur. Financing re-fers to the terms of sale, such as payment in thirty days, and risking is concerned with price guarantees, warranties, insurance, repair, after-sales service, etc. Orderingand payment are the flows concerned with the actual purchase and payment of the offering. Sometimes, more than one channel member is needed in order to per-form all these flows (ibid.).

The definitions of Stern and El-Ansary (1977) and Coughlan et al. (2001) postulate that marketing channels are inter-organisational arrangements. Today, though, di-rect marketing channels, including the Internet, compete against, become a substi-tute for, or complement conventional retail channels (Balasubramanian, 1998). O’Connor and Frew (2002) define the Internet as a distribution channel. Their definition, however, is too narrow, since it is limited to consumers and does not include the actual delivery to customers. In this thesis, the term marketing channels will be used, as it implies that issues will be addressed from a marketing perspec-tive. Hence, the following definition will be employed:

A marketing channel is a system involved, either directly or indirectly, with making some-thing of value available for use or consumption.

This definition does not exclude any of the eight flows described above. In addi-tion, if a marketing channel cannot perform these flows on its own, they can be accomplished indirectly by using another marketing channel. For instance, when distributing non-digital products bought on the Internet, the actual delivery, i.e. physical acquisition, of the product will be handled indirectly.

Besides the Internet, marketing channels consist of a variety of retail formats, com-pany sales forces, distributors, agents, call centres and catalogues. These can be clas-sified as direct marketing channels and indirect marketing channels. Direct marketing channels are those that flow directly from the company to the customers, e.g., the Internet or the company sales force. Indirect marketing channels, on the contrary, refers to the use of another company to reach customers, e.g., distributors or agents When adding an online presence to an existing channel, businesses become multi-ple marketing channel entities. So, how can multiple marketing channels be de-fined?

1.2.2 Definition of Multiple Marketing Channels The terminology used to describe what in this thesis is referred to as multiple mar-keting channels has varied considerably in the marketing literature, ranging from multimarketing (Weigand, 1977), dual distribution (Stern & El-Ansary, 1988), mul-tichannel marketing systems (Kotler, 1988), dual-multiple channels (Frazier, Sawhney & Shervani, 1990), hybrid marketing systems (Moriarty & Moran, 1990) and hybrid dis-

______ CHAPTER 1 ______________________________________________________________________________

4

tribution systems (Webb, 1997) to multiple channels (Frazier, 1999). Some important differences exist among these terms.

Some of the definitions include the term “same basic product,” or variations on that theme: multimarketing occurs “when a company uses separate channels to sell the same product to markets differing in some important way” (Weigand, 1977, p. 98); “The term dual distribution describes a wide variety of marketing arrangements by which a manufacturer or a wholesaler reaches its final markets by employing two or more different types of channels for the same basic product” (Stern & El-Ansary, 1988, p. 403), and when “more than one primary channel is used to sell the same product line to the same target market”, businesses are using multiple channels (Frazier, 1999, p. 232).

Kotler (1988) focuses on the targeting of customers: multichannel marketing takes place when “a single firm uses two or more marketing channels to reach one or more customer segments” (p. 368). Webb (1997) likewise focuses on targeting cus-tomers but also includes customer needs: a hybrid distribution system “is defined as a multichannel arrangement characterized by the execution of distribution tasks among a combination of distinct channels, direct and/or indirect, resulting in a va-riety of marketing mix offerings designed to satisfy the needs of diverse target mar-ket segments” (p. 39). In addition, in certain instances, “two or more of the distri-bution channels used by the supplier firm may cover the same target segment(s), al-though such overlap in the marketplace is not necessarily by design” (Webb, 1997, p. 2). This definition takes note of the fact that some companies use a mix of both direct and indirect channels to best serve customers. It also captures the possibility of targeting the same customer segment(s) with more than one marketing channel, an issue which Alba et al. (1997), among others, state is of particular concern since the introduction of the Internet. It further includes the possibility of selling differ-ent product versions or brands in different channels; these strategies will be dis-cussed further in Chapter 2. However, following from the discussion of marketing channels and distribution channels in Section 1.2.1, the term multiple marketing channels will be used in this thesis. Accordingly:

Multiple marketing channels are a multi-channel arrangement characterized by the execution of distribution tasks among a combination of distinct channels, direct and/or indirect, resulting in a variety of marketing mix offerings designed to satisfy the needs of diverse target segments. Two or more marketing channels used by the supplier may cover the same target segment(s), although such overlap in the marketplace is not necessarily by design.

So, why do companies employ multiple marketing channels?

_________________________________________________________________________ INTRODUCTION ______

5

1.3 Advantages with Multiple Marketing Channels ompanies use multiple marketing channels, among other reasons, to cut costs and increase sales growth through extended market coverage (Moriarty & Moran, 1990). One of the advantages of multiple marketing

channels is that each channel presents a set of unique characteristics, which offers businesses the opportunity to adapt to changing customer needs and purchasing patterns. Consequently, multiple marketing channels are needed when customers have different service needs, when customers respond differently to promotions, or when customers’ price responses differ (Anderson, Day & Rangan, 1997; Frazier, 1999). In other words, different customers with different buying behaviours will use the channels that best serve their needs. When purchasing, customers can move from one channel to another at different stages of a single transaction. The integra-tion of the Internet channel and a physical channel has been illustrated by Stein-field, Bouwman and Adelaar (2002) (see Figure 1.1).

Figure 1.1 Use of Physical and Virtual Channels in an Integrated Click-and-Mortar Business (Steinfield, Bouwman & Adelaar, 2002, p. 94)

For instance, a customer might gather information about a product online (A1), purchase it at a physical outlet (B0) and obtain after-sales service online (C1). The more integrated the channels become, the more paths are possible. Forrester (2004) found that 65 percent of all U.S. online consumers who have researched a product online purchased that product offline. Almost half of these consumers preferred the offline channel, since they wanted to see the item before purchasing it. Some mul-tiple channel businesses, i.e. Eddie Bauer (www.eddiebauer.com), even give cus-tomers the option of returning items bought through the web site or the catalogue, by mail or in any of their physical stores. There are several reasons for offering this service: it provides an enormous convenience to web shoppers, an advantage that pure Internet-based companies cannot match, and it helps the physical stores by getting more customers through the door (Gulati & Garino, 2000). In addition, customers purchasing through multiple channels spend more than their single-channel counterparts (Ashcraft, 2001). Once in stores, besides purchasing the product they originally researched online, multiple-channel shoppers on average spend over 30 percent more on impulse purchases (Forrester, 2004).

C

http://www.eddiebauer.com

______ CHAPTER 1 ______________________________________________________________________________

6

As can be seen, multiple marketing channels present several advantages, both for managers and for customers. So, what does a multiple marketing channel system look like?

1.3.1 Illustrative Example of a Multiple Marketing Channel System An example of a multiple marketing channel system is depicted in Figure 1.2.

Figure 1.2 Example of a Multiple Marketing Channel System

Mattel, the toy manufacturer, appears to work in this way when selling some of its Fisher-Price toys. For example, in November 2004, Sesame Street's E-L-M-O was available for US$ 30 at Mattel’s Fisher-Price online web site (www.fisher-pricestore.com), and for the same price in Fisher-Price’s print catalogue. It could also be bought through indirect marketing channels at www.buy.com, a pure Internet retailer, for US$ 24.99; at www.toysrus.com, the online business of Toys R Us, for US$ 19.95; at www.walmart.com, the online business of Wal-Mart, for US$ 19.72; and probably also in the physical stores of both Toys R Us and Wal-Mart.

Even though businesses might benefit from the use of multiple marketing channels, the system presents some challenges as well.

1.4 Challenges with Multiple Marketing Channels hen businesses introduce the Internet channel, channel conflict will oc-cur (e.g., Smith, Bailey & Brynjolfsson, 1999), since traditional chan-nels, e.g., resellers, retailers, distributors, dealers and even internal sales

organisations, will feel threatened by the online e-commerce (Friedman & Furey, 1999; Gilbert & Bacheldor, 2000). This is understandable, because adding the Internet makes the company’s Internet channel a competitor of its physical chan-

W

End Customers

ResellerCompany 2

PhysicalStore

Internet Store

Manufacturer

Pure Internet Reseller

ResellerCompany 1

PhysicalStore

Internet Store

CatalogueInternet

http://www.fisher-pricestore.com



http://www.buy.com

http://www.toysrus.com

http://www.walmart.com

_________________________________________________________________________ INTRODUCTION ______

7

nel(s) (Smith et al., 1999). Each faces a possible loss of revenue as well as competi-tion for ownership of customers. For instance, manufacturers can always price lower and, in some cases, even provide more efficient service than distributors, which will cause customers to migrate from distributors to direct channels (Friedman & Furey, 1999). Such competition is often labelled cannibalization. It occurs when one marketing channel grows at the expense of another channel, rather than providing incremental revenue for the firm (Webb, 1997). Existing business partners, therefore, will fear that the introduction of the Internet channel, in the end, will lead to the dismantling of their marketing channel (Gilbert & Bacheldor, 2000). There is a great fear of alienating resellers among companies that are establishing e-commerce sites, and especially among those that are establishing direct links to customers for the first time. The only greater fear is having no e-commerce presence at all (ibid.).

There are situations, though, in which channel conflict can be functional. Without conflict, systems can become passive, non-innovative, and, eventually, non-viable (Stern & Heskett, 1969). Channel conflict can even motivate channel members to adapt, grow and seize new opportunities (Stern & El-Ansary, 1988). In addition, a company with no revenue in conflict may be sacrificing coverage, thereby failing to attract new customers (Moriarty & Moran, 1990). According to Friedman and Furey (1999), channel conflict between a company’s direct channels is a manage-able problem. Such conflict is just another way of saying that the manufacturer of-fers customers multiple purchasing opportunities. The growth of one channel at the expense of another direct channel will have no net negative effect on revenue. Their conclusion, therefore, is that the only channel conflict worth caring about is that between a manufacturer’s direct channels and its resellers (ibid.). Webb (2002), on the other hand, states that conflicts among groups within the supplier firm and between two separate organisations are equally important, since conflict in one set-ting is likely to result in conflict in the other. Accordingly, channel conflict, re-gardless of setting, should not be ignored.

The reason for this is that conflict is direct, personal and opponent-centred behav-iour, and it can therefore degenerate quickly into actions intended to destroy, in-jure, or impede another party in an interdependent relationship (Reve & Stern, 1979; Stern & El-Ansary, 1988). Such dysfunctional4 conflict can be harmful, since the conflicting objectives of various channels can lead to internal conflicts over customers, raising the potential for customer confusion and dissatisfaction that, in the end, will lead to declining profitability for the company (Cespedes & Corey, 1990; Moriarty & Moran, 1990). Consequently, according to a survey of 50 con-sumer-goods manufacturers conducted by Forrester (Gilbert & Bacheldor, 2000),

4 Dysfunctional is used in the sense of having disadvantageous or unwanted consequences (cf. Gat-torna, 1978), i.e. in this thesis dysfunctional and dangerous conflict will be treated as synonyms.

______ CHAPTER 1 ______________________________________________________________________________

8

channel conflict is the most serious issue facing management in its online sales strategy.

1.4.1 Definition of Channel Conflict Although several definitions of channel conflict have been offered in the literature (e.g., Stern & Gorman, 1969; Lusch, 1976b; Robicheaux & El-Ansary, 1976; Stern & El-Ansary, 1977; Eliashberg & Michie, 1984; Gaski, 1984; Gaski & Nevin, 1985; Stern & El-Ansary, 1988), the common theme is that channel conflict exists if one channel member perceives another channel member to be engaged in be-haviour that prevents or impedes it from attaining its goal(s). Since the definitions are in essence the same, the following definition, adopted from Stern and El-Ansary (1988, p. 285), will be used in this thesis:

Channel conflict is a situation in which one channel member perceives another channel mem-ber or members to be engaged in behaviour that prevents or impedes it from achieving its goals.

Steinfield et al. (2002) claim that the avoidance of channel conflict is critical for click-and-mortar companies if they are to achieve the possible benefits of multiple marketing channels. It is, therefore, important to identify how it can be avoided, and how cooperative behaviour can be encouraged (ibid.).

1.5 Research Purpose n Chapter 1 it was established that the emergence of e-commerce has led more businesses to add the Internet to their existing marketing channels. For busi-nesses, the advantages of multiple marketing channels, such as reduced costs,

increased market coverage and increased sales, are striking (e.g., Moriarty & Moran, 1990; Forrester, 2004). However, multiple marketing channels present challenges as well. With the addition of the Internet channel, channel conflict will follow (Smith et al., 1999). Although some conflict can be beneficial, it soon can degenerate into actions intended to destroy, injure, or impede another party in an interdependent relationship (Stern & El-Ansary, 1988). Such dysfunctional channel conflict can be harmful, as it raises the potential for customer confusion and dissat-isfaction, which eventually leads to declining profitability for the company (e.g., Moriarty & Moran, 1990). Accordingly, channel conflict seems to be the most se-rious concern for management when adding the Internet channel (Gilbert & Bacheldor, 2000).

Despite the importance of this issue, and despite repeated calls for research on mul-tiple channel conflict (cf. Frazier, 1999; Mols, 2000; Steinfield et al., 2002; Webb, 2002; Webb & Hogan, 2002; Rangaswamy & Van Bruggen, 2005), the topic has

I

_________________________________________________________________________ INTRODUCTION ______

9

received minimal attention in the academic literature. Hence, the research purpose of this thesis is:

1.6 Disposition of the Thesis n this section, the disposition of this thesis will be provided (see Figure 1.3).Chapter 1 presents the purpose of the research. Chapter 2 includes a study of the producers’ perspective, i.e. the major part of the licentiate thesis that was

defended and published in 2005 (cf. Goldkuhl, 2005). It is included because it is needed to draw conclusions between the two studies, and also to give a richer pic-ture of the channel conflict in the companies studied.

Figure 1.3 Disposition of Thesis

Chapter 3 introduces the study of the resellers’ perspective on channel conflict and provides a bridge between the two studies. This chapter will also include the re-search questions for Study II, as well as the delimitations of the study. In Chapter

I

Chapter 3 Study II: Resellers’ Perspective

Chapter 2Study I: Producers’ Perspective

Chapter 1Introduction and Research Purpose

Chapter 4 Developing a Research Model

Chapter 5 Methodology

Chapter 6 Results & Analysis

Chapter 7 Discussion, Conclusions & Contributions

To explore and describe managers’ perspectives on conflict in multiple marketing channels, with a focus on the Internet

______ CHAPTER 1 ______________________________________________________________________________

10

4, the research hypotheses and the research model is developed. Chapter 5 dis-cusses different methodological questions and approaches for Study II and describes the methodology used to answer the research questions posed in Chapter 3. Chap-ter 6 contains the results and analysis of the results of Study II. Chapter 7 discusses these findings and presents the conclusions drawn from both Study II and Study I, followed by the limitations of Study II and the implications of the research both for managers and for future research.

__________________________________________ STUDY I: PRODUCERS’ PERSPECTIVE ______

11

2 STUDY I: PRODUCERS’ PERSPECTIVEThis chapter will include the first part of this doctoral research, i.e. Study I, in which produc-ers’ perspectives on channel conflict was studied. The chapter will start with a review of chan-nel conflict theories, followed by the research questions, frame of reference, methodology, em-pirical evidence, and analysis. The chapter will end with the conclusions drawn from Study I, along with suggestions for future research.

2.1 Managing Multiple Marketing Channelss early as 1969, Stern and Heskett recognized that in order to prescribe how channel conflict can be reduced or resolved, an analysis of its causes has to be done. Accordingly, first, theories about the causes of channel

conflict will be presented, followed by channel conflict issues that arise when add-ing the Internet, theories about how to assess the seriousness of channel conflict, and finally, theories about how to minimize channel conflict.

2.1.1 Causes of Channel Conflict Causes are the underlying bases of conflict which the parties in conflict may not even be aware of (Ong, Elliott & Armstrong, 1990). Stern and Heskett (1969), in a conceptual paper, proposed three primary causes of conflict, i.e. goal incompatibility,domain dissensus and differing perceptions of reality. These are still considered the major causes of conflict in marketing channels (cf. Coughlan et al., 2001).

Goal incompatibility, and subsequent conflict, is very common (Coughlan et al., 2001). The goals of one channel member very often are not compatible with the goals of another channel member (Stern & Heskett, 1969; Schmidt & Kochan, 1972; Kochan, Huber & Cummings, 1975; Perry & Levine, 1976). With the use of resellers, this becomes even more apparent. Even though the use of resellers is much cheaper and offers the business greater flexibility as compared to a business-owned sales force (e.g., Friedman & Furey, 1999), coordinating relationships with resellers is a challenge, as they are independent businesses with multiple suppliers and product lines (Coughlan et al., 2001). Both the reseller and the supplier want to maximise their profits. However, their ideas of how this should be accomplished differ (Friedman & Furey, 1999). For the reseller, it could mean charging custom-ers more while at the same time holding down expenses. The supplier, on the other hand, wants to see the reseller do almost exactly the opposite (Coughlan et al., 2001). Accordingly, conflict issues might include whether prices and service are being maintained at “reasonable” levels (Reve & Stern, 1979).

Domain dissensus often occurs when one channel member perceives that the other is not taking proper care of its responsibilities within its domain. This can mean doing the job in the wrong way, not doing the job at all, or trying to do another channel member’s job (Coughlan et al., 2001). Stern and El-Ansary (1992) assert

A

______ CHAPTER 2 ______________________________________________________________________________

12

that the four critical elements of a channel domain are the population to be served, the territory to be covered, the functions or tasks to be performed and the tech-nology to be employed. Conflict issues might include who has the right to repre-sent a particular product within a given territory (Stern & El-Ansary, 1977; Reve & Stern, 1979; Magrath & Hardy, 1987). That is, when manufacturers use several marketing channels simultaneously, channels sometimes find themselves competing to reach the same set of customers. This might lead to distributors and dealers be-ing bypassed via direct selling, which almost certainly will cause channel conflict (Magrath & Hardy, 1987; Bucklin, Thomas-Graham & Webster, 1997). Another common reason for conflict is pre- and post-sales support; that is, marketing chan-nels argue over who should do it, how it should be done, and how they should be compensated. A further reason is inventory, since suppliers often consider it a resel-ler’s duty to carry a large inventory, while resellers instead want to be restocked quickly from a central location (Coughlan et al., 2001).

Even though individual goals within the marketing system may correspond with one another, and domains may be well defined, channel conflict still may occur due to differing perceptions of reality, i.e. channel members having conflicting percep-tions of the same situation (Stern & Heskett, 1969). This is an important source of conflict, because it indicates that there will be different reactions to the same situa-tion, which will frustrate channel members and produce conflict. One major rea-son for such misperception is that different channel members are exposed to vary-ing information and influences, giving them different parts of the overall picture (Coughlan et al., 2001). In order to achieve needed coordination among channel members, good communication within the channel is necessary (Mohr & Nevin, 1990). Too much communication is just as bad as too little, since too much con-tact can overload channel members and have dysfunctional consequences. There-fore, one should examine communication in terms of the amount of contact in re-lation to how much contact is necessary to conduct activities adequately (ibid.).

It must be noted, though, that when studying the above described causes of chan-nel conflict, Rosenberg and Stern (1971) were not able to support the causes of conflict for each of their researched dyads (manufacturer-distributors, distributors-dealers and manufacturer-dealers); each cause, however, was significant in at least two dyads. Examples of conflict issues used in their questionnaire are presented in Appendix 1.

Etgar (1979), on the other hand, discusses the causes of conflict in terms of attitudi-nal or structural causes. Attitudinal causes are associated with disagreements about channel roles, expectations, perceptions and channel communications. Role refers to the behaviour that is prescribed for a certain job position. Conflict can emerge if a per-son does not act in accordance with an established role, or when channel roles are not well defined for all channel members. Expectations among channel members


13

can be different, due to differences in information availability, information process-ing capacities, or experience. Channel members also may have different perceptionsof the channel and its environment. For instance, dealers may not pay attention to conditions in markets in which they do not operate, while manufacturers may em-phasize the broader implications of inter-channel competition and demand. Com-munication refers to the fact that in a marketing channel there is constant communi-cation. Information about new products, promotional campaigns, technical inno-vations, etc., has to be transferred from the producer to the resellers, while resellers have to inform the producer about market conditions, customers’ reactions to products and so on. When communication is not working properly, misunder-standings will occur, incorrect strategies will be implemented, and mutual feelings of frustration will arise (ibid.).

Structural causes consist of three sets of factors: goal divergence, competition for scarce re-sources and drive for autonomy (Etgar, 1979). Goal divergence as a cause of conflict has been described earlier in this section. Competition for scarce resources occurs when the demand for resources in a channel exceeds the available supply. For instance, dif-ferent channel members, such as dealers and wholesalers, may compete for exclu-sive rights to a special market. Drive for autonomy means that one party tries to exer-cise control over another party, such as a manufacturer attempting to control re-tailers’ pricing through suggested retail price lists (ibid.). In his research, Etgar (1979) found that conflict was generated primary by attitudinal factors.

What Etgar (1979) refers to as drive for autonomy is referred to by other authors as power (e.g., Hunt & Nevin, 1974). Power can be described as the ability to get someone to do something he/she would not have done otherwise (Gaski, 1984). Power in channel relationships received considerable attention during the 1970s. Not all those who recognise a connection between power and conflict share the same perspective. Some researchers claim that conflicts arise in a channel because superiors attempt to control the behaviour of subordinates, who resist such control, while other researchers view power as a result of, or response to, conflict. But most researchers acknowledge that there is a causal relationship between power and con-flict, and that it can, and does, proceed in either direction (ibid.). The above men-tioned causes of conflict were all identified before the Internet existed. So, what are the causes when the Internet is added?

Causes of Channel Conflict and the Internet Marketing Channel Conflicts caused by goal incompatibility, domain dissensus and differing perceptions of re-ality can arise when an online channel is created alongside, for instance, a bricks-and-mortar retail channel (Coughlan et al., 2001). Three different situations in which conflicts can occur due to online sales are depicted in Figure 2.1.

______ CHAPTER 2 ______________________________________________________________________________

14

Figure 2.1 Dual Distribution with On-line Selling: Channel Structure Options(Coughlan et al., 2001, p. 463)

Manufacturer (Tup-perware)

Standard Channel (inde-pendent direct salespeople)

Consumers

Owned Internet Sales Channel (Tupperware.com)

Manufacturer (Cal-laway Golf)

Standard Channel (pro shops, bricks-and-mortar

sport-golf outlets) Independent Pure-Play Internet

Sales Channel (buy.com)

(b) Manufacturer sells through third-party pure-play on-line reseller (e.g., Callaway Golf selling through buy.com)

Manufacturer (Simon & Schuster, publisher)

Standard Channel (bricks-and-mortar

bookstores)

(c) Manufacturer sells through some standard channels that do operate their own on-line stores and some that do not (e.g., books sold through bricks-and-mortar retailers and Barnes & No-ble’s on-line site, bn.com)

Barnes & Noble

Consumers

Barnes & Noble (bricks-and-mortar

bookstores)

Barnes & Noble (internet sales channel)

(bn.com)

(a) Manufacturer has own on-line presence (e.g., Tupperware; dotted lines indicates common ownership)

Consumers


15

In case (a), the manufacturer itself creates a company-owned online presence, in direct competition with its traditional reseller channel. Tupperware, for instance, sells its plastic containers in this way, i.e. both through independent direct sales people and on the Internet. In this case, disintermediation is said to occur, i.e. the in-termediary is bypassed by the manufacturer. In case (b), besides selling though its standard channel(s), the manufacturer decides to sell through a pure Internet re-tailer, i.e. a retailer that does not possess any physical stores. Callaway Golf Clubs, for instance, can be bought at buy.com, a pure Internet retailer. Case (c) describes the situation when the manufacturer’s reseller starts up its own online operation. A book publisher selling through Barnes & Noble is an example of this situation. Of course, any combination of these online strategies is possible (ibid.).

Case (a) and the Causes of Channel Conflict When adding the Internet channel, conflict caused by goal incompatibility can occur if the manufacturer wants to maximize its profits over all its marketing channels, one of which is the online channel, since the Internet channel may offer the manu-facturer a higher margin on sales than the traditional bricks-and-mortar channel does (Coughlan et al., 2001). In addition, in order to cover the costs of establishing and operating the new channel, customers may have to be pushed to buy directly from the online channel, rather than through the resellers (ibid.).

As the Internet allows businesses to establish a direct channel to customers, the emergence of e-commerce has made domain dissensus an even more burning issue (Alba et al., 1997). That is, a manufacturer selling online may directly cannibalize the customers of the bricks-and-mortar channel, thereby causing conflict about the population to be served (Coughlan et al., 2001). The bricks-and-mortar channel, of course, is convinced that the customers who chose to shop online at the manu-facturer’s site would have shopped at the physical store if the web site did not exist. Furthermore, the Internet pushes old territorial borders aside, since it can take away business from any location in the market area served by the bricks-and-mortar channel, thereby causing domain dissensus over territorial rights (ibid.).

Domain dissensus over the functions and duties to be performed by channel mem-bers also can arise when the company sells through a company-owned Internet channel (Coughlan et al., 2001). A customer can visit a bricks-and-mortar store, inspect the product, ask the personnel some questions about the product, check the price, and then buy the product online. This is a classic example of free-riding, since the bricks-and-mortar retailer bears the cost of serving the customer, but it receives no compensation for it because it does not get the sale on the product. A further domain dissensus can occur if a customer who has bought from the manu-facturer’s Internet channel decides to return the product, and tries to do so at a bricks-and-mortar retailer. Generally, the retailer will not agree to repurchase the item, since otherwise it will bear the cost without compensation (ibid.).

______ CHAPTER 2 ______________________________________________________________________________

16

Differing perceptions of reality also frequently occur between manufacturers and resel-lers in case (a) (Coughlan et al., 2001). Manufacturers often believe that they are just expanding their market reach, generating sales from consumers who cannot or will not purchase the company’s products in any other way. Actually, as many as 38 percent of survey respondents who spent more on holiday gifts online in 1999 than in 1998 gave as their reason that they were able to purchase products online that were not accessible to them near where they lived or worked. Even so, bricks-and-mortar resellers probably will perceive that sales made on the manufacturer’s web site have been stolen from them (ibid.).

Case (b) and the Causes of Channel Conflict When the manufacturer decides not to sell online itself, but to use a pure Internet retailer (e-tailer) instead, conflict caused by goal incompatibility, domain dissensus and/or differing perceptions of reality might occur (Coughlan et al., 2001). In this case, however, the conflict might not be limited to the manufacturer and its resellers; it also might occur between the bricks-and-mortar retailer and its online counterpart directly. The two competing resellers are companies that probably have very dif-ferent goals. The e-tailer often has a strong need to generate sales volume and mar-ket share in order to maintain its flow of venture capital, as it seeks to grow and dominate the online channel for its product type. Therefore, the e-tailer might cut the price, or offer free delivery or some other incentive in order to sell. Bricks-and-mortar resellers, on the other hand, focus on a shorter-term horizon for profit and cost covering, charge higher prices, and find it hard to compete. Free-riding is, of course, also possible in this channel structure. Accordingly, the bricks-and-mortar retailer faces the higher cost structure arising from its provision of pre-sale service, further limiting its ability to compete on price with the e-tailer, who en-joys a lower marginal cost of sales. It should be noted that, in case (b), the manu-facturer has no direct control of the pricing or service offered by the different re-sellers. The manufacturer is, therefore, somewhat limited in its ability to prevent channel conflict between the traditional retailer and the e-tailer (ibid.).

Case (c) and the Causes of Channel Conflict Online sales also occur in case (c), but here it is one of the manufacturer’s resellers that operates both online and through bricks-and-mortar stores (Coughlan et al., 2001). It might seem as if this is the same conflict situation as in case (b), but there is a major difference: in case (c), the online reseller internalises, at least in part, the negative effects of aggressive online competition on the bricks-and-mortar channel, because it too will suffer from cannibalisation of the physical store sales by the Internet channel. That is, charging much lower prices online will lead customers to transfer from the stores to the Internet channel, whereby the combination reseller cannibalises itself. In addition, the combination reseller risks confusing its custom-ers if it offers different conditions in the different channels. A combination reseller is therefore more likely to offer a relatively fair price and fair conditions for its


17

online offerings, in comparison to the offering it gives at its physical store. Thus, a bricks-and-mortar retailer faces fewer potential conflicts from a combination re-tailer than from a pure Internet retailer (ibid.).

Communication and Coordination Webb (1997) conducted research of four companies that were using multiple mar-keting channels, including the Internet. It must be noted, though, that only one of the researched companies actually used the Internet as a sales channel. The case studies of these companies indicate that communication and coordination exert the most influence on conflict, and that domain similarity and goal incompatibility, in-stead of influencing conflict directly, are types of issues over which conflict-related disagreements occur (Webb, 1997).

This has been a description of the causes of conflict when companies add the Internet marketing channel to existing channels. But what issues are the parties struggling over?

2.1.2 Channel Conflict Issues When Adding the Internet Whereas causes were described as the underlying bases of conflict, issues can be re-garded as the symptoms of causes (Ong et al., 1990). When companies add the Internet marketing channel, existing business partners, e.g., resellers, retailers, dis-tributors, dealers and even the internal sales organisation, will resist (Gilbert & Bacheldor, 2000). And why not? Each faces a possible loss of revenue as well as competition for ownership of customers. For instance, manufacturers can always price lower and, in some cases, even provide more efficient service than distribu-tors, which will cause customers to migrate from distributors to direct channels (Friedman & Furey, 1999). Such competition is often labelled cannibalization. It oc-curs when one marketing channel grows at the expense of another channel, rather than providing incremental revenue for the firm (Webb, 1997). Existing business partners, therefore, will fear that the introduction of the Internet channel, in the end, will lead to the dismantling of their marketing channel (Gilbert & Bacheldor, 2000).

Multiple channels also compete with one another for internal resources such as capital, personnel, products and technology (Webb & Hogan, 2002). For instance, a company’s sales force, resellers and Internet channel may have conflicting inter-ests over issues related to budget allocation, revenue goals, customer assignments, the timing and nature of advertising, and promotional support and pricing (Webb, 2002).

When researching a hosiery company using multiple marketing channels, including the Internet, Webb (1997) found pricing, product availability, customer assignment, pro-motion, resource allocation and the reward system to be channel conflict generators.

______ CHAPTER 2 ______________________________________________________________________________

18

Pricing refers to different price levels and the use of coupons in different channels. Product availability means that certain hot items were not available in all channels, due to supply and demand. Customer assignment refers to the fact that there were crossovers between different channels, e.g., customers shopping both in outlets and department stores. Promotion applies to the inequality of expenditures on promo-tion for each channel. Resource allocation relates to both manufacturing and financial constraints. The researched hosiery company used two separate brands, and some-times both brands needed access to the plants simultaneously, which caused re-source constraints. The reward system caused conflict, since sometimes the objec-tives of the different channels were at odds with what was best for the corporation (ibid.).

According to Webb (1997), there is evidence that, of all these conflict issues, pric-ing is the biggest generator of channel conflict. For instance, the hosiery company’s two different brand-name groups frequently expressed frustration over discounted prices offered by the Internet marketing channel. They believed that such tactics had a negative impact on both their sales and the image of their brands (ibid.).

In seeking to build and manage multiple marketing channels, managers must first acknowledge and communicate the existence of channel conflict. Thereafter, the severity of the conflict should be assessed (Moriarty & Moran, 1990).

2.1.3 Assessing the Seriousness of Channel Conflict When the level of channel conflict is assessed, it is often considered to be a state, although it is also a process consisting of episodes or incidents (Coughlan et al., 2001). Rosenberg and Stern (1971) have developed a simplified model of this process. Even though the authors do not explain all the different parts of the model, they claim it is based on a multidisciplinary review of the conflict and channels literature, and that it emphasizes the major aspects of the conflict process (see Figure 2.2).

Figure 2.2 The Intrachannel Conflict Process (Rosenberg & Stern, 1971, p. 438)

Causes of conflict

Structural and attitudinal factors

Measurable level of conflict

Behavioralreaction(conflictresolutionstrategies)

Outcomes(behavioral and financial)


19

Each conflict situation is interpreted by channel members based on the history of their relationship, i.e. if conflict frequently occurs in the channel relationship, each new conflict situation will be judged by a channel member as if its counterpart is incompetent, operates in bad faith and so on (Coughlan et al., 2001). One of the characteristics of dysfunctional conflict is that, as time goes on, it is difficult to say what the conflict is about (Mack & Snyder, 1957). A positive relationship history, on the contrary, creates a positive future, since a conflict incident will be deem-phasised or interpreted tolerantly (Coughlan et al., 2001).

Four questions should be answered by management in order to get a better under-standing of which channel conflicts are really dangerous: 1) Are the channels really serving the same customers? 2) Do channel members mistakenly believe the channels are com-peting when they actually are benefiting from one another? 3) Is the deteriorating profitability of a complaining channel really the result of another channel’s progress? and 4) Will a chan-nel’s decline necessarily harm the company’s profits? (Bucklin et al., 1997)

Are the channels really serving the same customers? What first may seem to be channel conflict might instead be an opportunity to grow by reaching previously un-served customers (ibid.). In order to find out whether this is the case, managers should de-termine how much revenue the company has in conflict; revenue is in conflict whenever two or more channels simultaneously attempt to sell the same product to the same customer (Moriarty & Moran, 1990). As a rule of thumb, destructive be-haviour will occur when 10–30 percent of revenue is in conflict (ibid.)

Do channel members mistakenly believe the channels are competing when they actually are benefiting from one another? (Bucklin et al., 1997) New channels can seem to be in conflict when they actually are expanding product usage or building brand support. In the beginning, Nike’s opening of its NikeTown store in downtown Chicago was considered a major threat by resellers carrying Nike’s products. But this move probably was the reason for increased sales across all channels (ibid.).

Is the deteriorating profitability of a complaining channel really the result of another channel’s progress? (Bucklin et al., 1997) Selecting the right reseller within a channel is often as important as determining which channels to use. When a channel is not able to compete any longer, poor operations, instead of conflict, may be the real cause. Therefore, when a weak reseller is the only one complaining about conflict, the manufacturer should assess the likelihood that its business will fail, and, if so, how much revenue that would cost. Thereafter, the manufacturer should decide whether to support the threatened reseller. If the decision is not to support the re-seller, management must develop a migration strategy to replace lost profits by us-ing other, more viable, resellers. To prevent becoming dependent on unsuitable resellers, manufacturers should monitor the operations of channel partners and

______ CHAPTER 2 ______________________________________________________________________________

20

work to develop their skills and capabilities. It also may help to switch partners from time to time (ibid.).

Will a channel’s decline necessarily harm the company’s profits? (Bucklin et al., 1997) Sometimes economic shifts or customers’ changed preferences is the real cause of a channel’s decline. If a channel’s deterioration is caused by changes in customer preferences, management must prioritize the new channel without offending the declining channel. It is especially important not to provoke the declining channel if it continues to carry a significant volume. For instance, in the U.S., specialty pet food producers are actively aligning with two category killers, PETsMART and Petco. But they simultaneously support the economics of small pet shops, since the latter still represent 60 percent of specialty food volume (ibid.).

After the above questions are answered, and a conflict is still judged to be dysfunc-tional and a substantial amount of current or future volume passes through the of-fended channel, manufacturers must act to soothe the situation (see Figure 2.3.). As a rule of thumb, a channel in conflict that is not in decline and carries more than 10–15 percent of volume and/or profit needs attention (ibid.).

Figure 2.3 Decision-Making Framework (Bucklin, Thomas-Graham & Webster, 1997, p. 40)

According to Moriarty and Moran (1990), in order to assess the magnitude of the conflict, management must ask these questions: How much of management’s time is devoted to dealing with channel conflict? Where does the channel conflict occur? How do channel members and customers respond to it? Management needs to worry when it de-votes a good deal of time handling customer complaints or trying to settle internal disputes (ibid.). McDonald (1999) states further that the clearest sign that channel conflict is excessive is a higher than normal turnover rate of personnel and business partners. Also, by the time customers become aware of the conflict, it has really become serious.

Importance of threatened channel in terms of current or poten-tial volume or profitability

High Low

High(“fire”)

Low(“smoke”)

Prospect of de-structive conflict

Act to avert or address con-flict

Allow threatened channel to decline

Look for opportunities to re-assure threatened channel and leverage your power

Do nothing


21

Brown and Day (1981) measured channel conflict in a manufacturer-dealership channel. Coughlan et al. (2001) assert, based on their research, that in order to di-agnose the true level of channel conflict, management has to gather four kinds of information: counting up the issues, importance, frequency of disagreement and intensity of dispute.

Step 1. Counting Up the Issues. What are the major relevant issues between two par-ties in their channel relationship? They should be considered even if they are not in dispute at the moment, so as to chart the major aspects of the channel relation-ship. Examples of important issues might be inventory, allocation, and delivery of products (ibid.).

Step 2. Importance. For each issue, it is necessary to assess how important it is to the reseller. This can be done either by using one’s own judgment or by asking the re-seller. For instance, resellers may be asked to indicate on a scale of zero to ten (very unimportant to very important) how important each issue is to their profitability (ibid.).

Step 3. Frequency of Disagreements. For each issue, it is necessary to evaluate and per-haps collect data on how often the parties disagree over this issue. For instance, re-sellers may be asked to recall discussions with the manufacturer over an issue dur-ing the last year and to indicate on a scale of zero to ten (“never” to “always”) how frequently those discussions involved disagreements (ibid.).

Step 4. Intensity of Dispute. For each issue, it is necessary to evaluate and perhaps collect data on how much the two parties differ on the issue. For instance, resellers may indicate on a scale of zero to ten (“not very intense” to “very intense”) how strongly they disagree during a typical discussion with the manufacturer about the issue (ibid.).

These four kinds of information should then be combined to form an index of manifest conflict for each issue:

N

Conflict = Importancei x Frequencyi x Intensityii=1

Adding the product of all the issues collected in Step 1 forms an index of conflict. These estimates can be compared across resellers to assess where the most serious conflict occurs and why. This simple formula gives a good picture of the serious-ness of the conflict, since it allows the diagnostician to pinpoint exactly where and why the two parties are in opposition. In general, the more issues involved in a channel relationship, the greater the risk for conflict. There is no real conflict over any issue, though, if the opinions of the two parties rarely differ (low frequency),

______ CHAPTER 2 ______________________________________________________________________________

22

the issue is regarded to be of no great concern (low importance), or the two parties do not disagree much on the issue (low intensity). That is, if any of these compo-nents is low, the issue is not a real source of conflict, represented by the fact that multiplying by zero creates a product of zero (ibid.).

Magrath and Hardy ((1988) ref. to in Webb, 2002, p. 97) say that conflict is meas-ured empirically by the frequency and intensity of disagreements, weighted by the importance of the issue. Conflict frequency can range from minor disputes and occa-sional disagreements to long, drawn-out, bitter relations. The intensity of channel conflict, on the other hand, can range from easily forgotten minor flare-ups to ma-jor disagreements resulting in terminations or lawsuits. Conflict importance pro-vides a third dimension. Together, the combined frequency, intensity and impor-tance of disagreements makes it possible to classify the level of conflict as either low, medium, or high (ibid.).

Once management decides that the existing channel conflict is potentially danger-ous, it has to decide what to do about it (Bucklin et al., 1997).

2.1.4 Minimising Channel Conflict Several solutions to the channel conflict problem have been suggested in the litera-ture. These approaches can be categorized as institutionalised, pricing, product version, brand name, and compensation.

Institutionalised Approaches to Minimise Channel Conflict Institutionalised approaches refers to mechanisms that defuse disputes before they harden into hostile attitudes (Coughlan et al., 2001). Joint membership in trade associa-tions is one such approach. Another is exchange-of-personnel programs, which means that personnel can be traded for a specified time period; thereby, channel members will be encouraged to develop solutions rather than engage in conflict. Co-optation,the process of absorbing new ideas into the leadership or policy-making structure of an organisation (Stern & Gorman, 1969), is another mechanism for encouraging channel cooperation by, for instance, bringing representatives of different channels together to vent complaints and suggestions (Coughlan et al., 2001). The danger with the two latter approaches, though, is that each side might unintentionally re-veal sensitive information. Third-party mechanisms, i.e. mediation and arbitration, pre-vent conflict from arising or keep conflict within bounds (ibid.). Mediation is a process whereby parties get help to resolve their disputes. A third party, the media-tor, attempts to settle a dispute by, among other things, getting participants to talk to each other, helping them understand their problems, and suggesting possible so-lutions. The mediator typically has a fresh view of the situation and may perceive opportunities that those involved in the conflict cannot (Carver & Vondra, 1994; Coughlan et al., 2001). In arbitration, on the other hand, a third party makes a deci-sion on how to solve the conflict. This solution is final and binding for both parties


23

(Coughlan et al., 2001). Webb (1997), on the other hand, recommends that com-panies use a specified protocol for communication when problems do arise. Strong relational norms also can be a way to prevent channel conflict (ibid.). Heide and John (1992) have identified three dimensions of relational norms that seem to be especially important: flexibility, information exchange and solidarity. Flexibility refers to the willingness to adapt to changing circumstances; information exchanges implies the expectation that parties will proactively provide useful information to channel partners; solidarity means that channel members mutually try to build the relation-ship (ibid.).

Pricing Approaches to Minimise Channel Conflict Another suggested solution to the channel conflict problem is to deliberately set prices higher on the Internet than in retail stores, thereby raising awareness of the brand without really competing with the company’s retailers. This approach is util-ized by the toy manufacturer Mattel, which deliberately sets prices 15 percent higher on the Internet than in retail stores (Webb, 2002). A similar approach is to charge the same prices in all channels but to add shipping and handling fees in the online channel, which in the end causes online customers to pay a higher price. Nike Inc., manufacturer of sneakers and other athletic gear, is an example of this approach, as it lets its customers purchase online, but at full retail price plus ship-ping and handling fees (King, 1999).

Offering the same products in all available marketing channels at the same price is another suggested approach to avoiding channel conflict (Ancarani & Shankar, 2004). Omnitel Vodafone, for instance, offers its prepaid telephone credit cards in its traditional channels as well as on its web site, under the same conditions (Ancarani, 2002). King (1999) and Urban (2004) state further that taking business away from channel partners, such as retailers, by making a product available through the Internet at a lower price causes channel conflict. This conclusion is supported by Steinfield et al. (2002), who in their research of 18 Dutch click-and-mortar businesses found that, almost without exception, management stated the importance of resisting the temptation to bypass existing channels, or else channel conflict be the result. Besides, today’s shoppers want to view and touch products in a store and then purchase them on the Internet (Ashcraft, 2001), or, conversely, re-search products online and thereafter purchase them in a physical store (Forrester, 2004). They also want the possibility of purchasing articles on the Web or through the catalogue, and then return or exchange them at a conventional store. There-fore, offering clothes, for instance, at different prices in a company’s various mar-keting channels irritates customers (Ashcraft, 2001).

Nevertheless, some businesses offer lower prices online than through their other marketing channels. Cortese and Stepanek (1998), for instance, state that airlines mainly use the Internet as a way to sell unfilled seats, as they routinely send out e-

______ CHAPTER 2 ______________________________________________________________________________

24

mail alerts of last-minute fare specials. Other authors (e.g., Baker, Marn & Zawada, 2001; Pitt, Berthon, Watson & Ewing, 2001; Ancarani, 2002; Kung, Monroe & Cox, 2002; Steinfield et al., 2002) suggest that another reason for the price differ-ences is that the airline and banking industries have been able to lower the costs for conventional transactions (i.e. travel agent commissions, booking fees, etc). There-fore, they argue, it is hardly surprising that travel businesses normally offer online booking discounts, and that banks normally offer lower transaction fees and better interest rates for customers who conduct routine banking transactions online. Ex-amples of other firms that charge substantially lower prices for identical products on the Internet are Broderbound Software and Barnes & Noble (Zettelmeyer, 2000). Broderbound Software gives all customers who order on the firm’s web site a US$ 5 discount per order on any of the company’s products. Barnes & Noble, at its physical stores, offers hardcover New York Times bestsellers at a 30 percent dis-count off the publisher’s suggested price, softcover New York Times bestsellers at a 20 percent discount, and other hardcover books at a 10 percent discount; however, other softcover books are not discounted. On the Internet, though, Barnes & No-ble sells all hardcover books at a 30 percent discount and all softcover books at a 20 percent discount (ibid.). Indeed, Brynjolfsson and Smith (2000), in their compari-son of CDs and books sold through the Internet channel and through conventional retailers, found that Internet retailers charge lower prices than conventional retail-ers, whether prices are considered by themselves or include shipping fees, etc. Friberg, Ganslandt and Sandström (2000) also found that books and CDs are sold cheaper on the Internet (on average 15 percent lower); however, they found, in contrast to the previous study, that when a single item is purchased online, the to-tal cost for the item, including shipping fees, is the same as if the item were bought in a conventional store.

Smith et al. (1999) suggest that companies could also avoid channel conflict by sell-ing excess capacity through the Internet via auction. Airlines and hotels, for in-stance, can sell excess capacity in this way. Since the Internet channel then has an associated price uncertainty, customers will regard the products as different from those sold through physical channels (ibid.). Of course, auction pricing will in some cases lead to customers paying higher prices online (Lee, 1998).

Product Version Approaches to Minimise Channel Conflict Multiple products or versions give management the opportunity to target different market segments through different channels (Bucklin et al., 1997; King, 1999; Smith et al., 1999; Coughlan et al., 2001; Ancarani, 2002; Kung et al., 2002; Webb, 2002). Actually, the most obvious source of price dispersion, i.e. different prices charged for the same good at the same time, is product versions, since if the compared products differ in some way, customers will not be surprised if the prices also are different (Smith et al., 1999). By using multiple product versions, compa-nies avoid offering the same product for different prices in different channels (e.g.,


25

Ancarani, 2002). Procter & Gamble does this, since the products it sells through its Internet channel are not available in its traditional channels (Ancarani, 2002). Mat-tel, the toy manufacturer, is another example of a company that does this, since certain hot items are not offered at all on its web site (Webb, 2002). An additional example is Levi Strauss & Co., which became an e-commerce pioneer in 1998 when it launched Levi.com and offered online customers custom-made jeans de-livered to their doorstep (Gilbert & Bacheldor, 2000). In the beginning, the chan-nel partners were invited to sell apparel online as well, but in May 1999, Levi Strauss & Co. stated that it alone would sell its goods online (Collett, 1999). How-ever, just a year later, the company ceased selling online and turned all Internet sales opportunities back to retailers with Web stores, apparently because its rela-tions with its retailers needed to be strengthened (Gilbert & Bacheldor, 2000). However, the product variation does not have to be that substantial; just by intro-ducing minor product modifications, and by using different names and/or num-bers, manufacturers can create the illusion of differentiation (Bucklin et al., 1997). This approach, however, might not work in the Internet environment, since many Internet customers use price and feature comparison web sites instead of relying on model numbers (ibid.).

Another approach is to offer product bundles (Alba et al., 1997). The rationale for product bundles is to offer some complementary products that customers value and that are only available separately elsewhere, if at all. A company could, for instance, offer complementary recipes when selling wine online. Although the wine could be bought elsewhere, the customer will be ensured some satisfaction when serving the wine with a particular meal (ibid.).

Brand Name Approaches to Minimise Channel Conflict The use of a different brand name when offering products online is suggested as a way to minimise channel conflict (Bucklin et al., 1997; Webb, 1997; Coughlan et al., 2001; Steinfield et al., 2002). One of the goals of this approach is to compete with low-price competitors, while at the same time continuing to serve the high-price segment with a high-end product. In most cases, the price difference be-tween the brands is between 30 and 40 percent (Hilleke & Butscher, 1997). A Dutch bicycle dealer, for instance, has developed a separate online brand that is cheaper than the brands sold in its shop, since the business’s costs are lower for online purchases (Steinfield et al., 2002). By using a different brand name, competition between the channels might be reduced (Alba et al., 1997).

However, even if an e-tailer, in order to decrease competition, chooses to introduce its own private label, i.e. a brand that it sells exclusively, such an approach could lose its effec-tiveness due to the use of comparison sites by online customers (ibid.). Moreover, a brand is a search attribute that assures customers a certain level of product quality (Alba et al., 1997). Therefore, the choice of whether or not to use different brands in different

______ CHAPTER 2 ______________________________________________________________________________

26

marketing channels is largely a choice between flexibility and trust (Gulati & Garino, 2000). When extending an existing brand to the Internet, management loses flexibility, since a shared brand name makes it more difficult to use the Inter-net to target a different customer segment. It also makes it difficult to offer differ-ent products at different prices online, since such an offer might leave customers confused and distrustful. On the other hand, if the brand is both recognized and respected, extending it to the Internet gives instant credibility to a web site, a competitive advantage that disappears when using a separate brand name (ibid.).

Compensation Approaches to Minimise Channel ConflictControl of funds offers management a powerful tool to encourage cooperative be-haviour (Tuite, 1972). Indeed, considerable field evidence indicates that economic incentives are the best way to resolve conflicts (Coughlan et al., 2001). Friedman and Furey (1999) go so far as to state that restructuring the assignment of revenue credit is the easiest way to avoid channel conflict. In other words, in order to minimise dysfunctional channel conflict, it is important to fairly compensate offline channels (Gilbert & Bacheldor, 2000).

Channel conflict within a company can be handled by a corporate accountant who allocates costs and revenues, and a human resource manager who administers com-pensation (Coughlan et al., 2001). Channel conflict between interdependent com-panies is not so easy to settle. Suppliers have not paid enough attention to compen-sation mechanisms. The growth of e-commerce will make this a more serious problem, since free-riding will become extremely common. To cope, suppliers and resellers will be forced to develop new ways of doing business, such as adding flat payments (the equivalent of salary), fees for services (the equivalent of an expense account), or overrides, whereby one channel member is automatically compen-sated for sales made by another (the equivalent of a group bonus) (ibid.).

To avoid channel conflict when a channel is declining, manufacturers could con-sider improving the economics of the channel, which often improves its perform-ance at the same time. For instance, a reseller could be offered rebates if it satisfies certain requirements for value-added service, or margins could be adjusted be-tween products to reflect the services offered by distributors (Bucklin et al., 1997).

Some businesses offer the same prices online as their retailers do, with the excep-tion to only offer discounts through the retailer channel, since they want to reward customers who make the effort to come to their stores (King, 1999; Webb, 2002). Estèe Lauder, the cosmetics manufacturer, sells its Clinique products in this way. That is, Estèe Lauder sells directly to online customers, but it does not offer the kinds of promotions that its retailers, such as Macy’s, run several times a year, whereby each purchase greater than US$ 15 results in several free products (King, 1999).


27

When channel members, due to the addition of a new channel, no longer perform all channel functions, they cannot expect to receive a traditional margin or com-mission (Coughlan et al., 2001). Instead, they should be compensated only for the functions they perform (ibid.). Several companies seem to use this approach. When Ducati sold its motorcycles online, its dealers received a commission in spite of their minimal involvement in the sales process (cf. Jelassi & Leenen, 2003). The amount of the commission, though, was almost half the usual one, since the dealers did not assume any marketing or inventory costs (ibid.). Sun Microsystems does not give its customers access to the online Sun Store unless they already have a sales representative, who then receives a commission on these sales (Tsay & Agrawal, 2004). Orders placed at the web site of Ethan Allen Interiors, Inc. are filled at the nearest retail store, which receives 70 percent of the profits in ex-change (ibid.). Former Compaq, finally, when selling personal computers directly to end customers, compensated its resellers with an estimated 6 percent commis-sion for referring small business customers (Girishankar, 1998).

2.2 Research Questions & Frame of Reference n this section, the research questions will be developed and discussed, based on the research purpose and the literature review. Finally, a conceptual frame of reference, which will guide the data collection, will be developed.

2.2.1 Research Questions In chapter one it was concluded that the Internet is a new marketing channel which offers businesses several opportunities, including the possibility to increase sales growth through extended market coverage, and the possibility to adapt to the needs of more customers. When companies add the Internet channel to existing marketing channels, channel conflict seems to be the greatest concern (Gilbert & Bacheldor, 2000). Although not all channel conflict is dangerous, it can soon turn into harmful behaviour between two interdependent parties (Stern & El-Ansary, 1988). Such dangerous conflict is referred to as dysfunctional conflict in the literature (cf. Gattorna, 1978) and will, in the end, lead to declining profitability for the company (e.g., Moriarty & Moran, 1990). Accordingly, the research purpose of this study was stated as follows:

This research purpose is further developed into three research questions.

Research Question One According to the literature (e.g., Smith et al., 1999), channel conflict will occur when the Internet is added to existing marketing channels. Three causes for this

I

To explore and describe managers’ perspectives on conflict in multiple marketing chan-nels, with a focus on the Internet

______ CHAPTER 2 ______________________________________________________________________________

28

conflict have been described by Coughlan et al. (2001): goal incompatibility, domain dissensus and different perceptions of reality. These are the same causes that Stern and Heskett (1969) described in their conceptual paper. However, when testing these causes empirically, Rosenberg and Stern (1971) were not able to attain statistical significance for all the conflict causes for each of their researched channel relation-ships. In addition, these causes were identified long before the Internet was in-vented. When empirically researching the causes of channel conflict in multiple marketing channels, including the Internet, Webb (1997) found indications that other causes of conflict had more influence on conflict than those mentioned above. Therefore, it becomes important to investigate the causes of channel con-flict in multiple marketing channels, including the Internet, from a management perspective. Thus, the following research question:

Research Question TwoSome authors state that channel conflict is not necessarily negative and that it in fact can have positive implications for a business (e.g., Stern & El-Ansary, 1988; Moriarty & Moran, 1990). Furthermore, with few exceptions (cf. Bucklin et al., 1997; McDonald, 1999), little has been written about when channel conflict be-comes dysfunctional or dangerous. Accordingly, to get the full picture of channel conflict in marketing channels, it is also important to examine the positive aspects of channel conflict (Menon, Bharadwaj & Howell, 1996). Therefore, it becomes important to get management’s view on channel conflict: if and/or when it is re-garded as something positive for the company, and if and/or when it is regarded as having dysfunctional consequences.

As for how to measure the severity of the channel conflict, a number of methods have been suggested in the literature, ranging from very structured methods, e.g., finding out which major issues are involved in the conflict, weighted by impor-tance, frequency and intensity (cf. Coughlan et al., 2001), to more informal meth-ods, such as evaluating customers’ and channel members’ responses to the conflict (cf. Moriarty & Moran, 1990). Do managers use any of the methods suggested in the literature to measure channel conflict, or do they utilize some other way to as-sess the seriousness of conflict?

Based on the above discussion, the second research question is as follows:

How can the seriousness of channel conflict in multiple marketing channels, including the Internet, be described?

How can the causes of channel conflict in multiple marketing channels, including the Internet, be described?


29

Research Question ThreeIf management regards the channel conflict as dangerous, it has to decide how to handle it (Bucklin et al., 1997). One conflict-reducing category is that of institu-tionalised approaches, which refers to mechanisms that defuse disputes before they harden into hostile attitudes. The suggested institutionalised approaches range from exchange-of-personnel programs and third-party mechanisms to the building of relational norms. Even though these approaches might help prevent or reduce channel con-flict, they do not come without problems. For instance, as soon as parties from dif-ferent companies get too close together, there is a risk of unintentionally revealing sensitive information (Coughlan et al., 2001). Nevertheless, relational norms can be effective in preventing channel conflict (Heide & John, 1992). So, how can the in-stitutionalised approaches utilized by management be described?

Regarding pricing approaches, different authors suggest different strategies for mini-mising channel conflict when adding the Internet, ranging from identical prices in all marketing channels (e.g., Ancarani, 2002) to higher online prices (e.g., Webb, 2002). Unfortunately, managers do not get much help from the existing literature regard-ing the situations in which the different price levels are best utilized.

Multiple products, versions (Bucklin et al., 1997; Ancarani, 2002), product bundles(Alba et al., 1997) and using a different brand name when offering products online (e.g., Steinfield et al., 2002) also have been suggested as ways to minimise channel conflict, since, among other things, they give management the opportunity to offer different prices through different marketing channels (e.g., Hilleke & Butscher, 1997; Ancarani, 2002). It is suggested that the product variation can consist of only minor modifications. At the same time, it is not clear if these approaches would work in the Internet environment, due to the use of price and feature comparison web sites by customers (cf. Bucklin et al., 1997). In addition, by using a different brand name, the company’s product is deprived of the instant credibility that a well-known, well-regarded brand name offers; this might be another disincentive to applying a different brand name online (Gulati & Garino, 2000).

The development of compensation approaches has been disregarded by companies (Coughlan et al., 2001). The growth of e-commerce, though, makes such devel-opment even more important, since so called free-riding will increase greatly (ibid.). Therefore, it becomes important to see if companies introduce some kind of compensation when adding the Internet, and if so, how this works. Further-more, in the literature it is stated that compensation is the best approach to mini-mising channel conflict (Friedman & Furey, 1999; Coughlan et al., 2001). Is this true when adding the Internet marketing channel, or does management judge other approaches to be better?

______ CHAPTER 2 ______________________________________________________________________________

30

As pointed out by Mohr and Nevin (1990), good communication is necessary to achieve the needed coordination among channel members. However, managers do not get much help from the literature on how this should be accomplished. There-fore, how can the communication approaches that companies use in multiple mar-keting channels be described?

As can be seen from the above reasoning, there are different opinions in the litera-ture regarding which conflict-reducing approaches to use, and how those ap-proaches should be designed to achieve the purpose. So, which approaches does management utilize in order to minimise channel conflict in multiple marketing channels, including the Internet, and how are these approaches designed?

This leads to the following research question:

2.2.2 Developing a Conceptual Frame of Reference The purpose of this section is to build an appropriate frame of reference, based on the literature review and the presented research questions, for the study of channel conflict when adding the Internet. Since no empirical research is possible without clarifying how the researcher is to decide what constitutes evidence of the phe-nomenon to be studied (Potter, 1996), the concepts that will be included in the frame of reference will be defined below, together with a description of how the concepts will be measured.

Causes of Channel Conflict The first part of the frame of reference focuses on the causes of channel conflict in multiple marketing channels, including the Internet. As already stated, what Coughlan et al. (2001) claim are the major causes of channel conflict when adding the Internet are the same as those described in a conceptual article by Stern and Heskett (1969). Several other scholars have also described the causes of channel conflict, although most of their articles were written before the Internet existed, and the majority of those that include the Internet marketing channel are concep-tual. In fact, only one empirical study on the causes of channel conflict in multiple marketing channels, including the Internet, was found in the literature review (cf. Webb, 1997).

Nevertheless, there seems to be considerable agreement in the channel conflict lit-erature that incompatible goals among channel members is a cause of conflict (e.g., Stern & Heskett, 1969; Schmidt & Kochan, 1972; Kochan et al., 1975; Perry & Levine, 1976; Etgar, 1979; Coughlan et al., 2001).

How can the approaches that management utilizes in order to minimise channel conflict in multiple marketing channels, including the Internet, be described?


31

There also seems to be considerable agreement that domains dissensus, such as differ-ent channels competing to reach the same customers, is a cause of conflict (Stern & Heskett, 1969; Reve & Stern, 1979; Magrath & Hardy, 1987; Bucklin et al., 1997; Coughlan et al., 2001). It is true that Webb (1997) discusses coordination as one of the most influential causes of conflict. However, the aspect of coordination is in-cluded in the concept of domain dissensus, since the four critical elements of a channel domain are the population to be served, the territory to be covered, the functions or tasks to be performed, and the technology employed (cf. Stern & El-Ansary, 1992).

When it comes to different perceptions of reality, though, it seems that this cause of conflict should be replaced by inadequate communication. As pointed out by Etgar (1979), there is constant communication in marketing channels. When this com-munication does not work properly, misunderstandings will occur, which will lead to frustration among channel members (ibid.). Accordingly, if there are communi-cation problems within a marketing channel, channel members will perceive the reality differently. Using this viewpoint when analysing the table derived from Rosenberg and Stern’s (1971) empirical research (see Appendix 1), where different causes of conflict are classified, it becomes clear that what is labelled different percep-tions of reality is not really a cause; instead, it is an issue that derives from inadequatecommunication. If looking at the table from this perspective, it is possible to catego-rise some of their miscellaneous issues as caused by inadequate communication, al-though this is impossible to confirm with certainty without access to the original research data. Furthermore, when looking closer at Coughlan et al.’s (2001) discus-sion on differing perceptions of reality as a cause of conflict in an Internet context, it seems apparent that their example really is a matter of inadequate communica-tion. That is, in the survey they refer to, online customers responded that they were able to purchase products online that were not easily accessible to them oth-erwise. Thus, according to Coughlan et al. (2001), the manufacturer was not steal-ing sales from its bricks-and-mortar resellers when selling online, even though the bricks-and-mortar resellers probably thought they were. It should have been possi-ble to pass the results from this survey on to the resellers in order to defuse the conflict situation. In fact, not doing so is nothing but a matter of inadequate com-munication. Moreover, Webb (1997) found indications of communication to be one of the primary causes of channel conflict in multiple marketing channels, in-cluding the Internet.

Accordingly, the three causes of channel conflict that will be included in the frame of reference are goal incompatibility, domain dissensus and inadequate communication.The conceptualisation and the measures that will be used in this research to meas-ure the causes of channel conflict when adding the Internet are presented in Table2.1.

______ CHAPTER 2 ______________________________________________________________________________

32

Table 2.1 Measures of the Causes of Channel Conflict

Conceptual Area

Concept Conceptualisation Measure

Goal Incompati-bility

Conflict that arises be-cause different channel members have incom-patible goals

Assessment of the importance of incompatible goals among chan-nel members as a cause of con-flict (as perceived by the respon-dents)

Causes of ChannelConflict

Domain Dissen-sus

Conflict that arises be-cause one channel member is bypassed by another channel mem-ber

Assessment of the importance of one channel member being by-passed by another channel mem-ber as a cause of conflict (as per-ceived by the respondents)

Inadequate Communication

Conflict that arises be-cause communication is inadequate within the marketing channel

Assessment of the importance of inadequate communication within the marketing channel as a cause of conflict (as perceived by the respondents)

Assessing the Seriousness of Channel Conflict The second part of the frame of reference concerns the assessment of the serious-ness of channel conflict. A number of questions that can be asked in order to assess the seriousness of the conflict have been found in the literature. However, the word seriousness should not be misinterpreted as the equivalent of dysfunctional, since, according to the literature, conflict also can be functional (e.g., Stern & El-Ansary, 1988). The conceptualisation and the measures that will be used in this re-search to capture the seriousness of channel conflict when adding the Internet are presented in Table 2.2.

Moriarty and Moran (1990) suggest that management should assess how much of its time is spent dealing with channel conflict. Since all the issues connected to this question, as described by the authors, have reference to channel members’ and cus-tomers’ responses, this question is included in those categories of questions. The remaining questions can be divided into five different categories: importance of chan-nel, revenue in conflict, conflict issues, response of channel members and response of custom-ers.

Importance of Channel: Will a channel’s decline necessarily harm the company’s profits (Bucklin et al., 1997)?

Revenue in Conflict: Are the channels really serving the same customers? (Moriarty & Moran, 1990; Bucklin et al., 1997)


33

Table 2.2 Measures of the Seriousness of Channel Conflict

Conceptual Area


Importance of Channel

The relative importance of a marketing channel, from a manufacturer perspective

Assessment of the importance of the different marketing channels to the company (as perceived by the respondents)

Revenue in Conflict

The extent to which a com-pany’s different marketing channels are serving the same customers

Assessment of respondents’ views on the company’s differ-ent marketing channels serving the same customers (as per-ceived by the respondents)

Assessing the Seriousnessof ChannelConflict

Conflict Issues The issues involved in the channel conflict

Assessment of the major con-flict issues involved in the channel relationship (as per-ceived by the respondents)

Response of ChannelMembers

Channel members response to the conflict

Assessment of channel mem-bers response to the conflict (as perceived by the respondents)

Response of Customers

Customers’ response to the channel conflict

Assessment of customers’ re-sponses to the channel conflict (as perceived by the respon-dents)

Conflict Issues: What major conflict issues are involved in the channel relationship (Brown & Day, 1981; Coughlan et al., 2001)? The conflict issues that have been identified by previous empirical research of a multiple marketing channel system, including the Internet, are pricing, product availability, customer assignment, promotion, resource allocation and reward system (cf. Webb, 1997). There are indications that pricing is the largest conflict generator in multiple marketing channels (ibid.).

Response of Channel Members: How do channel members respond to the conflict (Moriarty & Moran, 1990; McDonald, 1999)?

Response of Customers: How do customers respond to the channel conflict (Moriarty & Moran, 1990; McDonald, 1999)?

Approaches to Minimise Channel Conflict The third part of the frame of reference concerns different approaches that can be used to minimise channel conflict in multiple marketing channels, including the Internet. Six different categories of conflict-reducing approaches will be included in the frame of reference: institutionalised, pricing, product version, brand name, compen-sation and communication approaches. The conceptualisation and the measures that will be used in this research to capture the approaches used by management to minimise channel conflict when adding the Internet are presented in Table 2.3.

______ CHAPTER 2 ______________________________________________________________________________

34

Table 2.3 Measures of the Conflict Reducing Approaches

Conceptual Area


Institutionalised Approaches

Institutional mechanisms, such as exchange-of-personnel, third-party mechanisms, and relational norms, intended to defuse disagreements before they escalate into conflicts

Description of the institu-tional mechanisms utilized by companies to defuse dis-agreements before they esca-late into conflicts (as de-scribed by the respondents)

ConflictReducingApproaches

PricingApproaches

Price levels used in different marketing channels to minimise channel conflict

Description of the price levels used in different marketing channels to minimise channel conflict (as described by the respondents)

ConflictReducingApproaches,contd.

Product VersionApproaches

Offering a product or prod-uct bundle that is not avail-able through all the com-pany’s marketing channels, to minimise channel conflict

Description of companies’ use of product version approaches (as described by the respon-dents)

Brand Name Approaches

Using a different brand name when offering prod-ucts online to minimise channel conflict

Description of companies’ use of brand name approaches (as described by the respondents)

Compensation Approaches

Compensating offline chan-nels for sales made online, to minimise channel conflict

Description of the compensa-tion approaches used by companies (as described by the respondents)

CommunicationApproaches

Communication mecha-nisms that enhance the ef-fectiveness of communica-tion within multiple mar-keting channels

Description of the communi-cation approaches used by companies (as described by the respondents)

Institutionalised approaches refers to mechanisms, such as exchange-of-personnel, third-party mechanisms and relational norms, intended to defuse disagreements be-fore they escalate into conflicts (Coughlan et al., 2001).

Pricing approaches: Some authors (King, 1999; Webb, 2002) state that the solution to the channel conflict problem is to set prices higher online. Other authors state that products cannot be made available on the Internet at a lower price than what the retailer offers, if a company wants to be able to avoid channel conflict (King, 1999; Urban, 2004); others state that prices should be identical in all marketing channels (Ashcraft, 2001; Ancarani, 2002; Ancarani & Shankar, 2004), whereas others (Baker et al., 2001; Pitt et al., 2001; Ancarani, 2002; Kung et al., 2002; Steinfield et al., 2002) have identified industries (financial services, airlines) where prices are deliberately set lower online. This strategy, though, has not been recommended as a


35

way to avoid channel conflict. Auction pricing, on the other hand, has been sug-gested for this purpose, since the Internet auction channel involves price uncer-tainty, which causes customers to regard the product as different from products sold through the physical channel (Smith et al., 1999). In other words, auction pricing offers companies an opportunity to sell products cheaper online without causing channel conflict (ibid.). At the same time, auction pricing also can lead to customers paying higher prices online (Lee, 1998). Accordingly, three pricing ap-proaches, all of which supposedly can be used to minimise channel conflict, have been identified from the literature review: higher online prices, identical prices in all marketing channels and auction pricing. A summary of these pricing approaches can be found in Table 2.4.

Table 2.4 Conflict Reducing Pricing Approaches when Adding the Internet

Pricing Approach Bibliography Product/Service Offered Higher Online Prices

King (1999); Webb (2002) Toys; Sneakers and other Ath-letic Gear

Identical Prices in all Marketing Channels

King (1999); Ashcraft (2001); Ancarani (2002); Steinfield et al. (2002); Anca-rani & Shankar (2004); Urban (2004)

Prepaid Telephone Credit Cards; Clothes

Auction Pricing Smith et al. (1999) Airlines; Hotels

Product version approaches concerns the fact that an offered product or product bun-dle is not available through all the company’s marketing channels (e.g., Bucklin et al., 1997; King, 1999; Ancarani, 2002; Kung et al., 2002).

Brand name approaches refers to the use of a different brand name when offering products online (e.g., Bucklin et al., 1997; Steinfield et al., 2002).

Compensation approaches refers to compensation of offline channels for sales made online (e.g., Webb, 2002; Jelassi & Leenen, 2003; Tsay & Agrawal, 2004).

Communication approaches refers to communication mechanisms that enhance the ef-fectiveness of communication within multiple marketing channels (Webb, 1997).

Unique Characteristics of the Internet The final ingredient of the frame of reference refers to the unique characteristics of the Internet, i.e. those not shared by other marketing channels. Several authors have claimed that the addition of the Internet marketing channel will cause conflict (e.g., Smith et al., 1999). However, whether the unique characteristics of the Internet contribute to channel conflict, or whether they can be used to minimise channel conflict, is not clear from the literature. Therefore, these aspects also will be taken into account when conducting the empirical research.

______ CHAPTER 2 ______________________________________________________________________________

36

2.2.3 Conceptual Frame of Reference A graphical presentation of the frame of reference helps to define what will, and will not, be studied (Miles & Huberman, 1994). The starting point for the concep-tual frame of reference is Rosenberg and Stern’s (1971) model of the intra-channel conflict process, which emphasizes its major components (see Figure 2.2, p. 18). In their model, however, nothing is mentioned about what the causes of conflict ac-tually are, how the conflict can be assessed and so on. Therefore, the three causes of channel conflict, i.e. goal incompatibility, domain dissensus and inadequate communi-cation, are included in the conceptual frame of reference (see Figure 2.4).

Furthermore, the different approaches that have been suggested in order to assess the seriousness of channel conflict, i.e. importance of channel, revenue in conflict, conflict issues, response of channel members and response of customers, will be added to the con-ceptual frame of reference, as well as the approaches to minimise channel conflict, i.e. institutionalised, pricing, product versions, brand name, compensation and communica-tion.

Additional questions will be asked in an effort to identify other conflict causes, conflict measurement approaches, and conflict management approaches. The com-ponents of the frame of reference labelled other represent these aspects of the re-search.

The Internet is not depicted anywhere in the frame of reference, but since the re-search will be concentrated on channel conflict in multiple marketing channels witha focus on the Internet, all questions deriving from the frame of reference will con-sider the impact of the Internet. However, many different variables are involved when it comes to the performance of a company. Access to financial or behav-ioural outcomes from the conflict process will probably be extremely difficult for an outsider, since such information is very sensitive to the companies. Therefore, the outcome part of the frame of reference will not be directly addressed in this re-search. Moreover, although some arrows are included in the frame of reference, the purpose of this study is not to explain causal relationships.


37

Figure 2.4 Conceptual Frame of Reference

2.3 Methodologyn this section, a description of the methodological considerations taken to an-swer the research questions will be presented. First, the research purpose will be presented, followed by the research approach, the research strategy, data I

Response of Channel Members

Level ofChannel Conflict

Conflict Issues

Assessing the Seriousness of

Channel Conflict


Pricing

Institutionalised

Product Versions

Compensation

Conflict Reducing Approaches

Brand Name

Causes of Channel Conflict

Goal Incompatibility

Domain Dissensus


Outcomes(Behavioural and

Financial)

RQ2

RQ1

RQ3

Importance of Channel

Revenue in Conflict

Other

Other

Communication

Other

______ CHAPTER 2 ______________________________________________________________________________

38

collection, data analysis, and finally a discussion of the trustworthiness of the re-search.

2.3.1 Research Purpose Research traditionally has been defined in terms of its purpose as exploratory, descrip-tive, or explanatory. Exploratory research is used to study a little-understood phe-nomenon, to identify or discover central categories of meaning (Marshall & Rossman, 1999), and/or when the goal is to develop relevant hypotheses and propositions for further research (Marshall & Rossman, 1999; Yin, 2003). Descrip-tive research is used when the purpose is to document and describe the phenome-non in question, whereas explanatory research is used to explain patterns related to the phenomenon of interest and/or to identify possible relationships shaping the phenomenon (Marshall & Rossman, 1999). Since few empirical studies of channel conflict in multiple marketing channels, including the Internet, have been con-ducted (see Section 2.3.2 for a further discussion on previous research), and since propositions for future research were developed, the research purpose was explora-tory. However, since existing theories within the channel conflict area were used in order to describe managers’ perspectives on conflict in multiple marketing chan-nels, including the Internet, the research purpose was also descriptive. In other words, the research purpose was both exploratory and descriptive.

2.3.2 Research Approach The research approach can be either quantitative, qualitative, or a combination of the two (Sullivan, 2001). Quantitative research involves data in the form of numbers, counts and measures, whereas qualitative research mainly involves data in the form of words, pictures, descriptions, or narratives (ibid.). Quantitative studies empha-size the measurement and analysis of causal relationships between variables, not processes (Denzin & Lincoln, 2000). Qualitative research, in contrast, is any type of research that produces findings not arrived at by statistical procedures or other means of quantification (Strauss & Corbin, 1998). It can, for instance, refer to re-search about organisational functioning (ibid.). Which approach to choose depends mainly on two factors: the state of knowledge of the research problem, and the re-searcher’s assessment of the nature of the phenomenon to be studied (Sullivan, 2001).

As for knowledge of the research problem, only five of the empirical studies of channel conflict in marketing channels (see Appendix 2) were conducted, at least to some extent, in a multiple marketing channel setting, including the Internet marketing channel (cf. Webb, 1997; Rohm, 2001; Webb & Hogan, 2002; Coelho, Easing-wood & Coelho, 2003; Jelassi & Leenen, 2003). However, in their research, Coelho et al. (2003) did not separate the Internet from other direct channels (such as direct mail, direct response advertising and telephone); the focus of Jelassi and Leenen’s (2003) study of a single company was not channel conflict per se; in his doctoral thesis, Webb (1997) studied only one company with regard to channel


39

conflict in the Internet context, and the paper by Webb and Hogan (2002) is based on that study; and although Rohm (2001) studied, among other things, channel conflict in the Internet context, only three of the five companies he studied actu-ally sold their products through multiple marketing channels, and only two of these included the Internet marketing channel. It can therefore be concluded that there are few studies and a lack of grounded research examining the influence of the Internet on multiple marketing channels and subsequent channel conflict.

Regarding the phenomenon to be studied, i.e. managers’ perspectives on conflict in multiple marketing channels with a focus on the Internet, it is important to find out what managers think or feel about this – or at least, what they say they think and how they say they feel. Such information is difficult to obtain without qualita-tive research (Strauss & Corbin, 1998).

Since knowledge of the research problem is judged to be limited, and since the na-ture of the research problem is best studied by qualitative research, a qualitative re-search approach was chosen for this study.

2.3.3 Research StrategyFive major research strategies exist: experiment, survey, archival analysis, history and case study (Yin, 2003). All of them have different advantages, and they are chosen depending on the type of research question(s), whether control over behavioural events is needed, and/or whether the focus is on contemporary events (ibid.). The research questions of this study were formulated as ‘how’ questions. When research questions focus mainly on ‘why’ or ‘how’, the best research strategy is most proba-bly case studies, histories, or experiments (Yin, 2003). The research problem of this thesis is contemporary, since the focus was on channel conflict when companies added the Internet marketing channel. The appropriate strategies for contemporary qualitative research are experiments and/or case studies (Yin, 2003). Experiments require control over the behaviour of the subjects under study, whereas case stud-ies are preferred when the relevant behaviour cannot be manipulated (ibid.). Since the researcher had no influence over managers’ perspectives on conflict in multiple marketing channels, and since this study aimed at exploring and describing manag-ers’ perspectives on this phenomenon, the case study research strategy was judged to be most appropriate for this research.

2.3.4 Case Study A case study can be described as a research strategy that focuses on understanding the dynamics in a single setting (Eisenhardt, 1989). The researcher can decide to conduct research on one single case or on multiple cases. A single case can be used, e.g., when it represents the critical case to test a well formulated theory, or when the case is extreme or unique (Yin, 2003). Multiple case studies involve, as the name implies, more than one case. The disadvantage of multiple case studies is that

______ CHAPTER 2 ______________________________________________________________________________

40

they require more time and resources than single case studies. The advantage, though, is that they provide more compelling evidence. The research, therefore, is considered more robust (ibid.). Since this research was not dependent on a single unique case, and since multiple case studies are judged to be more robust, a multiplecase design was chosen for this research.

Selection of Case Study Firms Even if one would like to, one cannot study everyone, everywhere, doing every-thing. The choices of whom to study, where, when, about what and why, place the limits on the conclusions one can draw (Miles & Huberman, 1994). The ones you choose to look at, or talk with, form the sample (Chisnall, 1997). Qualitative researchers generally work with small samples, which tend to be suited to the pur-pose of the study, i.e. samples that are purposively rather than randomly selected (Miles & Huberman, 1994), since the sampling is guided by the need to get access to relevant evidence about the phenomenon (Potter, 1996). This can be summa-rized by two words; relevant, which reflects a validity concern, and access, which re-flects a practical concern (ibid.).

The selection criteria for the sample to be used in this research were as follows: (1) the company must have at least two marketing channels; (2) one of these market-ing channels had to be the Internet; and (3) at least one of the marketing channels had to be owned by another company. In order to find companies with this pro-file, information was retrieved from published material (articles, case studies, etc.) as well as from research on the Internet.

Getting access to companies proved to be a challenge. Hewlett-Packard, Estée Lauder and Tupperware were only a few of the companies that declined to partici-pate in this research. In the end, Professor Jelassi, the author’s supervisor for the first part of this thesis, was able to convince Mattel and Ducati to agree to an inter-view. Scandinavian Airlines, on the other hand, was a positive exception; when the secretary of the responsible person was called, a time for an interview was reserved immediately. However, despite repeated attempts to get access to additional inter-views with other key informants at Scandinavian Airlines, the researcher was granted only one interview. In retrospect, this was not so strange, since the inter-view was conducted two days before the company was to present its new domestic offering, Nya Inrikesflyget.

Given the limited number of cases that usually can be studied, it makes sense to choose cases of polar types (Eisenhardt, 1989). The companies chosen for this re-search represent three diverse industries: Mattel, a toy manufacturer; Ducati, a mo-torcycle manufacturer; and Scandinavian Airlines, an airline business. Even though one of the case-study firms could not be included in this research for relevance rea-sons (see Section 2.3.4), the two remaining companies are of polar types, since one


41

is a manufacturer (Ducati) and the other is a service business (Scandinavian Air-lines).

2.3.5 The Empirical Investigation Data reduction can be defined as an analysis that sharpens, sorts, focuses, discards, and organises data so that conclusions can be drawn and verified (Miles & Huber-man, 1994). The process starts even before any data is collected, since the data col-lection is guided by the researcher’s decisions on research questions, conceptual frame of reference, data collection approaches and so on (ibid.). In the following sections, the data collection methods used in this research will be presented, fol-lowed by a description of how the data were analysed and reported.

Data Gathering Case studies usually combine data collection methods (Eisenhardt, 1989). Four typical methods for qualitative research are participation in the setting, direct observa-tion, in-depth interviewing and analysing documents and material culture (Marshall & Rossman, 1999). In this research, personal interviews, i.e. in-depth interviews con-ducted face-to-face with the respondents, was the main data collection method. Documents were another source of evidence. Examples include web pages, annual reports, newspaper articles, press releases, brochures, a case study of Ducati from 2000 (Leenen & Jelassi, 2005) and so on.

Interviews can take several forms. In case studies, they are most commonly open-ended, which allows the interviewer to ask respondents for facts as well as for their opinions about events (Yin, 2003). Qualitative in-depth interviews are typically more like conversations than formal interviews with predetermined response cate-gories (Marshall & Rossman, 1999). Since the research purpose was exploratory and descriptive, the interview guide was developed with the frame of reference as base, and the questions were open-ended (see Appendix 3).

Personal interviews were conducted at each company with persons who were con-sidered able to answer the questions. The first interview was conducted in Paris on the evening of 7 February 2005 with Mr Christophe de la Fouchardiére, who had worked as Sales Manager at Mattel France until September 2004. Unfortunately, Mattel started to sell directly to end customers on the Internet after Mr de la Fouchardiére left; therefore, he was not able to provide all the information neces-sary for this research. Accordingly, this interview was not included. Due to time and cost constraints, no additional interview with Mattel was set up. The interview still was useful, since it provided a very good test of the interview guide, as well as a deeper understanding of channel conflict.

The next case-study firm was Ducati Motor Holding, which is located in Borgo Panigale, Bologna, Italy. The afternoon of 8 February 2005 was spent with Mr

______ CHAPTER 2 ______________________________________________________________________________

42

Livio Lodi, Curator of the Ducati Museum, who arranged a visit to the factory. This was followed by a visit to the Ducati Museum, where a lot of interesting in-formation on the history of Ducati was provided. On the afternoon of 9 February, interviews were conducted with the Creative Director of Ducati Motor Holding SpA, David M. Gross; the General Manager of www.ducati.com, Patrizia Cianetti; and the Community Director of Ducati Motor Holding SpA, Annalisa Dimonte.

The final respondent was Klavs Pedersen, Director of Product Coordination for Scandinavian Airlines, Sweden. The interview took place on 14 February 2005 at the airline’s Swedish head office, which is located in Arlanda, Stockholm.

Each of the above interviews lasted between one and a half and two hours, and was conducted with the interview guide as the base around which questions cen-tred. Professor Jelassi, the supervisor of the researcher, participated in the inter-views with Mattel and Ducati. The interviews with Mattel and Ducati were con-ducted in English, whereas the interview with Scandinavian Airlines was con-ducted in Swedish. The questions and the order of questions were not followed to the letter; instead, they were used as a starting point for discussion. At the end of each interview, verifications were made to be sure that all subjects included in the interview guide had been covered. Two tape recorders were used for each inter-view to minimise the risk of non-functioning equipment. This also allowed the re-searcher to fully concentrate on the respondents’ answers instead of taking notes, which made it easier to realise when follow-up questions were needed.

Data Analysis and PresentationData reduction continues after the data is gathered, until the final report is com-pleted (Miles & Huberman, 1994). The tape recordings must be processed (ibid.). In this research, they were transcribed word for word, including incomplete sen-tences, laughs, etc. This led to almost forty single-spaced pages of text just for the Ducati case. One important step to cope with the staggering amount of data that results from qualitative research is to adopt within-case analysis (Eisenhardt, 1989). There is, however, no standard format for such analysis (ibid.). In this research, the frame of reference was used to determine what to analyse.

Miles and Huberman (1994) stress the importance of structure when analysing data. Accordingly, the text was structured and reduced under headings related to the research issues defined in the frame of reference. The order of the text that led to the conceptual frame of reference was used as a model for the order of the head-ings, to keep the main theme consistent throughout the sections of this study. Text that was related to the research but did not fit under any of the headings was put in a separate section, to avoid excluding unexpected results. To improve the readabil-ity of the research, figures were used to display the empirical findings.

http://www.ducati.com


43

After the within-case analyses were completed, a cross-case analysis was carried out. This analytical process is supported by Eisenhardt (1989), who states that the wrong conclusions can be drawn from within-case analyses due to information-processing biases. A good cross-case analysis can overcome these problems by look-ing at the data in many different ways, e.g., looking for within-group similarities coupled with inter-group differences (ibid.). Tables were used to help in this proc-ess. The results from the within-case analyses and the cross-case analysis were used to discuss and present the findings and conclusions from this research.

2.3.6 Issues of Validity and Reliability The research design is the logical sequence that connects the research questions posed, the empirical data collected, and the conclusions drawn (Yin, 2003). There-fore, the case study investigator must maximise the quality of the design, i.e. its va-lidity and the reliability (ibid.). Validity refers to how well a specific research method measures what it is supposed to measure (Chisnall, 1997), whereas reliabil-ity refers to how consistent the research is (Potter, 1996). Four tests are commonly used to establish the quality of any empirical social research: construct validity, inter-nal validity, external validity and reliability (Yin, 2003) (see Table 2.5).

Table 2.5 Case Study Tactics for Four Design Tests (Yin, 2003, p. 34)

Tests Case Study Tactic Phase of Research in which Tactic Occurs

Construct Validity - Use multiple sources of evidence - Establish chain of evidence - Have key informants review draft case

study report

Data collection Data collection Composition

Internal Validity - Do pattern-matching - Do explanation-building - Address rival explanations - Use logic models

Data analysis Data analysis Data analysis Data analysis

External Validity - Use theory in single-case studies - Use replication logic in multiple-case

studies

Research design Research design

Reliability - Use case study protocol - Develop case study database

Data collection Data collection

Construct validity refers to establishing correct operational measures of the concepts being studied (Yin, 2003). Conceptualisations and measures for addressing the re-search problem were developed (see Section 2.2.2). For the Ducati case, the main data collection method was personal interviews of three key informants. The Ducati Factory and the Ducati Museum also were visited, and they provided addi-tional information. Other sources of evidence were the Ducati 2004 yearbook, brochures, a case study from 2000 (Leenen & Jelassi, 2005) and the company web site. Follow-up questions were asked by e-mail. After the presentation of the em-pirical evidence was concluded, the draft was evaluated by each respondent. As al-

______ CHAPTER 2 ______________________________________________________________________________

44

ready explained, at Scandinavian Airlines the researcher was granted only one in-terview. The task of finding additional sources of information was therefore even more intense for this case. Additional sources of evidence were newspaper articles, press releases, the SAS 2004 annual report (SAS, 2005), the SAS web site, competi-tors’ web sites, etc. Follow-up questions were asked via e-mail and a telephone in-terview. When the empirical evidence presentation was completed, the respondent was sent a draft to review.

The perspective of this study was the producers’ perspectives. It is, however, possi-ble that the answers would have been totally different had the respondents been managers from the reseller side instead of producers.

Furthermore, it must be noted that conflict seems to be a sensitive subject. It is therefore possible that the respondents played down the seriousness of conflicts with their resellers so as not to jeopardise their relationships with them.

Internal validity is used only for explanatory and causal studies (Yin, 2003). Since this research was exploratory and descriptive, this aspect of validity will not be dis-cussed.

External validity deals with whether research findings can be generalised beyond the immediate study (Yin, 2003). The intention of this research was not to generalise the findings to all companies that bypass their resellers by use of the Internet mar-keting channel. Instead, the findings were used to generalise to theory. This kind of generalisation is referred to as analytical generalisation (Yin, 2003). However, this generalisation cannot take place unless the findings are replicated in another setting, so-called replication logic (ibid.). Therefore, in order to increase the external validity, two case studies were conducted with companies representing different types of industries (manufacturing company vs. service provider).

Reliability refers to whether the findings and conclusions would be the same if an-other researcher repeated the same study (Yin, 2003). Reliability is also about minimising errors and biases (ibid.). This research was discussed and evaluated sev-eral times by colleagues at internal research seminars at Luleå University of Tech-nology. In addition, throughout the research process, Professor Jelassi, supervisor of the producer study, revised the chapters and the interview guide. Three separate interviews with key informants were conducted at Ducati, which ensured that the evidence was accurate and consistent. Even though only one interview was con-ducted at Scandinavian Airlines, the respondent was deeply involved in the devel-opment of the company’s distribution strategy. During all the interviews, two tape recorders were used, which made taking notes unnecessary; this allowed the inter-viewer to focus on understanding the respondents’ answers and asking follow-up questions. This approach also provided tape recordings of the entire interviews; be-


45

cause the tapes did not end simultaneously, there was no missing data. Each of the interviews was written down word for word. Afterwards, the backup tape was lis-tened to and compared to the written text, to make sure that there were no obvi-ous misinterpretations of the respondents’ answers. During the data analysis stage, respondents were contacted by e-mail or telephone to clarify or elaborate certain issues. In order to increase the possibility of repeating this research, the goal of this thesis was to present how each step of the research process was carried out. All ma-terial related to the case studies is organised, documented and stored in a manner that enables others to retrieve them efficiently at some later date.

2.4 Empirical Evidence n this section, the empirical data collected from the two case studies will be presented. For each case, a background of the company will be presented. Thereafter, in accordance with the frame of reference, the findings related to

the causes of conflict, the approaches used to assess the seriousness of channel con-flict, and the approaches used to minimise channel conflict will be presented. Each case presentation will end with a discussion of the unique characteristics of the Internet, as well as a section on non-Internet-related channel conflict.

2.4.1 Case One: Ducati Motor Holding SpA Ducati Motor Holding SpA, hereafter referred to as Ducati, is an Italian manufac-turer of premium priced, high performance motorcycles. The company, which is located in Borgo Panigale, Bologna, Italy, was founded in 1926 by the Ducati brothers. At that time, the company produced industrial components for the grow-ing field of radio transmissions. In 1946, the Cucciolo was introduced, i.e. a small auxiliary motor for bicycles. Soon, the Cucciolo became a true miniature motor-cycle. In 1948, the company was taken over by the Italian government. By the mid 1950s, Ducati’s production included several motorcycle models. Since then, Ducati motorcycles have dominated the World Superbike Championships. In 1983, the company was purchased by the Cagiva Group. Under the new manage-ment, Ducati expanded its share of the motorcycle market, introduced new motor-cycle models, and intensified its commitment to racing. In 1995, despite product innovation and racing successes, Ducati entered a deep financial crisis. The com-pany was taken over by its present owner, the Texas Pacific Group, an American investment firm that brought in money and a team of international managers.

Shortly after the takeover, Ducati posted increasing sales and profits. The produc-tion, which was about 12,000 units/year in 1996, increased to about 40,000 units in 2000; at the same time the number of employees only increased from 600 to 1,200. The year 2000 was the peak of growth for the motorcycle market in gen-eral, since, among other things, it was the last year of certain tax incentives and subsidies for updating motorcycles for Europeans. It also was the year when the

I

______ CHAPTER 2 ______________________________________________________________________________

46

stock exchange crashed, which brought changes in the whole world economy. Since then, Ducati has been relatively stable in terms of volume, although the company has lost volume, so that last year it sold around 37,000 motorcycles. In other words, the company hit a ceiling in terms of sales and is therefore developing and launching new products that will make it possible to continue to grow. The factory has the capacity to produce around 70,000 motorcycles a year; it is just a question of adding shifts, or eventually building new production lines. But since the US accounts for around 16 percent of the company’s sales, the dramatic decline of the dollar has hurt the company’s profits. The company lost 200 of its employ-ees during 2004 and is planning to let go another 50 people during 2005.

EU regulations, together with environmental and noise regulations, have made it difficult to work with local suppliers of motorcycle components. Accordingly, Ducati purchases most of these components from major European suppliers in Germany, Sweden and Italy. Even so, all of Ducati’s motorcycles are produced in Bologna, Italy. Ducati’s major competitors are the Japanese motorcycle manufac-turers Honda and Yamaha, and sometimes Suzuki and Kawasaki.

“The reality is that anybody who manufactures a motorcycle or a scooter is competing with the Japanese, because the Japanese offer a direct competitor in everything that you can do. We have Superbikes, they have Superbikes. We have Naked bikes, they have Naked bikes. What we really are trying to do is to carve out the niches that our competitors don’t exactly have, because we do not want to go against their core strength. Therefore, we launched the Multistrada and the Sport Classic families. We’re good at creating excitement; we have European, particularly Italian heritage; we have a made in Italy product; we have the ability as a smaller niche-manufacturer to do things that other companies can’t do. For us, if we sell 5,000 or 6,000 units of a product family, that’s a big success. The Japanese look at num-bers that are hundreds of thousands, or tens of thousands at a minimum. So we have a lower breakeven point which allows us to do unique things. So, to the extent that we can create a world of excitement, a world of activities around our product, to justify the brand, to justify the premium price, this is what we have to do.”

– David M. Gross, Creative Director

Even though Ducati is a small company, it has the ability to develop competitive secure engines, both on a race level and on a street motorcycle level, while most competitors have to buy engines from other companies. The sound of Ducati’s en-gine is also very characteristic. Therefore, Ducati regards the engine as one of its core competencies. The other core competency is the brand. Ducati has trans-formed the company, and thereby the brand, from an industrial company into an entertainment company. Although the focus is on motorcycles, management claims that the company is also in the business of providing entertainment for its customers (Ducatisti, enthusiasts or fans, as Ducati refers to them).


47

Since 1996, Ducati has also built a licensing business that generates about € 2.2 million pure profit a year. Furthermore, the company has an apparel and accesso-ries business that brings in around € 30 million a year. In Italy, Ducati sells its products to its retail dealers directly from Bologna. In its other major markets, i.e. North America, UK, Germany, France, Japan, Benelux and Sweden, Ducati has wholly-owned subsidiaries. The Italian dealers and the wholly-owned subsidiaries are responsible for about 90 percent of the company’s sales. All other countries are handled by distributors, the major ones being South Africa, Australia, Switzerland and Spain. Thus, the sales are concentrated in Western Europe, North America, Japan and Australia. Ducati’s marketing channel system is illustrated in Figure 2.5.

Figure 2.5 Ducati’s Marketing Channel System

Since 1996, management has cut the number of dealers. Today, the global network consists of around 800 sales locations, about 200 of which are Ducati stores that sell most of the company’s volume. In Italy, for instance, the number of stores has been cut from 200 to 50. Most of those 50 stores are single-brand Ducati stores which sell around 13,000 units/year. In the US, Ducati has 200 sales locations, but these are multi-line dealerships which sell around 6,000 units/year. Although the company would like to have more single-brand stores, it cannot afford to support a store that is not selling enough on its own. Ducati also has a shop at its main office, and it has recently opened one at the airport in Bologna, as well as an outlet just outside the main office. In the main office and airport shops, the company does not sell motorcycles, but rather t-shirts, leather jackets, etc. In the outlet, on the other hand, besides apparel and accessories, the company sells motorcycles that have been used during Ducati events; the company uses around 250 motorcycles each year that they need to sell at the end of the year.

End Customers

DucatiInternet

Dealer 1 (e.g., Ibiza)

Dealer 2 (e.g., Stock-

holm)

Dealer 1(e.g., Luleå)

Dealer 2(e.g.,

Barcelona)

Wholly-Owned Subsidiaries (e.g., Sweden)

Distributors (e.g., Spain)

OutletMain Office & Airport Shops

ItalianDealers

______ CHAPTER 2 ______________________________________________________________________________

48

Ducati does not sell its products only through physical marketing channels. On 1 January 2000, at 00:01 a.m. GMT5, Ducati launched its MH900e, a special edition motorcycle, exclusively on the Internet, i.e. customers could only purchase the motorcycle online. The price of the motorcycle was set at € 15,000. After just 31 minutes, the first year’s production was sold out.

Figure 2.6 The MH900e (Ducati, 2005)

Shortly thereafter, Ducati increased its strategic commitment to the Internet with the establishment of www.ducati.com, an independent partner enterprise, which was located outside the main office. Besides the MH9000e, three other motorcycle models have been sold exclusively on the Internet, yielding a total of 3,000 bikes: the 996R, which sold for € 26,000, the Fogarty, which cost € 18,000, and the 999R, which sold for € 30,000. Ducati also sold a very high priced bicycle on the Internet, i.e. it was developed by the bicycle manufacturer Bianchi, but it was sold under Ducati’s brand name. Management did this as an experiment, because it wanted to understand the limits of the Ducati’s brand, but the selling of the bicycle was not very successful. Ducati also sells some of its apparel and accessories on the Internet. Apparel that is manufactured by Ducati is sold online. Some products de-veloped by Ducati’s licensing companies also are sold online, but not necessarily through the dealer network.

Accessories are divided into three categories: technical accessories, i.e. engine kits, wheels, and other important parts of the bike; light accessories, e.g., carbon fibre ac-cessories, aesthetic accessories; and technical gear, i.e. leather jackets and helmets. The technical accessories are not available online for (among other) legal reasons, since the company cannot sell parts directly to end customers if there is a danger of their being incorrectly assembled on the motorcycle. Such parts require the assistance of the dealer. Light accessories, on the other hand, are available online, since they do not involve the safety of the motorcycle. Technical gear, such as leather race suits,are not sold online, since, for example, it is difficult for customers to decide what size of leather suit or a helmet they need without trying it on.

5 GMT refers to Greenwich Mean Time.



49

www.ducati.com, hereafter called ducati.com, focuses mainly on community-related activities, i.e. online interaction with Ducati’s end customers. All kinds of interactions are involved, ranging from simply asking customers about their per-sonal lives to letting them publish articles, pictures, etc., on the site. Ducati.com has conducted online surveys, such as the Censimento Ducati; the company asked its online visitors forty-five questions and received 18,000 answers. Typically, online surveys are used to help confirm product decisions or modify them. In 2004, Ducati’s web site had more than eight million unique visitors, and the company had 140,000 active registered users. That same year, the renowned financial paper Sole 24 ORE praised ducati.com for having the best automotive site. Recently, ducati.com was integrated with Ducati as an internal department. It now belongs to the community section, which includes event management, company sponsor-ships, customer service, advertising, a creative centre, the Internet channel and trade shows. All personnel now work in a large, open working area at Ducati’s main office.

Causes of Channel Conflict When Ducati developed the MH900e, its sales department was not sure exactly what to do with it, since it thought that the motorcycle was a very controversial and exciting product. Accordingly, Ducati decided to sell the motorcycle on the Internet to get direct feedback from its customers, to create entertainment for its fans, to jumpstart the company’s Internet program, and to experiment outside the traditional network. It became clear to the company that the market for Ducati was very large and that the traditional network was very limited. Therefore, Ducati focused very much on the Internet; it thought that everything could be sold on the Internet, and that this was a fast way to sell products. But when Ducati sold motor-cycles online, channel conflict occurred with its dealers. The dealers wanted to be the ones to unveil new motorcycle models, and they wanted to be the contact point with Ducati’s customers.

“Think about any business that works this way. Porsche launches a new car. They do not sell reservations online. They send the car out to the dealers, and the dealers take deposits, and part of the excitement is that you get the brochure to your house, but you go to the dealer to see the car. So we need to make sure that in managing conflict in the channel, that the dealer is still the place. They do not make as much money on the sale of the bike as they do on the sales of the parts, the accessories, the experiences, the ride outs, service, so you want the customer in that environment.”


The customers who purchased motorcycles online had pretty much the same cus-tomer profile as the ones who purchased them at the dealership. At least during the sales of the MH900e, however, there were some differences: one marketing man-ager purchased a motorcycle and now keeps it in his office as a piece of art. Others


______ CHAPTER 2 ______________________________________________________________________________

50

saw the motorcycle as an investment. But most were just ordinary Ducati custom-ers who wanted to purchase a motorcycle, and since it was sold on the Internet they bought it there; if it would have been sold through the dealer network, they would have gone to a dealer. According to management, the channel conflict that occurred with its dealers was caused by inadequate communication.

“It was a problem of communication management. I am strongly convinced that 98 percent of problems with the dealer network can be easily solved with good communication and clear communication, and consistent communication. When the communication is not working properly, the dealer is reacting negatively, but this applies to everything: to the online sales, to the internal communication, to the availability of motorcycles, information that stops at one level, misunderstandings, or the time gap, that the customer has information before the dealer.”

– Patrizia Cianetti, General Manager, ducati.com

“The only thing that I see that causes conflict is when the communication is not homogene-ous, or when it is redundant. But this is not a problem of the channels; it is a problem caused by people who are not using the channels in a good way.”

– Annalisa Dimonte, Community Director

Assessing the Seriousness of Channel Conflict According to management, channel conflict occurs at different levels at Ducati: in-ternally, among different departments within Ducati, and externally, between Ducati and its resellers, between Ducati and its customers, and between the resel-lers and the customers. An example of channel conflict among different depart-ments within Ducati is when the press office comes out with its own press release about product positioning while, at the same time, the Internet comes out with its own product positioning. When there is no coordination between the different departments at Ducati, inter-functional disagreements arise.

Since management wants to entertain Ducati’s customers, it claims that it is trying to create conflict, to do things that are radical, exciting and controversial, so that the customers have a reason to believe in Ducati. The actions are all aimed at building a dynamic Ducati brand, since the Japanese are particularly adept at pro-ducing cheaper, more reliable motorcycles. In other words, Ducati needs to sustain its brand, to justify the premium price.

Management does not think that the Internet has caused more conflict with its re-sellers. Furthermore, it does not view the channel conflict that resulted from selling motorcycles online as dangerous to the company.

“I do not think so. We had some problems, some bad reactions, but I do not think it was a real critical problem for the company.”



51

But, as already stated, the dealers got angry and claimed that they would have been able to sell the MH900e at the same rate if the company had just given it to them. Although dealers were upset, none of them ended their relationships with Ducati due to the online sales of motorcycles, and the disagreements did not disturb Ducati’s customers much, according to management.

Ducati has not used any structured approaches to find out about the conflict with its dealers. Management spent a lot of time talking to dealers when they sold mo-torcycles online, as well as a lot of time talking internally, but they did no formal surveys, etc. Since Ducati thinks that one of the major obstacles to the company’s growth is the weakness of its distribution network, management decided that it did not want to do anything that would weaken the network any further. So for the time being, the online sales of motorcycles have almost stopped. The company is planning to sell motorcycles online occasionally, but only if it is a limited edition and a highly desirable, exciting product. But for the time being, Ducati is not planning to produce a very limited edition of motorcycles.

“We have not sold any bikes on the Internet for two years now, because it is very compli-cated, and because we did it several times so there was no longer any novelty, no news. Ithink that we will only sell motorcycles online if it is an element of distinction, an element to enrich the company in some way, to give value to our brand. Otherwise, if there is no nov-elty, something new, something that makes us somewhat distinguish from others, enrich our brand, I think we will not do it.”


“I strongly believe in the Internet channel. Now we are starting the regular work, not the ex-ceptional work that we saw in the past. We do not any longer look at the Internet channel as an opportunity to change the business, to do incredible stuff. The Internet is just another channel that the company has to manage, because if we do not manage it the dealers will start to sell online, something that will bring a lot of problems regarding boundaries of territories. Therefore, you can either decide to wait and see what will happen, or you can decide to go online and define a strategy of the Internet channel. There is no way to avoid Internet man-agement.”


“It is definitely clear that our Internet effort is continuing to grow in terms of apparel and ac-cessories. So it is very clear for me that we are going to sell more and more on the Internet over time, as long as we have a sense of the importance of the dealer in that mix. We want to make sure that the dealer is part of the success.”


______ CHAPTER 2 ______________________________________________________________________________

52

Approaches to Minimise Channel Conflict Ducati's management integrated ducati.com’s with Ducati because it wanted the message to be the same through all of its channels. To achieve this, management felt the need to gather all community personnel in the same building, at the same location. A base text that describes the positioning of a certain product is devel-oped by the Creative Director and another person. This text is then distributed through all the marketing channels, including the Internet, subsidiaries, distribu-tors, press communications and internal communications. This is done to increase the consistency of communication, whichever channel is used.

To improve communication with its resellers, Ducati has recently invested in a communication system, a basic extranet. The extranet allows Ducati to communi-cate with its subsidiaries and distributors, and these in turn can communicate with the dealers. One reason for the new communication system is to reduce conflict within the distribution network, but also to speed up and facilitate communication throughout the distribution network.

“It is so easy for us to publish something online. The customer sees this information, goes to the dealer, but since the dealer did not go to our web site they do not have the same informa-tion. With the use of the extranet, the dealer will get the information before it is published on our web site, so that the dealer can take as good care of the customer as possible.”


As soon as Ducati started to sell motorcycles on the Internet, management tried to involve the dealers in the sales. One reason was to minimise channel conflict, but also, since motorcycles are produced according to country specific regulations re-garding lights, cc, etc., Ducati needed to know where the motorcycles should be delivered. Furthermore, once a motorcycle is delivered, someone needs to give it oil and gasoline, address customer expectations and so on.

“To my knowledge, there is no way to sell a bike without the support of a dealer. You can not ship a bike with gasoline or oil inside. We need someone trained, who knows very much about the bikes, to refer the bike to the customer.”


“When you are selling online, it means that all channels have to cooperate, including the dealer. Because even though you purchase online, you go to the dealer to receive the physical bike. Customers see something on the Internet, and their expectations can be different from what they get in reality, so you have to be ready to manage the expectations of the customer in a positive and good way. This is very critical.”


“It is one thing to sell a motorcycle, but you need the customer allied with the dealer.”– David M. Gross, Creative Designer


53

The customer had to choose a reference dealer, which had to be someone within the country of delivery. Then the customer had to pay a reservation fee online, by credit card. Due to customers’ credit card limits, the reservation fee was not linked to the price of the motorcycle; for instance, 10 percent of the purchase price is a lot of money (e.g., the 999R sold for € 30,000) and not many people have such a high credit card limit. The information that a person had paid the reservation fee was submitted electronically to the subsidiary or the distributor, who thereafter transferred the information to the selected dealer. If the dealer refused to authorise the sale of the motorcycle, the subsidiary or distributor transferred the request to another dealer, because Ducati cannot force a dealer to accept a transaction. When a dealer accepted the transaction, which he or she could do by simply clicking yes to the question, the customer received this information, together with a statement that the dealer would soon contact the customer.

The dealer received a commission on the sale, but this commission was lower than it would have been had the customer bought a motorcycle directly from the dealer, i.e. through the physical channel, since the dealer had no marketing costs, no inventory costs or risk, no pre-sale service costs, etc., for the online sale.

As already stated, Ducati’s web site has more than eight million unique visitors each year. Ducati, therefore, has increased its online offerings of apparel and acces-sories, in order to sell more to these visitors. The routine is the same for these online sales of accessories or apparel as for motorcycles, i.e. customers have to choose a reference dealer, who then receives a commission on the sales. The ap-parel or accessories are delivered directly to the end customer, without going through the dealership. Since the dealer is not involved at all in these sales, Ducati refers to this commission as ‘money for nothing’. But dealers get more than money as compensation for the sales; they also get access to information about the custom-ers. Ducati encourages the dealers to contact its online customers, to invite them to events, etc.

The online prices for apparel and accessories are the same as the suggested dealer prices, since Ducati applies the MSRP, the Manufacturer’s Suggested Retail Price. For instance, if the MSRP is € 100, this is the price that the dealer will show to the customer, but this price includes shipment costs and duty. If the customer pur-chases the same product online, the base price will still be € 100, but then the shipment costs and duties will be added. So, in the end, customers pay more for the same product when purchasing it online. Ducati does this to show the dealers that they are not trying to compete with them, since they view the dealers as one of the company’s biggest assets. Ducati wants the dealers to be happy and rich, to grow, to open new stores, etc. But Ducati also feels the need to exploit the Inter-net channel; therefore, it gives the dealers all information on the online sales, to make sure that the situation is transparent. More specifically, Ducati transfers all

______ CHAPTER 2 ______________________________________________________________________________

54

customer data to the reference dealer, and on a monthly basis it publishes all the results from the Internet sales.

Since the products in reality cost more on the Internet than at the dealers, there is no channel conflict about pricing. The problem is that with this pricing approach it is not interesting for customers to purchase products online, unless they live far away from a dealer. As far as money is concerned, the Internet is not currently an important channel for Ducati.

“We sell 40,000 bikes a year, but there are eight million people searching for us online. There is something that we can do. There is something that we have to do. What we have said is that we are looking for the right way to exploit this potential. We do not know if this is the right way. Probably not. Meanwhile we are trying to understand what opportunities there are to merge the processes of selling offline through dealers and selling online directly to customers.”


Unique Characteristics of the InternetIn 2000, Ducati’s management thought that the Internet would offer the opportu-nity to speed up the company’s procedures, since the Internet can transfer informa-tion rapidly to customers. However, it proved to be the offline processes that brought the maximum speed to the company. In other words, the company’s pro-cedures defined the speed of the Internet channel, not the other way around. Man-agement also thinks that, from the beginning, it should have concentrated more on offline processes, especially logistics, invoicing, etc. This is because, apart from the online part of the purchase, where the customer uses the catalogue, the “shopping cart” and so on, it is a standard sales process. Ducati spent a lot of time and energy trying to change the process in order to have proper management of end customer sales, because when Ducati sells to its dealers it sells to other companies, and this involves different kinds of orders and parcel management. It took management a great deal of time and energy to understand how to merge the processes and then to do the merging. So management thinks that it from the start should have fo-cused more on defining the offline processes, because then it would have spent less time, less money and less energy on developing the e-commerce.

Furthermore, in the beginning, Ducati’s web site was extremely sprawling. Today, Ducati emphasises focus and increased quality, which management wishes it had done from the beginning. Now the site is closely connected with sales goals, be-cause even though Ducati has managed to build a strong brand name, the brand name has to sell motorcycles. That is, each year a goal, i.e. a percentage of growth in web visitors and registered users, is established. Ducati also has set goals related to brand management, including style of content, richness of graphics, pictures, etc. The web site is organised around the concept of totem, meaning that different


55

customers have different entry points to the company. Some “Ducatisti” like Ducati because of the design, whereas others like it because of the company’s his-tory. Therefore, Ducati’s totems are the design, the history of the company, the engine, the sound of the engine, etc. The company wants to get its customers to enter through all these totems, or gates, so they will understand and appreciate the whole world of Ducati.

Ducati also discussed the online solution of Harley-Davidson. In the US, Harley-Davidson has developed a kind of online mall, i.e. every dealership has its own “shop” on the Harley-Davidson web site. Customers have to select a dealer, which means that when they enter the catalogue, they look at the product prices the dealer has selected. Customers purchase, i.e. receive the invoice, from the selected dealer. The product is shipped from the dealer if it is in stock; otherwise, it is shipped from the central warehouse. Ducati thinks that this online model is very interesting but difficult for it to utilize it, since Ducati has relatively few dealers, and these are located in many different countries with different laws and rules. For instance, in some countries it is illegal to offer products to end customers at a price below MSRP.

The Internet gives Ducati information about customer preferences, since, for in-stance, customers in chat rooms discuss Ducati and its products. Even though it is not possible to listen to everything, the Internet gives Ducati a strong sense of what people think. Management also conducts many online surveys to get information about new products and new ideas, as well as to take the pulse of what is happen-ing in the dealer network. As already stated, customers are very willing to partici-pate in these surveys. Ducati uses Siebel’s CRM platform as support in this work. Since Ducati is a small company, management regards the Internet as crucial to everything, because it is the most efficient, economical way for it to communicate with its customers, or potential customers, at a very low cost. The Internet also gives Ducati the opportunity to serve customers who are located far away from a dealer.

Management used to talk about going public with ducati.com as a separate com-pany, but no one talks about that today. However, management will continue to develop the basic premise, which is that the Internet will continue to provide in-formation for management, bring the company closer to its customers, and be the first place where people see the motorcycles. The idea is that the Internet market-ing channel always will be richer, more dynamic, more exciting and more relevant to the customer experience, than will all the other marketing channels.

“For me, the physical and virtual worlds are complementary, and help each other to enrich the company, to enrich our activities and the value of the company.”


______ CHAPTER 2 ______________________________________________________________________________

56

Non-Internet-Related Channel Conflict As already stated, Ducati sells used motorcycles through its own outlet, just outside the main office. In order not to get into conflict with its resellers, Ducati shows new motorcycles at this outlet, but it does not sell them. If customers are interested in purchasing new motorcycles, Ducati refers them to the dealers. In addition, when customers purchase used motorcycles, they are offered the option to pick up the motorcycle at any Ducati dealer. The transportation of the motorcycle to the dealer is offered to the customer or the dealer without cost. The reason is that Ducati wants to link customers to their dealers, because the story of the customer starts when purchasing the motorcycle, but the maintenance period, the repur-chase, the handling of customer expectations, and everything else, must be contin-ued in Ducati’s dealer network. Management thinks that this exchange with the customer is positive, because the customer gets to see a link between the main of-fice and the dealer. It also is a way for management to better understand the taste or attitude of the customer. Because the outlet personnel are employed by Ducati, the company gets direct access to customers’ reactions without the intermediation of other channels.

Moreover, Ducati has fan clubs all around the world. Harley Davidson, the Ameri-can motorcycle manufacturer, has forced its clubs to be attached to dealers. Ducati decided not to do this, since it does not want to act like a policeman and it likes the idea of a big, heterogeneous ‘tribe’. The dealers want the clubs, which consist of Ducati customers, to be attached to their businesses, since it is the dealers who resolve customer issues such as service problems, warranty problems and connec-tions with the company. Thus, having clubs outside the dealer network has created conflict.

Other conflict issues also arose. For instance, dealers became angry when they re-ceived information that a motorcycle model was still available for sale, when in re-ality it had sold out. Furthermore, customers sometimes got access to information before the dealers did. This made dealers angry, since they were not able to man-age the customer properly or exploit the fact that the customer decided to spend time at their dealership.

Ducati also refers to conflict with its customers. Since the core customers are more or less mechanics who can take the engine apart and put it back together, they of-ten know more about the motorcycles than do Ducati’s marketers, sales people and executives; this can create conflict. In addition, Ducati mentioned the conflict that occurs among different customer segments. Ducati’s customers have different entry points to the brand. Some love racing, whereas others love Ducati’s design, etc. Management thinks that these diverse interests create conflict. Although manage-ment thinks that this conflict is very positive for Ducati, and even celebrates it as a sign of customer engagement, it wants to keep it from getting too intense. There-


57

fore, it has to be very clear about communications. The previously described base text (see Section 2.4.1, p. 52), distributed though all marketing channels, is used to achieve this objective.

2.4.2 Case Two: Scandinavian Airlines Scandinavian Airlines, hereafter referred to as SAS, was formed in 1946 through the merger of three national carriers in Denmark, Norway and Sweden. Today, SAS is the fourth largest European airline group, and it is the Nordic region’s larg-est listed airline and travel group in terms of number of passengers and operating revenue. SAS has around 1,445 daily departures to 146 destinations in Europe, North America and Asia. The company is a member of Star Alliance, the world’s largest airline alliance. Together with its Star Alliance partners, SAS offers more than 770 destinations worldwide. In addition to its involvement in Star Alliance, SAS since 1996 has had a strategic alliance with Lufthansa, coordinating routes, timetables and capacity. The SAS Group also includes companies that support air-line operations, as well as a business area that operates hotels.

Most large European airlines, including SAS, were once at least half-owned by the government. But during the 1990s, the European Union forced a deregulation of the European airline industry. Since this deregulation, the industry has suffered from overcapacity, and there have been dozens of bankruptcies in Europe. The terrorist attacks on 11 September 2001 were a major shock that further weakened the industry. The outbreak of SARS in 2003 was yet another bad stroke. In addi-tion, particularly in the northern skies, strong competition has led to a collapse in prices. As a result of all these factors, European airlines have not been profitable since 1998, and the last few years have been rather problematic financially for SAS.

Although there can be advantages in being a large company, such as the ability to exploit synergies, the formerly totally integrated SAS faced increased costs and thus did not achieve competitive advantage. Therefore, to reduce complexity and to clarify responsibility for its earnings, SAS established in October 2004 three wholly-owned subsidiaries in Denmark, Norway and Sweden, plus SAS Interna-tional. The three subsidiaries focus on their respective domestic markets, or opera-tions in, to and from their countries. SAS International is run as a separate business. It is responsible for intercontinental airline operations to the US and Asia, and for sales outside Scandinavia. A recovery program has also been implemented, which reduced the cost level by more than SEK 10 billion6 in 2004 compared to 2003, due, among other things, to a reduction by 2,130 employees in 2004.

6 On 29 March 29 2005, SEK 1,000 amounted to € 108 or US$ 139.60.

______ CHAPTER 2 ______________________________________________________________________________

58

SAS’s competitors are airlines in Scandinavia and the rest of Europe, the major ones being Air France/KLM, British Airways, Finnair, FlyNordic and Flyme. Other competitors include Maersk Air, Sterling, Malmö Aviation and Iberia.

Management distinguishes four different marketing channels: travel agencies and company owned channels, which in turn can be classified as offline or online. The understanding of the Internet as a sales channel has increased a great deal during re-cent years, and several travel agencies, just as SAS, have spent a lot of money to develop online sales systems to increase efficiency. The major travel agencies have their own web sites, which serve as a complement to their traditional channels, so-called mixed agents7. The largest Swedish travel agency, American Express (formerly Nyman & Schultz), for instance, has introduced online booking as its standard pro-cedure. If customers instead prefer manual service (e.g., telephone), it costs extra. SAS sees this as a sign that the market is getting more mature and that the under-standing of the Internet is increasing. Some smaller travel agencies operate only through traditional channels, so-called traditional agents. Then there are travel agen-cies that operate only online, so-called Internet agents. SAS has its own telephone sales centres, in addition to its web site, www.scandinavian.net. SAS’s marketing channel system is illustrated in Figure 2.7.

Figure 2.7 SAS’s Marketing Channel System

The majority of SAS’s sales are made through travel agencies; in 2004 agents made between 56 percent (in Norway) and 84 percent (SAS International) of all book-

7 To make it clearer to the reader, the different types of travel agencies have been labelled. Please note that these labels are not used by SAS!

End Customers

Traditional Agents

Scandinavian Airlines

Internet Agents

TelephoneInternet

Internet Channel

Tele-phone

Channel

PhysicalChannel

Tele-phone

Channel

PhysicalChannel

MixedAgents

http://www.scandinavian.net


59

ings, whereas telephone sales accounted for between 9 percent (in Denmark) and 23 percent (in Norway). The Internet channel had between 6 percent (SAS Inter-national) and 21 percent (in Norway and Sweden) of the total bookings. SAS’s Internet sales through its own web site have increased substantially in recent years, but this increase comes according to management primarily from SAS telephone sales, not so much from sales through travel agencies.

SAS offers three different types of tickets for three different kinds of customers: comfort customers, productivity customers and low fare customers. Comfort customers are those whose priorities are extra comfort and service, e.g., comfort on board, access to lounges and better in-flight service. Productivity customers focus on flexible travel, efficiency and saving time. They want to utilize their travel time productively. Low fare customers do not want to pay for extras and want the lowest fare possible.

SAS has also introduced the brand name Snowflake on several of its international flights, through which it wanted to enter the low-cost carrier market. A number of airlines have entered that market during the last few years, offering a different product for which customers pay a very low price. Customers get exactly what they pay for, namely the right to fly; everything but the actual flight costs extra. Since this concept is based totally on price, it does not offer the opportunity to de-velop product advantages; therefore, it is completely different from SAS’s original concept. Accordingly, SAS launched Snowflake, since management believed that it could not accommodate low-cost products within and among its regular brand and products. Snowflake, however, departs from ordinary SAS airports, not, like most low-cost carriers, from remote airports. In a short time, Snowflake has become a well-recognized and reputable brand among customers. Destinations offered on Snowflake’s web site include, for instance, London, Frankfurt, Berlin, Brussels, Düsseldorf, Milan, Nice and Rome. These destinations, and several others oper-ated by Snowflake, also are served by at least one of the low-cost European carriers Ryan Air and FlyNordic.

When a travel agency books an SAS ticket, it usually goes through the General Distribution System, i.e. the GDS reservation system. It consists of reservation or-ganisations such as Amadeus, Galileo and so on. SAS has to pay a fee for each booking that goes through this reservation system. Snowflake, however, is an in-dependent SAS-owned web site (www.flysnowflake.com) which, just like other low-cost carriers such as Ryan Air, is not connected to the GDS booking system. The same applies to SAS’s own web site. It is therefore cheaper for SAS when cus-tomers book a travel on the SAS web site or Snowflake, rather than through travel agencies, since SAS does not then have to pay the GDS booking fee.

The basic idea is that all products should be available in all channels, provided it is cost efficient. Snowflake is therefore not available in the most expensive channels,

http://www.flysnowflake.com

______ CHAPTER 2 ______________________________________________________________________________

60

since that is not possible from an economic point of view. This means that Snow-flake’s tickets are available only on the Internet, through its own web site. Since Snowflake fares are so cheap, it would not be profitable to include Snowflake in the GDS reservation system, because the reservation fees, in relative terms, would be too high.

SAS also has been working to stimulate e-behaviour by encouraging customers to switch from paper tickets to e-tickets, which is the cheapest form of distribution. E-tickets provide ticket-less travel, i.e. customers do not need a paper ticket to check in. Instead, reservations are stored electronically in the SAS reservation sys-tem, and customers can use, for instance, their credit card or their frequent flyer card to make use of their reservation.

“Everyone gains from using e-tickets; we, the travel agency, and the customer. Customers must realise that they do not need a piece of paper when flying Stockholm-Luleå. But to change people’s behaviour and their way of thinking takes time. It does not happen over-night. On the other hand, no one will buy a ticket to, for instance, Borneo, on the Internet, and even fewer will go to Borneo on an e-ticket. If you are on Borneo, you want to be able to show a paper ticket at the check-in desk.”

– Klavs Pedersen, Director of Product Coordination, SAS Sweden

”In a few years, paper tickets are expected to disappear completely.” – The SAS Group’s Annual Report & Sustainability Report 2004, p. 19

On 30 March 2005, Scandinavian Airlines Sweden commenced sales of 300,000 low-price tickets costing SEK 450 on all Swedish domestic destinations flown to in 2005. The fare includes all taxes and fees, and it applies to bookings made at www.sas.se. It is also possible to book these tickets through travel agencies or by telephone, but then the price might be slightly different due to other service fees. From the same date, all domestic Swedish tickets have been exclusively e-tickets, so if customers prefer a paper ticket, they must resort to SAS’s telephone sales or to a travel agency and pay an extra fee (SEK 100 for SAS’s telephone sales).

“We know that people want to fly with SAS. With our new low fares and the highest num-ber of destinations and departures, it will be attractive for everyone to fly with us.”

– Anders Ehrling, President of SAS Sweden

The launching of low-price tickets was very successful; after just five days, SAS had sold 50,000 of its 300,000 tickets.

“The ability to fly simply and more cheaply with SAS was clearly attractive. The response from passengers, customers and travel agencies has been overwhelming. Or, as one traveller expressed it yesterday, ‘Now we can return to flying with SAS!’”

– Anders Ehrling, President of SAS Sweden

http://www.sas.se


61

Causes of Channel Conflict When SAS started to sell on the Internet, management was aware that it would lead to conflict with the travel agencies. This was not something new to the Inter-net marketing channel; on and off, there have been conflicts between SAS and its resellers, even before the Internet was introduced. One reason is that SAS previ-ously tried to increase its own telephone sales, which led to conflicts with the agencies.

Starting on 1 January 2003, SAS removed the commission on sales of tickets to travel agencies; at the same time the company reduced the prices of its tickets by the same percentages as the agencies previously received in commissions. SAS was one of the first European airlines to make this change, although several other air-lines soon followed suit. The commission used to be compensation that it was hoped would make travel agencies choose SAS instead of a competitor. Many travel agencies, however, had reduced the prices to customers by the same amount as the commission and then charged them a service fee; thus, in the end, the cus-tomers paid just as much as before. Management therefore viewed the commission as just a circulation of money. Another reason for the removal of the commission was that SAS wanted to adapt to customers’ changing habits. Specifically, custom-ers wanted to manage simple travel arrangements themselves, via the Internet. When SAS removed the commission, travel agencies saw it as a move to make it more expensive to purchase tickets through them than through SAS directly, and especially via SAS’s Internet channel, something that caused conflict. However, SAS states that even though adding the Internet caused some channel conflict, management does not think that the Internet per se has caused any more channel conflict than the company experienced before it was added.

“The Internet by itself has not caused any more channel conflict. It is very much about going through a travel agency or bypassing them, and that takes place both with or without the Internet. When the Internet was added it became clearer for travel agencies what we are doing as an airline company. They can enter our web site and see our prices, our product offerings, how we handle mass communication. They can see that we are always communicating that a customer can either contact us or a travel agency. It is sort of a tacit understanding that it does not matter which channel customers choose, as long as they fly with SAS.”


Customers using the different marketing channels have very different profiles. The majority of the customers who purchase tickets through SAS’s direct channels are private individuals; there are very few business customers, and most of these repre-sent small businesses. The routes sold through these channels are mainly simple, i.e. from A to B and back. The majority of customers that purchase through travel agencies, on the other hand, are companies and customers who purchase complex routing; the more complex the routing, the more consultation a customer needs.

______ CHAPTER 2 ______________________________________________________________________________

62

Even so, the basis for conflict in the travel market is the fight for customers. Travel agencies believe that SAS is trying to steal customers from them, that it is trying to make customers purchase directly from SAS. SAS thinks that the channel conflict, generally speaking, is caused by inadequate communication, which in turn leads to insecurity; travel agencies are not sure which customer segment purchases tickets where. Management also states that conflicts very often occur due to ignorance of what the other party is doing, or why.

“Insecurity creates conflict. Conflict is often caused by lack of knowledge, i.e. that you do not really know what is happening. Then it is easy to become defensive.”


Assessing the Seriousness of Channel Conflict SAS does not think there is much conflict in its relations with travel agencies. In addition, management does not think that the conflict that occurred when the Internet channel was introduced was a serious problem for the company. SAS did not see any risk in bypassing its resellers when selling online, since some travel agencies, e.g., Seat24 (www.seat24.se), sell only online, and they have access to the same prices that SAS offers, just like all travel agencies that sell online. SAS’s large partners, i.e. American Express, BTI Nordic, etc., have their own online solutions.

When SAS removed the sales commission to travel agencies, however, dangerous conflict occurred. Although, as already stated, some travel agencies used the com-mission as intended, others reduced the prices for tickets by the same amount as the commission and then charged customers a services fee. There were still other travel agencies that charged customers a service fee but kept the commission the same; for those agencies, the change was even bigger. Accordingly, the removal of the commission forced travel agencies that did not charge anything for their service to introduce a service fee. As previously stated, some travel agencies also saw the removal of commissions as a move to make it cheaper to purchase tickets through SAS’s Internet channel than through travel agencies. According to the newspapers, the resistance from travel agencies was very strong; several of them, as a protest, tried to avoid booking customers with SAS. Management, however, states that be-fore the change was made, SAS held long meetings with travel agencies, the Asso-ciation of Swedish Travel Agents SRF8 (Svenska Resebyråföreningen), and major customers (companies). As a result, the large travel agencies understood the neces-sity for change. Although some smaller travel agencies protested, most of them ac-cepted it.

8 SRF is an organisation of which the majority of Swedish travel agents belong. SRF’s business con-cept is to create the right conditions for travel agencies to develop profitable relationships with cus-tomers and suppliers.

http://www.seat24.se


63

In the beginning, the online prices offered through SAS’s Internet channel were lower than prices offered through other marketing channels. That is, SAS used to have SAS online-only prices, in order to arouse interest in online sales and to push development in that direction. But then SAS changed its approach, making those prices available through both the SAS web site, and the travel agencies’ web sites, i.e. online only. One reason for this is that the customers who did not have Inter-net access called SAS telephone sales and expected to get the same price when pur-chasing through that channel. SAS’s telephone sales channel thought that, since it was a part of SAS, it also should be able to sell that SAS product. But management could not offer SAS telephone sales the Internet product without offering it to the travel agencies as well. Therefore, management has almost ended with lower online prices, although SAS sometimes introduces campaigns with online-only prices, these campaigns are also available to online travel agencies, i.e. to both Internet agents and mixed agents (see Figure 2.7, p. 58). SAS has not offered SAS-only prices for a long time now, in order to avoid channel conflict. However, since 31 October 2004, EuroBonus9 customers earn 500 extra points if they book their bonus flight on SAS’s Internet channel; a deal that is not offered through any other marketing channel.

Although SAS could offer lower prices through its own marketing channels, as these are not connected to the previously discussed GDS reservation system, man-agement purposely has chosen not to offer lower prices in these channels in order to avoid channel conflict. In addition, all travel agencies are offered the same prices, since it is not in the interest of SAS to give certain travel agencies special fa-vours; all SAS cares about is that a travel agency sells tickets on SAS, and not on its competitors. Besides, end customers would get very upset if they realised that the basic price would have been different had they bought a ticket through another travel agency. In short, SAS does not have any conflict with its resellers about prices, since the company offers uniform prices in all marketing channels.

“If we have said that we want to have SEK 1,000 for a seat between A and B, and we get SEK 1,000 for that seat, we do not really care if the customer bought it through A, B, C, or D. What we live on is that the customer is onboard, flying with SAS. The important thing is that we have a working distribution network with full coverage, not that customers use a certain part of the distribution network. If all our aircraft were full, well then we could start to prioritise channels based on effectiveness. But, today, the most important thing is to get customers aboard.”


There are other conflict issues, however. Travel agencies think that they could provide customers with better service had Snowflake been incorporated in the

9 EuroBonus is the name of SAS’s frequent flyer program.

______ CHAPTER 2 ______________________________________________________________________________

64

GDS system, since it would be much faster and easier for them to find Snowflake in their ordinary reservation system. But SAS’s purpose has never been to make Snowflake available everywhere, just keep costs down. SAS also sees this as a po-litical statement toward the GDS reservation system; with today’s price trend, air-lines cannot afford to pay the high reservation fees that this system charges.

Another conflict issue is when travel agencies do not have their own online solu-tion. Traditional agents feel bypassed by those who offer online sales, especially since they do not have access to some of the products that are sold only online. In addition, although SAS tries to be clear in its communication with travel agencies, the agencies have sometimes misunderstood the meaning of a message.

Management states that, on the odd occasion, conflict between SAS and travel agencies has led to customer confusion. Since customers have a relation both with the travel agency and SAS, they get confused when they are stuck in the middle of the conflict.

SAS uses no structured approaches to find out about conflict with its resellers. However, the large travel agencies are visited about once a week or twice a month. Although the contact is not as frequent with the smaller travel agencies, SAS thinks that it keeps a close dialogue with these resellers also. In addition, SRF have regular meetings with SAS. These dialogues give SAS a good sense of when conflict exceeds appropriate levels.

Today, as far as the Internet is concerned, management does not experience mani-fest channel conflict. The Internet is becoming ubiquitous, since more people are using it to book airline and theatre tickets, search for information, etc. The under-standing of the Internet as a marketing channel has increased in recent years, which has led to less conflict. SAS does not see anything positive about channel conflict, and therefore the goal is to avoid conflict with its channel partners.

Approaches to Minimise Channel Conflict In order to avoid channel conflict, management states that it is important that both parties are clear in their communication. It also has to do with trust, which is built through good communication and good relationships. Therefore, SAS keeps a rather close dialogue with travel agencies, especially since these represent the major part of SAS’s sales and will continue to be very important to the company in the foreseeable future. The communication with large travel agencies such as American Express, BTI, CWT, etc., is handled through personnel assigned to the companies’ main offices. These persons make sure that agreements are followed, i.e. they de-velop basic plans, work extensively with communication, help SAS promote prod-ucts, and make sure that SAS also considers them. SAS also uses sales agents, nor-mally small businesses that are responsible for providing information to several


65

smaller travel agencies. The sales agents visit these smaller travel agencies within a certain district, making sure that the head of the agency as well as the personnel are informed about SAS’ products, activities, etc. SAS puts a lot of effort into working together with travel agencies on activities directed at end customers. Even though SAS does not, and cannot, dictate how travel agencies market themselves, or what products they market, SAS works with incentives and sometimes sponsors different marketing campaigns. By doing so, SAS hopes to create win-win situations for it-self and the agencies. Thus, the work of the sales agents is also concerned indirectly with minimising and removing potential conflict.

Since SAS regards communication as an important factor in reducing conflict, some years ago the company introduced an extranet to which travel agencies have password access. All sales information is now published through the extranet, in contrast to the old approach in which all information was sent to travel agencies by fax. Since travel agencies are more or less pelted with information, the use of fax as an information channel is very inefficient. The extranet offers travel agencies the ability to search for information in one place, when they need it and with instant access. It also makes it easier for SAS to highlight important issues.

As already stated, dangerous channel conflict occurred when SAS removed the sales commission. SRF and SAS, therefore, held a number of meetings. Those meetings led to some changes in terms and conditions. However, the travel agen-cies, especially the small ones, still felt that SAS was making it more expensive for customers to purchase tickets through them than directly from SAS. Therefore, SAS simultaneously introduced service fees on sales made through its direct chan-nels. The service fees were set at about the same price as the previous commissions. The travel agencies, of course, decided themselves how much they would charge customers for services provided.

“We set our service fees at a level that fairly corresponded to the previous commissions on sales. At the same time, we said that it is not our business how much the travel agencies charge their customers. They have to decide that themselves.”


If a ticket cost SEK 1,000 at SAS, it cost the same amount at the travel agency. In the beginning, for domestic Swedish travel SAS additionally charged SEK 100 as a service fee when customers purchased the ticket on the Internet. Some travel agencies duplicated this fee, without changes. Thus, at least there were no conflicts about prices. Others charged a higher fee, since they provided customers with ad-ditional services. On the other hand, some charged a lower service fee than SAS, since their business concept was to sell high volumes at low prices. Thus, the com-petition is really among the travel agencies, since they all receive the same prices

______ CHAPTER 2 ______________________________________________________________________________

66

from SAS. Today, SAS’s Internet channel charges SEK 50 for domestic Swedish travel. The fees for European or intercontinental travels are higher.

In recent years, SAS’s strategy has been to support its sales through all marketing channels, including travel agencies, because it is the customers who decide where to buy products. Customers who prefer a travel agency when purchasing a ticket are likely to keep that preference; that is how it has been so far, at least. In other words, as long as a customer purchases a ticket with SAS, management does not care where the customer bought it.

“If a customer purchases an air ticket at a travel agency, and that travel agency sells an SAS-ticket, then both the travel agency and SAS are satisfied. Hopefully the customer receives a good product, and then everyone is satisfied.”


Management states that the travel agencies are aware that customer profiles are dif-ferent in the different marketing channels, among other things, since SAS regularly informs them about this.

Unique Characteristics of the InternetManagement states that to make SAS’s extranet usable, it is important not to pub-lish everything; otherwise, it would be impossible for users to process all the in-formation. SAS utilizes a sort of 80/20-rule: Information that is relevant to most agencies is published on the site, but information about how to handle an odd re-quest, which occurs maybe every second year, is not published. That has been a learning experience for SAS; in the beginning, since it is so easy to publish some-thing on the extranet, the web site was too detailed.

SAS is also using the Internet to communicate with its EuroBonus customers. Fur-thermore, business customers will soon get access to a special corporate site, where they will be able to access their contracts, terms, statistics, etc.; such information, instead of being paper-based, will be web-based. Thus, companies will have access to the information at any hour of the day or night, and reservations can be made when it is most convenient for customers. SAS still prints a small timetable for Swedish domestic flights, but it no longer prints timetables with all SAS departures, as it did until a couple of years ago. The timetables for all scheduled flights are in-stead available online.

Management, however, wishes that SAS’s ordinary web site had been more user- friendly from the beginning.


67

“I wish that we have had a more user-friendly web site already from the beginning. Our web site used to be too complicated. It must be easy to book, easy to understand, and easy to handle.”


On 31 March 2005, SAS launched a new web site that is much more user-friendly than the old one. Customers no longer have to fill in unnecessary information just to research travel options, and the new web site remembers the inserted informa-tion in case visitors want to research another option during the same session, or if customers return to the SAS web site after visiting another site; the old site did not have these features.

The Internet has forced SAS to think in a totally different dimension. Management has to make sure that SAS’s products are available in a channel that more and more people, especially young people, are using to make purchases. The Internet has brought a different way to search for products. Management thinks that the Inter-net as a whole reduces both brand loyalty and distributor loyalty. SAS can see that customers are researching its offerings, but that they also go to competitors’ web sites to see what these can offer. Price has become a more important factor than it used to be, at the expense of other product features. The introduction of compari-son sites has accelerated this development. For instance, customers search for travel between A and B. The alternatives are almost always listed according to price, with the cheapest price first. It is therefore much more difficult on the Internet to con-vey the message to customers that SAS offers better service, that they will get bet-ter seats, better food, access to lounges, etc. All these augmented services become of minor importance. Management therefore sees that the Internet contributes to the impoverishment of products, i.e. it makes them ‘thinner’.

In addition, the Internet contributes to the dis-intermediation of travel agencies, since customers do their own shopping. Also, there are so many comparison sites, such as www.expedia.com, which make it easier for customers to plan their own travel. The competition in the travel agency business is intense. Although some customers still prefer to book a package tour with, e.g., Fritidsresor (affiliated com-pany of Tui), many research the Internet and then book a seat and a hotel, i.e. they arrange their own package tour. SAS thinks that although decreasing prices is good for customers in some ways, it also means that customers have to spend more time to find what they want, and that they often also get deprived of the added value that they previously took for granted. When SAS took away free food on some of its flights, some customers got very upset. But, with today’s price pressure, it is too expensive to include food in the cheapest fares.

http://www.expedia.com

______ CHAPTER 2 ______________________________________________________________________________

68

Non-Internet-Related Channel Conflict As previously stated, management does not think that SAS has experienced much conflict with travel agencies. Occasionally, however, SAS has encountered channel conflict that was dangerous to the company. This conflict has all been about who owns the customer according to management, and for the most part it has not been related to the Internet. For instance, SAS used to have contracts with large corpo-rate customers, which travel agencies perceived as a restriction of competition. SAS still has contracts with some of these large corporate customers, but the conditions have changed, and travel agencies are involved to make sure that everything works smoothly.

“I really do not feel that we have experienced much conflict with travel agencies. Rarely, however, we have had dangerous conflict with travel agencies, but it did not have so much to do with the Internet. The conflicts were about other issues, discussions about who owns the customer.”

– Klavs Pedersen, Director of Product Coordination

2.5 Analysisirst, a within-case analysis of each case, based on the frame of reference, will be presented. A cross-case analysis, in which the cases are compared to the frame of reference and to each other, will then be presented.

2.5.1 Within-Case Analysis – DucatiIn this section, the empirical evidence from the Ducati case will be analysed based on the frame of reference regarding the causes of channel conflict, the assessed se-riousness of channel conflict, and the approaches to minimise channel conflict.

Causes of Channel Conflict Table 2.6 includes an overview of the causes of channel conflict when the Internet is added, including the conceptualisation, the measures, and Ducati’s assessment of the role of the different causes of channel conflict. The grey colour indicates that findings related to the causes of conflict do not match the frame of reference.

As expected from, for instance, Smith et al. (1999), Ducati experienced conflict with its dealers when it started to sell on the Internet. Three different causes of conflict were included in the frame of reference: goal incompatibility, domain dissensus and inadequate communication.

Ducati’s dealers wanted to unveil new motorcycle models and be the meeting point with customers. Ducati, on the other hand, wanted to get direct feedback from its customers, to create entertainment for its customers, to jumpstart its Inter-

F


69

net program, and to experiment outside the company’s ordinary distribution net-work. Since the goals of the dealers and Ducati were incompatible, conflict occurred.

Also, conflict caused by domain dissensus occurred when Ducati started to sell on the Internet. Coughlan et al. (2001) maintain that this kind of conflict can occur if, e.g., bricks-and-mortar resellers provide customers with service without compensa-tion, since the sales goes through the Internet channel, so called free-riding.

Table 2.6 Within-Case Analysis of the Causes of Channel Conflict – Ducati

Conceptualisation- Causes of Channel Conflict

Measure Used Causes of Channel Con-flict – Ducati


Assessment of the role of incompatible goals among channel members as a cause of conflict (as perceived by the re-spondents)

Conflict caused by goal in-compatibility occurred be-tween Ducati and its dealers.

Conflict caused by domain dissensus occurred between Ducati and its dealers, despite compensation of dealers for Ducati’s Internet sales.

Domain Dissensus

Assessment of the role of one channel member being bypassed by another channel member as a cause of conflict (as perceived by the respondents) Conflict caused by domain

dissensus had not increased due to the adding of the Internet channel. Conflict caused by inade-quate communication oc-curred between Ducati and its channel members,

Inadequate Commu-nication

Assessment of the role of inadequate communication within the marketing channel as a cause of conflict (as per-ceived by the respondents) and was judged to be the

major cause of conflict by management.

In the case of Ducati, however, dealers received compensation for the service they provided. Despite this, the dealers were upset because they wanted to provide the service of unveiling motorcycles, and felt bypassed by Ducati, since they were robbed of this chance. The dealers wanted to be the place where the action oc-curred, and they wanted to be the starting point of the relationship with the cus-tomers. This is in line with Stern and El-Ansary (1988), who discuss functions or du-ties to be performed as one of four critical elements in the concept of channel do-mains, i.e. the role of each channel member in the eight marketing flows (see Chapter 1, Section 1.2.1 for a description of the marketing flows). In other words, by selling motorcycles online, Ducati took over the function of being the first meeting point with customers, a function which used to belong to the dealers; thus, dealers felt bypassed. Dealers also felt bypassed by Ducati because they thought that they could have sold the motorcycles themselves if they had just had

______ CHAPTER 2 ______________________________________________________________________________

70

the opportunity. Then they would have received a full commission on the online sale instead of the reduced commission. This represents the domain element popula-tion to be served (cf. Stern & El-Ansary, 1988).

Nevertheless, according to management, the major cause of channel conflict when Ducati added the Internet marketing channel was inadequate communication. That is, when communication was not clear, consistent, homogeneous, and timely, or when it was too redundant, conflicts occurred between Ducati and its dealers, be-tween Ducati and its customers, between dealers and customers, and among the in-ternal departments at Ducati.

In the literature, it is stated that the Internet will cause much more domain-related conflict than before (e.g., Alba et al., 1997). Ducati, however, stated that although it had experienced some channel conflict due to the addition of the Internet, it had not experienced more conflict with its resellers than before that channel was added. As already mentioned, inadequate communication was the major cause of conflict when Ducati sold motorcycles online.

Ducati did not discuss any causes of conflict other than those included in the frame of reference.

Assessing the Seriousness of Channel Conflict The empirical findings related to assessing the seriousness of channel conflict are presented in Table 2.7. The grey colour indicates that findings do not match the frame of reference.

The suggested approaches suggested for measuring channel conflict range from very structured (i.e. counting up the conflict issues as weighted by their impor-tance, frequency of disagreements, and intensity of dispute (cf. Coughlan et al., 2001)) to more informal approaches (i.e. evaluating customers’ and channel mem-bers’ responses to the conflict (cf. Moriarty & Moran, 1990)). Ducati did not use any structured approaches, such as formal surveys, to measure the level of conflict with its dealers. Instead, when the Internet marketing channel was added, man-agement spent a lot of time talking internally as well as to dealers. Management states that the channel conflict that occurred when selling motorcycles online was not dangerous to the company. Some problems occurred with its dealers, but nothing that management judged as critical. To verify this statement, five ques-tions, which were included in the frame of reference to judge the seriousness of channel conflict, will be examined with reference to the collected data. The ques-tions are: importance of channel, revenue in conflict, conflict issues, response of channel members and response of customers.


71

Table 2.7 Within-Case Analysis of the Seriousness of Channel Conflict – Ducati

Conceptualisation - Assessing the Seri-ousness of Channel Conflict

Measure Used Seriousness of Channel Conflict – Ducati

The reseller channels are extremely im-portant for Ducati, i.e. Ducati cannot sell motorcycles without the help of dealers. Importance of Chan-

nel

Assessment of the different marketing channels’ impor-tance to the company (as perceived by the respon-dents) Before launching online sales of motor-

cycles, conflict reducing measures were implemented to reduce conflict with dealers. Most of the customers who purchased motorcycles through the Internet chan-nel would have bought them through dealers instead, if possible. Therefore, although no real revenue was in con-flict, the dealers’ perceived that the channels were serving the same custom-ers.

Revenue in Conflict

Assessment of respondents’ view of the company’s dif-ferent marketing channels serving the same customers (as perceived by the respon-dents)

Customers purchasing apparel and/or accessories online do not have a dealer in their vicinity.

Four conflict issues (unranked) were mentioned by Ducati:

Product availability Customer assignment Reward system Communication problems

Conflict Issues

Assessment of the major conflict issues involved in the channel relationship (as perceived by the respon-dents) Previous research identified pricing as

possibly the biggest generator of chan-nel conflict, but pricing was not a con-flict issue for Ducati.


Assessment of channel members’ response to the conflict (as perceived by the respondents)

In the beginning, the reactions from dealers were not judged as causing real problems for Ducati. This changed over time.

Response of Custom-ers

Assessment of customers’ re-sponse to the channel con-flict (as perceived by the re-spondents)

Customers were not disturbed much by channel conflict between Ducati and its resellers.

Effect on Brand

Assessment of how the channel conflict affected the company’s brand (as per-ceived by the respondents)

The most important consequence of creating channel conflict, due to by sell-ing motorcycles online was its positive impact on Ducati’s brand.

______ CHAPTER 2 ______________________________________________________________________________

72

Regarding importance of channel, the reseller channels are extremely important to Ducati; it cannot sell motorcycles without the support of dealers, since, for in-stance, someone with training has to refer the motorcycle to the customer. Bucklin et al. (1997) state that any channel in conflict that is not in decline and carries more than 10 to 15 percent of volume and/or profit needs attention. In agreement with this, before launching online sales of motorcycles, Ducati implemented several conflict reducing approaches, which will be discussed further under the next head-ing. Although management knew that these sales would bring conflict with resel-lers, the advantages of jumpstarting the company’s Internet program, and the ex-perimentation outside the traditional network, were judged to be more advanta-geous that the consequences of the conflict with resellers. Even so, after selling motorcycles online four times, Ducati realised that it was too hazardous to risk their relationships with their dealers. Therefore, the online sales of motorcycles have come to a standstill. On the other hand, the online sales of apparel and acces-sories have continued, but, according to management, for the time being they are not important in terms of money.

Concerning revenue in conflict, i.e. whether two or more channels simultaneously are serving the same customers with the same product (Moriarty & Moran, 1990), the motorcycles sold on the Internet were not available in any other marketing channel. However, management thought that, with a few exceptions, the custom-ers who purchased a motorcycle online had the same profile as those who pur-chased a motorcycle through a dealer. In other words, had the motorcycle been sold through the dealer network instead, most of the same persons would have bought it. So, somewhat contradictory to theory, although there was no real reve-nue in conflict, since the motorcycles sold online were not sold through the dealer channel, i.e. the channels did not serve the same customers simultaneously, almost the same customers who purchased motorcycles through the Internet channel would have bought them through dealers instead, if possible. Hence, dealers’ per-ceptions were that Ducati was stealing their customers, i.e. that the channels were serving the same customers. Regarding sales of apparel and accessories, customers purchasing those products online were located far from a dealer. Thus, because the geographical profile of these online customers is different from that of those who purchase through dealers, there is no revenue in conflict.

The major conflict issues involved in Ducati’s channel relationship with its resellers were product availability, the reward system, customer assignment and communication prob-lems. Product availability concerns the dealers’ claim that they would have sold the MH900e (and its successors) at the same rate if the motorcycle(s) had just been given to them. This also produced discontent with the reward system, since if the dealers had sold the motorcycles themselves, they would have received the full commission instead of the reduced one they received from the online sales. Cus-tomer assignment concerns the fact that dealers wanted Ducati to force fan clubs to


73

be attached to them, which Ducati had decided not to do. Ducati did not want to act like a policeman, and management liked the idea of a big heterogeneous ‘tribe’. Communication problems occurred when dealers were misinformed by Ducati and when customers got access to information before dealers did. Webb (1997) identi-fied pricing, product availability, the reward system, customer assignment, promo-tion and resource allocation as the major conflict issues in a marketing channel sys-tem including the Internet. Three of these conflict issues (pricing, promotion and re-source allocation) were not mentioned by Ducati. It did however mention communi-cation problems, a conflict issue that was not identified in Webb’s research. More-over, in Webb’s (1997) research, product availability referred to the difficulty to match supply and demand on new products, meaning that sometimes certain hot items were not available through all channels. For Ducati, on the other hand, it meant that certain products were available only through the Internet channel. Also, in Webb’s research, the reward system referred to the conflict between the objec-tives of the various channels and what was best for the company, whereas for Ducati, it referred to dealers’ dissatisfaction with the commissions received on online sales. Furthermore, in Webb’s research, customer assignment referred to the crossover between different channels, i.e. customers shopping in both outlets and department stores, whereas for Ducati it referred to dealers’ desire to attach fan clubs to them. According to Webb (1997), pricing was perhaps the greatest conflict generator, but, as already stated, pricing was not mentioned by Ducati as a channel conflict issue.

As for response of channel members, the clearest sign that channel conflict is exceeding appropriate levels, according to McDonald (1999), is a higher than normal turn-over rate of channel partners. When selling motorcycles online, Ducati got bad re-actions from some of its resellers. But although some dealers were upset, none de-cided to stop selling Ducati because of the online sales of motorcycles. Manage-ment did not judge the reactions from dealers to be a critical problem for the com-pany, but after selling four motorcycles online, it decided not to do anything that might further weaken the dealer network. This is one reason why motorcycles have not been sold online for two years, and it is also why management has no plans to sell motorcycles online in the near future. Ducati still sells apparel and ac-cessories, but the response from dealers on these sales has not been negative.

Regarding response of customers, management said that the channel conflict between Ducati and its resellers when selling motorcycles online did not disturb Ducati’s customers much. Instead, the conflict was very positive for Ducati, since the online sales of motorcycles created excitement for the ‘fans’ and also allowed Ducati to get direct feedback from customers.

______ CHAPTER 2 ______________________________________________________________________________

74

One issue, the effect on brand due to the conflict, could not be included in any of the above described categories. That is, Ducati chose to take the conflict with its resellers about selling motorcycles online, among other things to create positive publicity and excitement for Ducati’s fans. In other words, selling the MH900e online created much positive response from the media and customers, which con-tributed to building Ducati’s brand. Accordingly, the most important consequence of Ducati creating the channel conflict with its resellers was the positive impact on its brand.

Hence, it seems that the conflict resulting from selling the MH900e was not a seri-ous problem for Ducati. Instead, it created the desired excitement for the com-pany’s customers, it allowed management to get direct feedback from customers, and to experiment outside the traditional network, and it helped to build Ducati’s brand. However, after four online sales of motorcycles, there was no longer any novelty, so the desired publicity did not transpire. Also, the profile of the online customers was almost the same as that of offline customers, so the company was not able to reach many new customers with its online sales of motorcycles. Since dealers are so critical to selling motorcycles, regardless of the channel used, man-agement decided that it was not worth risking its relationship with dealers any longer. In other words, although Ducati said that it did not experience dangerous channel conflict due to the online sales of motorcycles, it seems that, in the course of time, management judged the channel conflict to be dysfunctional. Thus, the online sales of motorcycles were paused. The online sales of apparel and accesso-ries, on the other hand, did not cause any conflict between Ducati and its dealers. These sales gave Ducati the opportunity to reach customers who were located far from dealers and also allowed it to keep control of the Internet channel; therefore, the online sales of accessories and apparel has continued. As management states, there is no way to avoid Internet management, i.e. if Ducati does not manage the Internet channel, its dealers will start to sell online, which would cause territorial conflicts.

Approaches to Minimise Channel Conflict Short descriptions of the findings on conflict reducing approaches are presented in Table 2.8. The grey colour indicates findings that did not match the frame of refer-ence. Six different approaches that can be used to reduce channel conflict were identified in the frame of reference: institutionalised, pricing, product version, brand name, compensation and communication approaches.

Institutionalised approaches refer to exchange-of-personnel programs, mediation, ar-bitration, relationship-building activities, etc. Ducati utilizes some institutionalised approaches. Since it is not possible to sell motorcycles without the support of deal-ers, customers had to choose a reference dealer within the country to which the motorcycle should be shipped. Although accessories and apparel sold online are


75

sent directly from Ducati to end customers, the same approach is used for those sales, i.e. customers have to choose a reference dealer, thereby reducing channel conflict. This approach is discussed with regards to compensation (e.g., Tsay & Agrawal, 2004), but the customer choosing a reference reseller is actually an insti-tutionalised approach, which does not necessary involve compensation. The results from Ducati’s online sales are made public to dealers once a month. Furthermore, Ducati sells used motorcycles (those used for various events) through its outlet in Bologna. Customers can pick up the purchased motorcycle at no charge at any Ducati dealer, i.e. the transportation is free for both for the customer and the dealer. Ducati also shows new motorcycles at this outlet but it does not sell them; instead, potential customers are referred to dealers. These approaches seem to be in line with two of the three relationship-building dimensions identified by Heide and John (1992): information exchange, which refers to channel members’ belief that all relevant information should be shared freely, frequently, quickly and thor-oughly; and solidarity, which refers to channel members’ belief that the channel re-lationship should bring mutual benefits.

Three different pricing approaches, all of which can supposedly reduce channel con-flict, were identified in the frame of reference: higher online prices, identical prices in all marketing channels and auction pricing. The MH900e (and its successors) were available only through the Internet channel; therefore, pricing was not an issue. A large portion of Ducati’s accessories and apparel, however, is available through all the company’s marketing channels. Ducati utilizes the MSRP, which means that the online prices for apparel and accessories are the same as the prices offered through dealers. Ducati states that pricing does not cause conflict with its dealers, since, once shipping costs and duty are included, the total purchase prices offered online are higher than those offered by dealers.

Product version approaches, i.e. when an offered product or product bundle is not available through all the company’s marketing channels, has also been suggested as a way to reduce channel conflict (e.g., Bucklin et al., 1997). Some of Ducati’s ac-cessories and apparel are not available through all marketing channels. The reasons are mainly legal and practical. That is, Ducati cannot legally sell anything directly to end customers that can be dangerous if improperly mounted on a motorcycle. For practical reasons, Ducati cannot sell, for instance, leather race suits to custom-ers online, since they have to try them on to see which size they actually need. Thus, these products are available only through dealers and, in line with theory, channel conflict did not occur between Ducati and the resellers. Ancarani (2002) claims that a producer can sell products online that are not offered through tradi-tional channels without causing channel conflict. However, although Ducati of-fered the MH900e (and its successors) only through the Internet channel, conflict with its dealers occurred. These conflicts are one reason why motorcycles have not been offered through the online channel for the past two years.

______ CHAPTER 2 ______________________________________________________________________________

76

Table 2.8 Within-Case Analysis of Conflict Reducing Approaches – Ducati

Conceptualisation – Conflict Reducing Approaches

Measure Used Conflict Reducing Approaches – Ducati

Online customers have to choose a reference dealer, within the country of product deliv-ery.Results from Ducati’s online sales are made public to dealers monthly.

Free transportation of used motorcycles from Ducati’s outlet to dealers, if desired by a cus-tomer.

Institutionalised Ap-proaches

Description of the in-stitutional mecha-nisms utilized by companies, intended to defuse disagree-ments before they es-calate into conflict (as described by the re-spondents)

Showing new motorcycle models in Ducati’s outlet, but not selling them. Prospective cus-tomers are referred to dealers.

Pricing Approaches

Description of the price levels used in different marketing channels to minimise channel conflict (as described by the re-spondents)

No conflict about pricing, since total pur-chase prices of apparel and accessories offered online are higher than total purchase prices offered by dealers.

Product Version

Description of com-panies’ use of product version approaches

Channel conflict occurred between Ducati and its resellers when the MH900e was sold only through the Internet channel

Approaches (as described by the respondents)

Accessories and apparel available only through the dealer channel has not caused channel conflict between Ducati and its re-sellers.

Brand Name Ap-proaches

Description of com-panies’ use of brand name approaches (as described by the re-spondents)

Ducati did not use different brand names in different channels.

Reduced commission to the reference dealer for delivering motorcycles sold online to end customer.

Compensation Ap-proaches

Description of the compensation ap-proaches used by companies (as de-scribed by the re-spondents)

Reduced commission to the reference dealer, without its actual involvement, for online sales of accessories and apparel.

Information about online customer to refer-ence dealer was provided.


77

To reduce communication conflicts, ducati.com was integrated with Ducati.

A base text describing the positioning of a product is distributed through all marketing channels to increase consistency of commu-nication, thereby reducing channel conflict.

An extranet is used to reduce communica-tion conflict with resellers.

Communication Ap-proaches

Description of the communication ap-proaches used by companies (as de-scribed by the re-spondents)

The web site is designed to minimise conflict between different customer segments.

Ducati did not use different brand names in different marketing channels.

The development of compensation mechanisms of resellers has been a neglected by companies, according to the literature (cf. Coughlan et al., 2001). Ducati, how-ever, had already developed a method for compensating resellers for online sales before the Internet channel was added. The compensation consisted of a commis-sion that was lower than the ordinary commission for a motorcycle. Although the dealers had no costs for marketing, inventory, etc., they were at least compensated for delivering the motorcycle to the customer. Dealers also are compensated for online sales of accessories and apparel. Also in this case, the commission is lower than the commission for sales made directly by the dealer. In this case, however, the dealer is not involved at all in the sales; the only reason the dealer receives the commission is that the customer chose that dealer as reference dealer (the customer must choose a dealer for an online purchase). Besides compensation in the form of money, dealers also get access to information about the customers. This offers them the opportunity to contact those online customers who chose them as reference dealer, in order to invite them to events, etc.

Although the necessity of good communication within a marketing channel system for the purpose of avoiding channel conflict is discussed by researchers (Etgar, 1979; Mohr & Nevin, 1990; Coughlan et al., 2001), no suggestions were found in the literature about how this should be done. Ducati, however, utilized several communication approaches. For example, management integrated ducati.com with Ducati in order to reduce the conflict due to differences in the communicated messages, and now all community personnel are gathered in the same large, open working area at Ducati’s main office. To make sure that a product is positioned the same way in all marketing channels, management developed a base text about the positioning, which then was distributed through all marketing channels, both in-ternally and externally. The use of base texts also helps Ducati avoid conflict with and among its customers. Many customers are more educated about the motorcy-cles than are Ducati’s marketers, sales people, and executives; also, customers have

______ CHAPTER 2 ______________________________________________________________________________

78

different entry points to the brand. Therefore, the communicated message has to be very clear to keep conflict from becoming too intense. Ducati also uses an ex-tranet to improve communication with its resellers, which has reduced conflict within the distribution network. To minimise conflict among different customer segments, Ducati’s web site is designed according to the concept of totem, in order to get customers to appreciate the whole world of Ducati, regardless of which en-try point (the engine, the design, racing, etc) they have to the company.

2.5.2 Within-Case Analysis - SAS In this section, the empirical evidence from the SAS case will be analysed, based on the frame of reference, regarding the causes of channel conflict, the assessed seri-ousness of channel conflict, and approaches to minimise channel conflict.

Causes of Channel Conflict The empirical findings on the causes of channel conflict when adding the Internet are presented in Table 2.9. The grey colour indicates that findings do not match the frame of reference.

Table 2.9 Within-Case Analysis of the Causes of Channel Conflict – SAS

Conceptualisation – Causes of Channel Conflict

Measure Used Causes of Channel Conflict – SAS


Assessment of the role of incom-patible goals among channel members as a cause of conflict (as perceived by the respondents)

Conflicts caused by goal incom-patibility occurred between SAS and the travel agencies.

Domain Dissensus

Assessment of the role of one channel member being bypassed by another channel member as a cause of conflict (as perceived by the respondents)

Conflicts caused by domain dis-sensus occurred between SAS and its channel members,

Conflicts caused by inadequate communication occurred be-tween SAS and travel agencies, Inadequate

Communication

Assessment of the role of inade-quate communication within the marketing channel as a cause of conflict (as perceived by the re-spondents)

and was judged by management to be one of the major causes of conflict with travel agencies.

When SAS added the Internet channel to its existing marketing channels, man-agement states that even though some channel conflict did occur, the Internet per se did not contribute to it. Even so, conflict occurred both with SAS’s other direct channels and with its resellers, i.e. the travel agents. The three causes of conflict in-cluded in the frame of reference were goal incompatibility, domain dissensus and inade-quate communication.


79

As pointed out by Friedman and Furey (1999), both suppliers and resellers want to maximise their profits, which might lead to goal-related conflict. This is what hap-pened when SAS added the Internet channel, i.e. conflicts occurred because of in-compatible goals; SAS wanted to increase sales through its Internet channel, whereas travel agencies, of course, wanted to increase sales through their own channels. In offering Snowflake only through an SAS-owned web site, management’s main goal was to keep costs down. Travel agencies, however, reacted to this, since their goal is to provide customers with the best service. Therefore, they think that it would have been better had Snowflake been incorporated into the GDS reservation sys-tem, as it is much easier for them to research their ordinary reservation system than a separate web site. Another reason for conflict caused by incompatible goals is that travel agencies wanted to keep the sales commissions, which were a large source of income. SAS, on the other hand, decided to remove the commissions, since they were just a circulation of money and the company wanted to adapt to customers’ increasing preference to handle simple travel themselves.

According to Stern and El-Ansary (1988), the four critical elements of a channel domain are the population to be served, the territory to be covered, the functions or tasks to be performed and the technology employed. Management states that the source of con-flict in the travel market is the fight for customers. As for the population to be served,customers who purchase through different channels have rather different profiles. Those who purchase through travel agencies are mainly businesses, or private per-sons who purchase complex routes, whereas those who purchase through SAS’s di-rect channels mainly purchase simple routes from A to B. However, it is the cus-tomer who decides where to purchase a ticket, which means that there are no in-disputable boundaries between different customer segments. Travel agencies, therefore, believe that SAS, through its direct channels, is trying to steal their cus-tomers. Regarding territory to be covered, the only limitation is that some products are available only through the Internet channel. However, this applies to SAS as well as the travel agencies. Concerning functions or tasks to be performed, if customers want paper tickets for Swedish domestic flights, they have to contact SAS’s tele-phone channel or a travel agency, since only e-tickets are sold on the Internet. Hence, the technology employed in this case is the Internet. However, even though these products are available only on the Internet, conflict arose because SAS’s tele-phone channel and the traditional travel agencies (which lacked an Internet chan-nel) felt bypassed and, therefore, also wanted access to e-tickets. In other words, conflict caused by domain dissensus occurred both between SAS’s direct channels, and between SAS and its travel agencies.

Conflicts caused by inadequate communication also occurred between SAS and its re-sellers. When travel agencies did not know which customer segment purchased tickets where, or why SAS was doing something, channel conflict occurred.

______ CHAPTER 2 ______________________________________________________________________________

80

Therefore, management generally attributed the channel conflict to inadequate communication.

As already stated, it is often implied that when the Internet channel is added, there will be more conflict caused by domain dissensus than before (e.g., Alba et al., 1997). Even though SAS management said that domain dissensus and inadequate communication were the major causes of the channel conflict, it claimed that add-ing the Internet channel did not increase the conflict.

Assessing the Seriousness of Channel Conflict The empirical findings on the assessment of the seriousness of channel conflict are presented in Table 2.10. Findings that do not match the frame of reference are marked in grey.

SAS’ representatives visit travel agencies quite often. In addition, SAS holds regular meetings with the Association of Swedish Travel Agents, SRF. Although this is not a structured approach to measure channel conflict, as suggested by Couglan et al. (2001), the meetings with travel agencies and SRF give SAS a good understanding of when conflicts exceeded appropriate levels. Management thinks that SAS has not experienced much conflict with travel agencies and that the dangerous conflicts that have occurred had little to do with the Internet marketing channel. Neverthe-less, five questions were included in the frame of reference to assess the seriousness of channel conflict: importance of channel, revenue in conflict, conflict issues, response of channel members and response of customers.

Bucklin et al. (1997) claim that a channel in conflict needs attention if it is not in decline and/or carries more than 10 to 15 percent of volume and/or profit. Since between 56 percent (in Norway) and 84 percent (SAS International) of SAS’s res-ervations were made through travel agencies, the travel agency channels were very important for SAS. Sales through both the SAS and travel agencies’ Internet chan-nels have increased in terms of market share, but according to SAS this increase came mainly from SAS’s telephone sales, and not so much from travel agencies. In accordance with Bucklin et al. (1997), SAS took several measures, discussed under the next heading, to reduce the conflict that occurred when the sales commission was removed.

Regarding revenue in conflict, i.e. whether two or more channels simultaneously serve the same customers (Moriarty & Moran, 1990), management says that the customers who purchase tickets through SAS’s direct channels are quite unlike those who purchase through travel agencies; hence, for the most part, the channels serve different customer segments.


81

Table 2.10 Within-Case Analysis of the Seriousness of Channel Conflict – SAS

Conceptualisation –Assessing the Seri-ousness of Channel Conflict

Measure Used Seriousness of Channel Conflict – SAS

The travel agency channels are very im-portant for SAS, since between 56% and 84% of SAS’ reservations are made through travel agencies. Importance of Channel

Assessment of the dif-ferent marketing chan-nels’ importance to the company (as perceived by the respondents)

Conflict reducing measures were taken af-ter removing the commission on sales to travel agencies.

Revenue in Conflict

Assessment of respon-dents’ views on the company’s different marketing channels serving the same cus-tomers (as perceived by the respondents)

Customers purchasing through SAS’s di-rect channels are different from customers purchasing through travel agencies’ chan-nels.

Five conflict issues were mentioned by SAS:

PricingCustomer assignment Product availability Reward system Communication problems

Conflict Issues

Assessment of the major conflict issues involved in the channel relation-ship (as perceived by the respondents) Previous research identified pricing as the

possibly largest generator of channel con-flict. But the removal of the commission, i.e. the reward system, was the most dan-gerous conflict issue for SAS.


Assessment of channel members’ responses to the conflict (as per-ceived by the respon-dents)

The reactions from travel agencies when removing the commission was dangerous to SAS.


Assessment of custom-ers’ responses to the channel conflict (as per-ceived by the respon-dents)

Customers were not much disturbed by channel conflict between SAS and the travel agencies.

Effect on Brand

Assessment of how the channel conflict affected the company’s brand (as perceived by the re-spondents)

The channel conflict that occurred when removing the commission on sales to travel agencies most likely had a negative impact on SAS’s brand.

______ CHAPTER 2 ______________________________________________________________________________

82

The major conflict issues involved in SAS’s channel relationship with the travel agencies were pricing, product availability, the reward system, customer assignment and communication problems. Pricing refers to the fact that, after the Internet was intro-duced, prices offered through SAS’s Internet channel were lower than prices of-fered through other marketing channels, since SAS wanted to arouse interest in online sales. However, both SAS’s telephone channel and the travel agencies were upset by this approach. Therefore, SAS has ended lower online prices, except for a few campaigns which are also available to Internet agents and mixed agents, i.e. travel agencies with an Internet channel. Product availability concerns the fact that some products are sold only through the Internet, a policy that traditional agents, i.e. travel agencies without an Internet channel, resist. Also, SAS’s telephone chan-nel thinks that they should have access to Internet products as well, since they are part of SAS. In addition, offering Snowflake only through a company-owned web site, rather than within the GDS reservation system, has caused conflict with travel agencies. According to management, when the reward system was changed, i.e. the commission removed, it caused dangerous conflict between SAS and travel agen-cies. One reason for this was that travel agencies saw it as a move to bypass them, i.e. to move ticket sales from them to SAS’s Internet channel. According to the newspapers, it even motivated travel agencies to stop booking SAS tickets as a pro-test. But, according to management, after long discussions with travel agencies’ representatives, and changes in approach such as the introduction of service fees from SAS’s side, most travel agencies understood the necessity for the change. Thus, today, SAS offers no sales commissions. The travel agencies have adjusted and are now charging customers service fees instead. While this conflict was in full progress, however, newspapers wrote several articles about it, which made custom-ers (and others) aware of the conflict. This confirms McDonald’s (1999) contention that the conflict was quite serious. Customer assignment refers to the crossover be-tween different channels, i.e. customers purchasing tickets through SAS’s direct channels as well as through travel agencies. Even though SAS tries to be clear in its communications with travel agencies, conflicts sometimes occur, because travel agencies misinterpret the message; thus, communication problems was also a conflict issue. In his research of a company using multiple marketing channels, including the Internet, Webb (1997) identified pricing, product availability, the reward sys-tem, customer assignment, promotion and resource allocation as the major conflict issues. As can be seen, two of these issues, promotion and resource allocation, were not mentioned by SAS. Communication problems, on the other hand, formed a conflict issue that, although mentioned by SAS, was not identified in Webb’s research. Furthermore, the meaning of product availability in SAS’s case is different from that in Webb’s (1997) research, since for Webb product availability refers to the diffi-culty in matching supply and demand for new products. This meant that on occa-sion certain hot items were not available through all channels, while for SAS it meant that some products were only available through the Internet channels. Also the reward system has a different meaning for SAS than the one provided by Webb


83

(1997). Whereas for Webb it refers to the objectives of the different channels being at odds with what was best for the company, for SAS it means that travel agencies were dissatisfied with the removal of the commission on sold tickets. Webb cites pricing as perhaps the greatest generator of channel conflict. In SAS’s case, how-ever, the reward system was identified as the issue causing the most dangerous con-flict between SAS and the travel agencies.

Concerning response of customers, SAS said that on only a few occasions were cus-tomers confused because of the conflict between SAS and the travel agencies. In other words, the channel conflict between SAS and its resellers disturbed SAS’s customers very little. However, since so much had been written in the media about the conflict that occurred when the commission was removed, customers became very aware of the conflict.

When SAS removed the sales commission, the media mostly reflected the travel agencies’ perspective on the conflict, reporting that travel agencies even tried to avoid booking customers with SAS. Therefore, the publicity most likely had a negative effect on SAS’s brand. However, the effect of channel conflict on a com-pany’s brand is not discussed in the channel management literature.

Approaches to Minimise Channel Conflict The empirical findings related to SAS’s conflict-reducing approaches, when it added the Internet, are presented in Table 2.11). Findings that do not match the frame of reference are marked in grey. In the frame of reference, six different ap-proaches to reduce channel conflict were included: institutionalised, pricing, product version, brand name, compensation and communication approaches.

Institutionalised approaches refer to joint membership in trade organisations, ex-change-of-personnel programs, mediation, arbitration, relationship-building activi-ties, etc. Joint membership in trade associations (Coughlan et al., 2001) can be compared to SAS’s regular meetings with the Association of Swedish Travel Agents (SRF), the organisation representing most of the Swedish travel agencies. SAS also keeps close contact with travel agencies; it visits them often (sometimes as often as twice a month), works closely with them in implementing activities aimed at end customers, and tries to create win-win situations for itself and the travel agencies. Visits to smaller travel agencies are handled by sales agents. These meet-ings aim first of all to inform travel agencies about new products, campaigns, etc., but since management thinks that trust is important in preventing conflict, and that trust is built through good communication and a good relationship, these meetings indirectly also help to minimise or prevent channel conflict. This institutionalised approach seems to be in accordance with two of the three relationship-building dimensions identified by Heide and John (1992): information exchange, which refers to channel members’ expectations that all relevant information should be shared

______ CHAPTER 2 ______________________________________________________________________________

84

freely, frequently, quickly and thoroughly; and solidarity, i.e. that channel members mutually try to build the relationship.

Table 2.11 Within-Case Analysis of Conflict Reducing Approaches – SAS

Conceptualisation – Conflict Reduc-ing Approaches

Measure Used Conflict Reducing Approaches – SAS

Institutionalised Ap-proaches

Description of the institutional mechanisms utilized by com-panies to defuse disagreements before they escalate into con-flict (as described by the re-spondents)

SAS engaged in relationship building activities, i.e. regular meetings with travel agencies, and meetings with SRF.

SAS used to offer lower prices through its own Internet channel than through all other channels, which caused channel conflict. SAS introduced a service fee on its sales when the commission to travel agencies was removed, to make it easier for travel agencies to charge for their services.

Pricing Approaches

Description of the price levels used in different marketing channels to minimise channel conflict (as described by the respondents)

Today SAS offers identical prices in all marketing channels and, some-times, lower online prices through all Internet channels.

Product Version Ap-proaches

Description of companies’ use of product version approaches (as described by the respon-dents)

SAS’ uses the product version ap-proach to target different customer segments through different marketing channels.The brand name Snowflake was in-troduced to compete with low-cost carriers, on a web site detached from the GDS reservation system, which caused ‘conflict’ with travel agencies.

Brand Name Ap-proaches

Description of companies’ use of brand name approaches (as described by the respondents)

In a short period of time, Snowflake became a well-known and reputable brand.

Compensation Ap-proaches

Description of the compensa-tion approaches used by com-panies (as described by the re-spondents)

SAS did not compensate travel agen-cies for sales made through SAS’s di-rect channels.

In advertising, reference is given both to SAS directly and to travel agencies. Communication Ap-

proaches

Description of the communi-cation approaches used by companies (as described by the respondents)

An extranet is used to reduce com-munication conflict with travel agen-cies.


85

Concerning pricing approach, in the beginning, prices available at the SAS web site were lower than those offered to travel agencies or through SAS’s telephone sales, since SAS wanted to arouse interest in online sales. This approach, however, caused channel conflict. SAS decided to offer the same online prices to travel agen-cies with Internet sales, even though it is cheaper for SAS when customers pur-chase tickets through SAS’s direct channels, which are not connected to the GDS reservation system. SAS’s telephone channel and those travel agencies without Internet sales were not content with this approach, however. SAS therefore has ended online-only prices, except for certain campaigns that are also available to travel agencies with Internet sales. Moreover, in order to avoid conflict, SAS has not offered SAS-only prices for a long time. In addition, to minimise conflict when removing the commission on travel agencies’ sales, SAS introduced a service fee on sales made through its direct channels. However, SAS does not charge this service fee to make it more expensive to purchase tickets through its channel, as suggested by King (1999). Instead, the purpose of the service fee was to make it easier for travel agencies also to charge a service fee. Furthermore, SAS’s pricing approach is not very similar to those identified in the literature review; prices are not always identical in all marketing channels, and they are not always lower online. Since all travel agencies are offered the same base prices as the ones SAS of-fers through its Internet channel, provided they utilize the Internet marketing channel, SAS does not experience any dangerous conflict with its resellers about pricing, although online prices sometimes are lower.

In the literature, multiple products or versions was suggested as a conflict-reducing mechanism (e.g., Ancarani, 2002), since it gives management the opportunity to target different customer segments through different channels. SAS utilizes the product version approach, i.e. it only offers e-tickets through its Internet channel on Swedish domestic flights. If customers prefer a paper ticket, they have to turn to SAS’s telephone sales or a travel agency and pay an extra fee. In other words, SAS’s product version approach serves as a way to target different customers through dif-ferent marketing channels, as suggested in the literature.

Regarding different brand names in different marketing channels, SAS offers Snow-flake, a brand aimed at the low-cost carrier market, solely through its own web site, www.flysnowflake.com. According to Gulati and Garino (2000), the decision of whether or not to use different brands in different marketing channels largely comes down to a choice between flexibility and trust. In SAS’s case, however, Snowflake quickly became a well-known and reputable brand. It, therefore, seems that offering different brands in different marketing channels does not always have to involve a choice between flexibility and trust. The brand name approach has been suggested as a way to avoid channel conflict (e.g., Coughlan et al., 2001). This was, however, not the major reason why SAS introduced Snowflake; it was introduced because SAS wanted to compete in the low-cost carrier market. Also,


______ CHAPTER 2 ______________________________________________________________________________

86

management did not think it was possible to accommodate such a product within the SAS brand, because a low-cost carrier is totally different from what SAS previ-ously offered. Customers now purchase the right to fly from A to B, and if they want more than the flight per se, they have to pay extra. Therefore, to keep costs down, the Snowflake web site was launched. By using this approach, SAS does not have to pay any reservation fees, since the web site, just like those of its low-cost competitors’, is not connected to the GDS reservation system. However, this ap-proach makes it more difficult for travel agencies to make a reservation; instead of the traditional approach of booking tickets through the GDS system, they need to go to Snowflake’s web site to effect a reservation. Some travel agencies objected that this made it more difficult to serve customers well. Thus, when SAS decided to offer a different brand through a separate web site, it did not come without fric-tion.

Compensation to travel agencies for sales made through SAS’s direct channels was not utilized by SAS.

Concerning communication approaches, management states that to avoid conflict with channel partners it is important to communicate clearly. The extranet, which SAS introduced some years ago, helps SAS in its communication with travel agencies. For instance, all sales information is published through the extranet. The extranet also makes it easier for SAS to highlight certain important issues. Furthermore, in its advertising, SAS tells its customers that they can contact either SAS directly, or a travel agency, to make a reservation. The latter method seems to be in accordance with one of the three relationship building dimensions identified by Heide and John (1992), i.e. solidarity, which refers to channel members’ belief that the channel relationship should bring mutual benefits.

2.5.3 Cross-Case Analysis The cross-case analysis is based on the within-case analyses presented earlier in this chapter. The cross-case analysis is, like the within-case analyses, structured in ac-cordance with the frame of reference. It will start by comparing the causes of channel conflict when adding the Internet, followed by assessing the seriousness of channel conflictand approaches to minimise channel conflict.

Causes of Channel Conflict The findings from both cases are presented schematically in Table 2.12. Departures from the frame of reference are marked in grey.

In agreement with theory (e.g., Smith et al., 1999), channel conflict occurred for both Ducati and SAS when the Internet was added to their existing marketing channels, even though both companies stated that the Internet per se had not con-tributed to more channel conflict. When analysing the data related to the causes of


87

conflict when adding the Internet, some similarities and dissimilarities can be iden-tified between the two cases. The three causes of conflict included in the frame of reference, i.e. goal incompatibility, domain dissensus and inadequate communication, oc-curred in both cases. No other causes of conflict could be identified for either case. Both Ducati and SAS stated that adding the Internet marketing channel did not contribute to more conflict with its resellers. Furthermore, contrary to the litera-ture (e.g., Alba et al., 1997), both Ducati and SAS claimed that conflicts caused by domain dissensus had not increased due to adding the Internet.

Table 2.12 Cross-Case Analysis of the Causes of Channel Conflict

Conceptualisation –Causes of Channel Conflict

Causes of Channel Conflict – Ducati

Causes of Channel Conflict – SAS

Goal Incompatibility Conflicts caused by goal in-compatibility occurred between Ducati and its dealers.

Conflicts caused by goal incom-patibility occurred between SAS and the travel agencies.

Conflicts caused by domain dis-sensus occurred between Ducati and its dealers,

Conflicts caused by domain dissen-sus occurred between SAS and its channel members.

despite compensation to dealers for Ducati’s Internet sales.

Domain dissensus was judged by management to be one of the ma-jor causes of conflict with travel agencies.

Conflicts caused by domain dis-sensus were mostly related to the population to be served,

Conflicts caused by domain dissen-sus were mostly related to the population to be served.

as well as functions or tasks to be provided.

Domain Dissensus

Conflicts caused by domain dis-sensus had not increased due to the adding of the Internet channel, according to manage-ment.

Conflicts caused by domain dissen-sus had not increased due to the adding of the Internet channel, ac-cording to management.

Conflicts caused by inadequate communication occurred be-tween Ducati and its channel members.

Conflicts caused by inadequate communication occurred between SAS and travel agencies.


Inadequate communication was judged to be the major cause of conflict by management.

Inadequate communication was judged by management to be a ma-jor cause of conflict with travel agencies.

However, dissimilarities also could be identified in the comparison of the cases. Ducati judged inadequate communication to be the major cause of conflict with its dealers, while SAS judged both inadequate communication and domain dissensus to be the major causes of conflict with travel agencies. Even though dealers were com-

______ CHAPTER 2 ______________________________________________________________________________

88

pensated for Ducati’s online sales of motorcycles, contrary to theory (cf. Coughlan et al., 2001) domain dissensus caused conflicts. The dealers felt bypassed by Ducati, since they wanted to be the starting point of the relationship with customers. In other words, domain conflicts occurred about functions or tasks to be provided. SAS, however, did not discuss this aspect of domain dissensus. For SAS, although cus-tomers decide the channel through which they want to purchase tickets in the travel market, and although the travel market is rather segmented, domain conflicts resulted from discussions of who owns the customer, i.e. the population to be served.This aspect was discussed also by Ducati, since dealers thought that they could have sold the motorcycles themselves if they had just been given the opportunity, thereby earning the full commission instead of the reduced one they received on Ducati’s online sales. Hence, also in the Ducati case, domain conflicts resulted from discussions of who owns the customers, i.e. the population to be served.

Assessing the Seriousness of Channel Conflict The findings on the cross-comparison of the cases are summarised in Table 2.13.Those that do not match the frame of reference are marked in grey.

Neither Ducati nor SAS used any of the structured approaches suggested in the lit-erature to measure the seriousness of channel conflict. Moriarty and Moran (1990) suggest a rather informal approach for assessing the magnitude of the conflict, i.e. asking questions about how much revenue a company has in conflict, where the conflict is, how channels and customers react to it, and how much time manage-ment spends dealing with the conflict. The approach utilized by both Ducati and SAS to find out about conflict seems to have been especially informal, since both companies stated that they spent a lot of time talking with channel partners, and, in SAS’s case, meeting with the trade association. However, none of them discussed any other ways to measure the level of conflict with its channel partners besides these meetings. Nevertheless, to assess the seriousness of channel conflict, five questions were included in the frame of reference: importance of channel, revenue in conflict, conflict issues, response of channel members and response of customers.

The reseller channels are extremely important for both for Ducati and SAS. It is not possible for Ducati to sell motorcycles without the assistance of dealers, and at least 56 percent of SAS’s tickets are sold via travel agencies. According to Bucklin et al. (1997), a channel in conflict needs attention if it is not in decline and/or car-ries more than 10 to 15 percent of volume and/or profit. Both Ducati and SAS addressed the conflicts with their resellers. There is a difference, though, between the companies: Ducati implemented several conflict-reducing measures beforelaunching online sales of motorcycles, whereas SAS introduced its conflict-reducing measures after the sales commission to travel agencies was removed. (These conflict-reducing approaches will be further discussed under the next head-ing).


89

Table 2.13 Cross-Case Analysis of the Seriousness of Channel Conflict

Assessing the Seriousness of ChannelConflict

Assessing the Seriousness of Channel Conflict – Ducati

Assessing the Seriousness of Channel Conflict – SAS

Approach to measure conflict with channel partners

Meetings with channel partners, but no formal approach to measure level of conflict with channel partners.

Meetings with channel partners and trade association, but no formal ap-proach to measure level of conflict with channel partners.

Importance of

The reseller channels are very impor-tant for Ducati, since Ducati cannot sell motorcycles without the help of dealers.

The travel agency channels are very important for SAS, since between 56% and 84% of SAS’s reservations are made through travel agencies.

Channel Before launching online sales of mo-torcycles, conflict reducing measures were implemented to reduce conflict with dealers.

Conflict reducing measures were taken after removing the commis-sion on sales to travel agencies.

The customers purchasing through the Internet channel had the sameprofile as regular customers.

Customers purchasing through SAS’s direct channels are different from customers purchasing through travel agencies’ channels.

Therefore, although the channels were not simultaneously serving the same customers with the same prod-uct, dealers’ perceived that there was revenue in conflict.

Revenue inConflict

Customers purchasing apparel and/or accessories online are customers who do not have a dealer in the vicinity

Four conflict issues were mentioned by Ducati:

Five conflict issues were men-tioned by SAS:

PricingCustomer assignment

Conflict Issues

Product availability Customer assignment Reward system Communication problems

Product availability

Reward system Communication problems

Previous research identified pricing as possibly the biggest generator of channel conflict, but pricing was not a conflict issue for Ducati.

Previous research identified pricing as possibly the biggest generator of channel conflict, but pricing was not the biggest conflict issue for SAS.The removal of the commission, i.e. the reward system, was the most dangerous conflict issue for SAS.

______ CHAPTER 2 ______________________________________________________________________________

90

The reactions from travel agencies when removing the commission were dangerous for SAS. Response of

Channel Mem-bers

In the beginning, the reactions from dealers were not judged as causing real problems for Ducati. This changed over time.


Customers were not disturbed much by the channel conflict between Ducati and its resellers.

Customers were not disturbed much by the channel conflict be-tween SAS and the travel agencies.

Effect on Brand

The most important consequence of creating channel conflict caused by the online sales of motorcycles was its positive impact on Ducati’s brand.

The channel conflict that occurred when removing the commission on sales to travel agencies probably had a negative effect on SAS’s brand.

Concerning revenue in conflict, i.e. whether two or more channels simultaneously try to sell the same product to the same customers (Moriarty & Moran, 1990), there are similarities, but also dissimilarities, between the cases. In SAS’s case, cus-tomers purchasing through travel agencies and through SAS’s direct channels, in-cluding the Internet, have very different profiles. Therefore, there is not much revenue in conflict. Also, in Ducati’s case, since the motorcycles sold online were not available through the dealer channels, there was no real revenue in conflict. However, since the profile of customers purchasing motorcycles on the Internet was similar to the profile of those purchasing them through dealers, the dealers felt that they could have sold the motorcycles themselves had they just been given the opportunity. In other words, in contrast to theory (cf. Moriarty & Moran, 1990), even though the channels were not simultaneously serving the same customers with the same product, the dealers perceived revenue in conflict, i.e. they per-ceived that Ducati took sales away from them by its Internet sales. For online sales of apparel and accessories, on the other hand, there was no revenue in conflict, since customers purchasing these products online were located far away from deal-ers and had no other access to these products.

There are several similarities between the cases, as well as dissimilarities, in the ma-jor conflict issues involved in the channel conflict. Product availability, the reward sys-tem, customer assignment and communication problems were mentioned as conflict issues in both cases. In each case, product availability meant that some products were sold only through the Internet channel. This differed from Webb (1997), where it meant that it was difficult to match supply and demand for new products and, thus, at times certain hot items were not available through all channels. The meaning of the reward system as a conflict issue differed between the cases. Ducati discussed it in terms of dealers’ dissatisfaction with the received commission on online sales, i.e. sales made by another party (Ducati) through another channel (the Internet), whereas for SAS it referred to travel agencies’ dissatisfaction with the removal of their sales commission, i.e. sales they made themselves through their own channels.


91

For (Webb, 1997), the reward system has yet another meaning, i.e. that the objec-tives of the different channels were at odds with what was best for the corporation. Also customer assignment had different meanings for the two cases: for Ducati, it meant that dealers wanted Ducati to force fan clubs to be attached to them, which Ducati, for various reasons, declined to do; for SAS, as for Webb (1997), customer assignment referred to the crossover between channels, i.e. customers purchasing through both reseller channels and the company’s direct channels. Both Ducati and SAS mentioned communication problems as a conflict issue; a conflict issue that was not identified by Webb (1997). However, since the motorcycles were not available through the dealer channels, and since prices offered online on apparel and acces-sories were higher than those offered through dealers, pricing was not a conflict is-sue for Ducati. For SAS, on the other hand, pricing was a conflict issue in the be-ginning, since prices offered through SAS’s Internet channel were lower than prices offered through other channels. Today, however, SAS does not experience conflict on pricing, since the company offers lower online prices only on rare oc-casions, and those prices are also available to travel agencies with an Internet chan-nel. Therefore, contrary to theory (cf. Webb, 1997), pricing was not the biggest conflict generator in either of the cases.

As for response of channel members, neither company regarded the reactions from re-sellers due to the addition of the Internet channel as dangerous to the company. However, in Ducati’s case this changed over time; after selling four motorcycles online, management determined that the sales no longer generated the desired publicity and decided not to weaken the dealer network any further. So, for the time being, they are not selling motorcycles on the Internet. McDonald (1999) states that the clearest sign of excessive channel conflict is a higher than normal turnover of channel partners. Although no dealer ended its relationship with Ducati, management saw the reactions of dealers as a problem. Apparel and acces-sories are still sold online, but for these sales the response from dealers has not been negative. In SAS’s case, management saw the negative response from travel agen-cies as a response to the removal of commissions, and not so much as related to the Internet. However, according to the newspapers, the conflict was very much re-lated to the Internet channel, since travel agencies saw the removal of the commis-sion as SAS trying to bypass them by making tickets sold through SAS’s Internet channel cheaper than tickets they sold themselves. According to the newspapers, travel agencies, in response to this conflict, even tried to make customers use air-lines other than SAS. It seems that for Ducati the response from channel members to Ducati’s direct sales on the Internet became a problem over time, whereas for SAS, the corresponding response probably became dangerous as soon as the com-mission on sales was removed.

Regarding the response of customers, neither Ducati nor SAS claimed that channel conflict has disturbed customers very much. However, SAS received a lot of nega-

______ CHAPTER 2 ______________________________________________________________________________

92

tive press when it removed its sales commission to travel agencies, and this made customers very aware of the conflict.

In fact, both companies received a lot of publicity when they added the Internet channel. In Ducati’s case, the articles focused on the fact that Ducati was selling high-priced motorcycles on the Internet and that the online sales were very suc-cessful; the focus was not on the channel conflict. For SAS, on the contrary, the articles focused on the conflict with travel agencies. The difference was that the publicity affected Ducati’s brand in a positive way, whereas, for SAS, it probably had a negative effect on its brand. How channel conflict affects a company’s brand, as an approach to assess its seriousness, is not mentioned in the literature.

Approaches to Minimise Channel Conflict The empirical findings on the cross-case analysis of the approaches companies used to minimise channel conflict are summarised in Table 2.14. Findings that do not match the frame of reference are marked in grey. Six different approaches that can be used to minimise channel conflict were included in the frame of reference: insti-tutionalised, pricing, product version, brand name, compensation and communication ap-proaches.

Institutionalised approaches, such as relationship-building activities, etc., were dis-cussed by both Ducati and SAS. Coughlan et al. (2001) discuss joint membership in trade associations as one such approach, which can be compared to SAS meeting with the travel agencies’ trade association as well as with travel agencies. The use of reference resellers has previously been discussed in the channel management litera-ture (e.g., Tsay & Agrawal, 2004), but then in the sense of compensating resellers. The basic function, in which customers have to choose a reference reseller on online sales, is, however, an institutionalised approach. Ducati utilized this ap-proach. The other institutionalised approaches utilized by Ducati (i.e. regular pub-lication to resellers of the company’s online direct sales, exposure of products sold only through resellers in the company’s direct channel, and free transportation to resellers of end customers’ purchases of motorcycles through the company’s direct channel) have not previously been discussed in the channel management literature, at least not in the context of minimising channel conflict. Although no similarities among the institutionalised approaches can be found between the cases, it might be inaccurate to say that all of the described approaches represent dissimilarities, be-cause the question was open-ended and the respondents might have forgotten some of the approaches in their answers.

When it comes to pricing, the companies in this study used different approaches. Obviously, pricing was not an issue when Ducati sold motorcycles on the Internet, since these were not available through other channels. Apparel and accessories, though, are sold both on the Internet and through dealers. Although the prices of-


93

fered online are the same as those offered by dealers, once shipping costs and duty are included the total purchase prices offered online are actually higher. Ducati does not experience any conflict with dealers about pricing, which is in accordance with the frame of reference, in which higher online prices was identified as one con-flict-reducing approach. When SAS started selling on the Internet, it offered lower prices through its own Internet channel than through all other channels, which caused channel conflict. SAS introduced a service fee on its sales to mitigate the conflict that occurred when eliminating the commission to travel agencies. Service fees were not discussed in the literature as a conflict-reducing approach. SAS also changed its pricing approach by offering the same lower online prices to all Inter-net channels, including those of travel agencies, but this approach also caused con-flict. Today, SAS offers the same prices through all marketing channels, except for certain campaigns that are also available to travel agencies that have Internet sales. Since all travel agencies are offered the same base price that SAS offers, provided they are selling through the Internet, SAS experiences no dangerous conflict with its channel members about pricing. This pricing approach, in which online prices usually are the same as in other marketing channels, but occasionally lower, could not be found in the conflict-reducing literature.

Both companies used product version approaches, i.e. when an offered product or product bundle is not available through all the company’s marketing channels. The results of such approaches, however, differed between the companies. Although product versions have been suggested as a way to reduce channel conflict (e.g., Ancarani, 2002), conflict still occurred when Ducati did not offer motorcycles through all marketing channels. That is, when motorcycles were sold only online, Ducati’s dealers got very upset, which is one reason why motorcycles have not been sold online for the past two years. For legal and practical reasons, among oth-ers, some of Ducati’s accessories and apparel are only available through dealers. This product version approach obviously has not caused any conflict between Ducati and its resellers, in accordance with the literature. SAS only offers e-tickets through its Internet channel, so if customers want to purchase a paper ticket, they have to turn to SAS’s telephone sales or a travel agency and pay an extra fee. SAS’s product version approach, therefore, serves as a way to target different customer profiles through different marketing channels, which has been suggested as another conflict-reducing approach (e.g., Bucklin et al., 1997). Ducati’s product version approach when selling motorcycles on the Internet, on the other hand, has not tar-geted different customer segments, since customers purchasing motorcycles online have almost the same profile as customers purchasing through dealers.

As for the use of brand name, the companies also were different; Ducati did not use a different brand name in different marketing channels, whereas SAS introduced the brand name Snowflake to compete in the low-cost carrier market. Although the purpose of introducing Snowflake was not to reduce channel conflict, the lit-

______ CHAPTER 2 ______________________________________________________________________________

94

erature suggests that the use of a different brand name in different marketing chan-nels can reduce conflict (cf. Coughlan et al., 2001). For SAS, however, the intro-duction of Snowflake led to ‘conflict’ with travel agencies, since Snowflake is only sold through its own web site (www.flysnowflake.com) and not through the GDS reservation system, which travel agencies normally use in making reservations. Gu-lati and Garino (2000) claim that the decision of whether to separate a brand in dif-ferent marketing channels is to a large extent a choice between flexibility and trust. According to management, Snowflake became a reputable brand in very short time, so, contrary to theory, it seems that using a separate brand name in different marketing channels does not always create problems of trust.

Table 2.14 Cross-Case Analysis of Conflict Reducing Approaches

Conceptu-alisation – Conflict Reducing Approaches

Conflict Reducing Approaches – Ducati

Conflict Reducing Approaches – SAS

Regular meetings are held with travel agencies and their trade association.

Online customers have to choose a ref-erence dealer, within the country of product delivery. Results from Ducati’s online sales are published to dealers once a month.

Free transportation of used motorcycles is provided from Ducati’s outlet to dealers, if a customer wishes.

Institutiona-lisedApproaches

New motorcycles are shown, but not sold, at Ducati’s outlet. Prospective cus-tomers are referred to dealers. There is no conflict about pricing, since the total purchase price online for ap-parel and accessories is higher than that offered by dealers.

PricingApproaches

When SAS offered lower prices through its own Internet channel than prices offered through all other channels, channel conflict occurred.

SAS introduced a service fee on its sales when the commission to travel agencies was removed.

Today, SAS usually offers identical prices in all marketing channels and, sometimes, lower online prices through all Internet channels.



95

Accessories and apparel that are only available through the dealer channel have not caused channel conflict be-tween Ducati and its resellers.

Product VersionApproaches

Ducati’s product version approach of selling motorcycles through only the Internet caused channel conflict.

Ducati’s product version approach re-garding online sales of motorcycles didnot target different customersthrough different channels.

SAS’ product version approach does target different customers through different channels.

Ducati does not use different brand names in different channels.

The brand name Snowflake was in-troduced to compete with low-cost carriers, on a web site detached from the GDS reservation system; this caused ‘conflict’ with travel agencies.

Brand Name Approaches

In a short period of time, Snowflake became a well-known and reputable brand.

Compensa-tion

Reduced commission to dealers for re-ferring motorcycle sold online to end customer, i.e. reduced commission for reduced service involvement.

No compensation to travel agencies for sales made through SAS’s direct channels, i.e. no commission, be-cause of no service involvement.

Approaches Reduced commission to reference deal-ers on online sales of accessories and apparel without their actual involve-ment in those sales. Ducati provides information about online customers to reference dealers. An extranet reduces communication conflicts with resellers.

An extranet reduces communication conflicts with resellers.

A base text, describing the positioning of a product, is distributed through all channels to increase consistency of communication, thereby reducing channel conflict.

Communi-cationApproaches

To reduce communication conflicts, the Internet business was integrated with the parent company.

The web site is designed to minimise conflict between different customer segments.

In advertising, reference is given to SAS directly and to travel agencies.

______ CHAPTER 2 ______________________________________________________________________________

96

Concerning compensation, Ducati and SAS used totally different approaches. Ducati’s dealers received a commission for Ducati’s online sales, whereas SAS did not compensate travel agencies for their online sales. The big difference between the companies is that Ducati’s dealers actually perform the service of referring mo-torcycles to customers; they are compensated for their services, whereas travel agencies are not involved at all in SAS’s direct sales. Since the service of referring motorcycles sold online to customers does not involve all the traditional services expected by dealers (marketing, inventory, etc.) dealers receive a lower than usual commission on these sales. As Coughlan et al. (2001) point out, when channel members no longer perform all the channel functions, they can no longer expect to receive a full commission. However, the compensation approach used by Ducati for the online sale of accessories and apparel contradicts this theory, since, although dealers are not involved in these sales at all, they receive a reduced commission on them. But compensation does not only come in the form of money; dealers also receive information about the online customers who choose them as reference dealers, which offers them the opportunity to contact these customers.

Both Ducati and SAS utilize an extranet to improve their communication with resel-lers, which both companies referred to as reducing channel conflict. Ducati also discussed other communication approaches, namely, the integration of ducati.com with Ducati and the use of a base text about the positioning of a product, to make sure that the communicated message is the same across channels. Furthermore, to reduce conflict between customer segments, Ducati’s web site is designed with the hope that visitors will appreciate the whole world of Ducati, regardless of their en-try point to the brand (the engine, the design, racing, etc.). SAS also pointed out that in its advertising it always tells its customers that they can make a reservation by contacting either SAS directly or a travel agency. SAS does this in order to communicate to customers and travel agencies that it does not matter where the customers purchase their tickets; all SAS cares about is that they travel with SAS. This method seems to be in agreement with what Heide and John (1992) call soli-darity, which refers to channel members’ belief that the channel relationship should bring mutual benefits.

2.6 Conclusionshis section provides an overall conclusion for the findings of this study. The findings and conclusions will be discussed first, followed by manage-rial implications, and, finally, suggestions for future research.

2.6.1 Findings and Conclusions The focus of this research has been channel conflict when adding the Internet. Al-though some channel conflict was described by the companies, it must be noted that both Ducati and SAS emphasized that they have a very good working rela-

T


97

tionship with their resellers, although on rare occurrence they have encountered problems. With that said, it is time to return to the purpose of this thesis, namely:

Based on the literature review, three research questions were posed:

1) How can the causes of channel conflict in multiple marketing channels, including the Internet, be described?

2) How can the seriousness of channel conflict in multiple marketing channels, including the Internet, be described?

3) How can the approaches that management utilizes in order to minimise channel conflict in multiple marketing channels, including the Internet, be described?

The findings and conclusions based on these research questions will be discussed in this section, which is divided into three parts: causes of channel conflict, assessing the se-riousness of channel conflict and approaches to minimise channel conflict.

Causes of Channel Conflict The findings from this study confirm that channel conflict will occur when com-panies add the Internet marketing channel to existing marketing channels. The study, however, has led to the identification of both similarities and dissimilarities between the companies included in this study with respect to the causes of channel conflict. In both cases, goal incompatibility, domain dissensus and inadequate communica-tion were described as the primary causes of conflict. No other causes of conflict were discussed by the respondents.

In agreement with theory, conflicts caused by goal incompatibility occurred, since the business goals of both the manufacturer and the service provider were different from those of their resellers.

Although it has been suggested that conflict caused by domain dissensus will increase with the adding of the Internet channel (cf. Alba et al., 1997), both companies claimed that such conflicts had not increased. This does not mean, though, that conflicts caused by domain dissensus related to the addition of the Internet did not occur. In the case of Ducati, domain-related conflicts occurred because several functions or tasks usually provided by the reseller were taken over by the manufac-turer. So, even though resellers were compensated for the manufacturer’s online sales, channel conflict occurred, contrary to theory (cf. Coughlan et al., 2001), since resellers felt bypassed in giving service to customers. In other words, by sell-ing products directly to end customers on the Internet, the manufacturer took over


______ CHAPTER 2 ______________________________________________________________________________

98

functions usually provided by resellers, which interfered with the resellers’ relation-ships with customers. Another reason for domain- related conflicts in this case is that the profile of customers purchasing online was so similar to that of customers purchasing through resellers. This led to conflicts about who owns the customers, i.e. population to be served. However, although the profiles of customers purchasing online and through resellers were different, discussions about who owns the cus-tomers also occurred in this case. In fact, although the customer profiles were dif-ferent in different channels, domain dissensus was judged to be among the major causes of conflict.

The manufacturer (Ducati) judged inadequate communication to be the major cause of channel conflict. When communication was not clear, consistent, homogene-ous, timely, or too redundant, communication-related conflicts occurred between the manufacturer and its resellers, as well as among various divisions within the company. The service provider (SAS) also judged communication conflicts to be one of the major causes of channel conflict; lack of knowledge in the channel rela-tionship leads to insecurity and, hence, conflict.

Hence, quite contrary to Webb’s (1997) findings, although inadequate communi-cation, as well as the part of domain that is related to lack of coordination, are im-portant as causes of channel conflict, this study indicates that the other parts in-cluded in the concept of domains, as well as incompatible goals, are also important causes of conflict when multiple marketing channels are utilized. To be precise, no matter how much the two companies included in this study would have commu-nicated or coordinated their operations with their resellers, occasionally conflict still would have occurred: (1) since some of the business goals of the manufacturer or service provider are different from those of their resellers, and (2) since, when the manufacturer started selling directly to end customers, domain conflicts were caused because some relationship-building activities were taken over from the re-sellers.

Assessing the Seriousness of Channel Conflict Neither of the companies included in this study used any of the structured methods suggested in the literature to assess the seriousness of channel conflict. Instead, meetings with channel partners, and in one case with the resellers’ trade associa-tion, provided the companies with a good understanding of when conflicts with channel members exceeded appropriate levels. Even so, the answers that were given to the questions included in the frame of reference helped achieve an under-standing of the conflict that occurred between the companies and their channel members.

It is apparent that the reseller channel is very important for both companies in-cluded in this research. For example, the manufacturer cannot sell its motorcycles


99

without the assistance of resellers, and most of the service provider’s sales go through resellers. The manufacturer still decided to sell motorcycles on the Inter-net, since it wanted to create excitement for presumptive customers and create publicity to build the brand. It therefore took several measures before launching its Internet sales to make sure that the resellers were involved in the sales process. The actions of the service provider, on the contrary, were judged by its resellers as an attempt to bypass them by means of direct sales through the Internet channel, so-called disintermediation. As a result, some resellers even tried to make customers purchase offerings from other service providers. Furthermore, the media wrote a great deal about this conflict, so customers became very aware of it. The service provider judged the conflict to be dysfunctional and therefore made changes in its Internet sales policy to reduce the conflict, but only after it occurred.

Although the service provider’s customers decide whether to purchase through the company’s direct channels or through the resellers’ channels, the different channels, for the most part, do not serve the same customers, as customers purchasing through the different channels have rather different profiles. The manufacturer sells apparel and accessories through both its web site, as well as through resellers. Also in this case, the profiles of customers purchasing through the different channels are different. However, the customers who bought motorcycles online had pretty much the same customer profile as those who purchase them through resellers. Therefore, as the customer profiles are nearly the same, and even though resellers do not sell the same motorcycle models that were sold through the manufacturer’s web site, the resellers thought that the channels were serving the same customers. In other words, somewhat contradicting theory (cf. Moriarty & Moran, 1990), re-sellers felt that revenue was in conflict because the profiles of the customers are the same, even though the manufacturer and reseller channels did not provide the same products to the same customers simultaneously; the result was conflict.

The manufacturer received a lot of publicity, but the focus of this publicity was not the channel conflict. Instead, as intended, the focus was on the product and the amazing fact that a manufacturer was selling high-priced motorcycles online. So, although the online sales created conflict with the resellers, this conflict was judged to be less important than promoting the brand and building the excitement that the sales created. In other words, the publicity had a positive effect on the manu-facturer’s brand, although it probably had a negative effect on that of the service provider. This issue, i.e. how the channel conflict affects a company’s brand, is not discussed in the literature.

However, when Ducati had sold some motorcycle models online the novelty dis-appeared, and the manufacturer did not receive much more publicity. At that point it was judged more important to manage the relations with resellers. That is, when the online sales no longer created excitement, it seems that management judged

______ CHAPTER 2 ______________________________________________________________________________

100

the channel conflict to be dysfunctional to the company. Therefore, motorcycles have not been sold online for the past two years. However, the online sales of ap-parel and accessories continue, since these make it possible for the company to reach customers who are located far from resellers and they have not created chan-nel conflict for the company. These sales also allow the manufacturer to keep con-trol of the Internet channel, because, as management states, if it does not manage the Internet channel, its resellers will.

Previous research on a multiple marketing channel system that includes the Inter-net identified several conflict issues: pricing, product availability, customer assignment, promotion, resource allocation and the reward system (Webb, 1997). Resource allocation was not mentioned as a conflict issue in the present study. However, since the in-terviews were open, the companies may not have mentioned resource allocation; however, it still might be a conflict issue. A conflict issue that was not identified in previous research but was mentioned by both companies is communication problems.The other issues mentioned by at least one of the companies are product availability, customer assignment, reward system and pricing. These conflict issues all relate to the previously identified causes of conflict, i.e. goal incompatibility, domain dissensus and inadequate communication. The service providers’ pricing and the companies’ reward systems relate to goal incompatibility, and product availability and customer as-signment relate to domain dissensus. Communication problems obviously relate to inadequate communication. Since the conflict issues described by the respondents can be related to all three conflict causes included in the frame of reference, this re-search does not support the suggestion made by Webb (1997), i.e. that goal in-compatibility and domain similarity rather than being causes of conflict represent types of conflict issues. Instead, this research seems to support the conclusion of Ong et al. (1990) that conflict issues are symptoms of the causes of conflict.

Moreover, neither of the companies included in this research thought that pricing was the biggest conflict issue. In fact, the manufacturer stated that it had no chan-nel conflict about pricing at all. Although the service provider had experienced some channel conflict about pricing when introducing the Internet channel, man-agement stated that the company no longer had any pricing-based channel conflict, since its approach to pricing had changed. Management instead stated that the larg-est conflict issue was the removal of compensation on sales to its resellers, which, due to the addition of the Internet channel, some of the company’s resellers saw as a move towards disintermediation. In other words, this research does not support previous research, which suggests that pricing is the single issue over which most of the conflict is generated (cf. Webb, 1997). On the contrary, this study indicates that different companies experience different conflict issues when using multiple marketing channels, including the Internet, and also that the conflict issues change over time.


101

The service provider did not find anything positive about channel conflict, and therefore stated that the goal is to avoid conflict with its channel partners. The manufacturer, on the other hand, stated that it tries to create conflict, do things that are radical, exciting and controversial, to build the company’s brand in order to justify the premium price of its products. Therefore, even though management knew that it would lead to conflict with its resellers, the manufacturer decided to sell high-priced motorcycles online. However, since the resellers are so crucial for the company, management took several measures beforehand to reduce the con-flict as much as possible.

Approaches to Minimise Channel Conflict Both the manufacturer and the service provider utilized several approaches to minimise channel conflict. The service provider built its relations with resellers through regular meetings with them, as well as with their trade association. This institutionalised conflict-reducing approach is discussed in the literature. The manu-facturer, on the other hand, utilized institutional approaches which have not been described, at least in that sense, in the literature. That is, the manufacturer’s online customers had to choose a reference dealer within the country of product destina-tion (who later received compensation on these sales), and results from the com-pany’s online sales were published to resellers monthly.

In accordance with the frame of reference, the manufacturer experienced no con-flict about pricing, since, although the base prices were the same for accessories and apparel regardless of whether they were offered through resellers or online, once shipping costs and duty were included the total purchase price was higher online. The service provider previously experienced conflict about pricing, since prices of-fered through the company’s web site were lower than prices offered through other channels. However, the company had introduced a service fee on its direct sales, and it also had changed its pricing approach. Now the company offers the same prices through all Internet channels. Sometimes Internet prices are lower than prices offered through traditional channels, but then the lower prices are offered also through resellers’ Internet channels. Thereby, the company experiences few pricing conflicts. This pricing approach, however, is not discussed in the literature.

Both the manufacturer and the service provider used the product version approach, i.e. some of the companies’ offerings were not available through all of its marketing channels. Ancarani (2002) claims that companies can avoid channel conflict by re-sorting to e-commerce only for products not available in traditional channels, re-ferring to Procter & Gamble. Nevertheless, Ancarani (2002) mentions nothing about the kind of products Procter & Gamble offers solely through its Internet channel. In the present research, however, even though the manufacturer decided to sell online only those motorcycle models that were not available through deal-ers, channel conflict still occurred. The dealers thought that they could sell these

______ CHAPTER 2 ______________________________________________________________________________

102

motorcycles had they been allowed to, especially since this product version ap-proach did not target different segments, as the customer profiles of online and off-line customers were practically the same. The service provider’s product version approach, on the other hand, targeted different customer segments, as suggested in the literature (cf. Bucklin et al., 1997). The results of this study therefore indicate that when producers sell to end customers highly desirable products directly that resellers could have sold just as well, channel conflict will occur, especially if the product version effort does not target different customer segments.

The use of different brand names in different channels has also been suggested as a way to reduce channel conflict. The service provider utilized this approach, al-though its primary objective was to compete with low-cost carriers and not to re-duce channel conflict. Even if some resellers complained about not having easy ac-cess to the low-cost brand’s offerings, no real conflict occurred due to this ap-proach. It has been argued that the decision to separate brands in different market-ing channels is a choice between flexibility and trust (Gulati & Garino, 2000). However, when the service provider introduced its low-cost brand, it became well known and reputable in a short period of time. Hence, contrary to theory, it seems that using separate brands in different marketing channels does not always raise trust issues.

Compensation of resellers for a company’s direct Internet sales has also been sug-gested as an approach to reduce channel conflict (cf. Coughlan et al., 2001). In agreement with theory, the service provider did not compensate its resellers, since they were not involved in its online sales. The manufacturer compensated its resel-lers with a lower than usual commission, since the resellers no longer were per-forming all the traditional services, such as stock holding or marketing; this also is in agreement with theory. However, contrary to theory, the manufacturer also compensated resellers for online sales of accessories and apparel, even though the resellers were not involved at all in these sales. The manufacturer also ‘compen-sated’ its resellers with information about the online customers who chose them as a reference reseller.

Both the manufacturer and the service provider used an extranet to improve com-munication with resellers; both companies cited this as reducing communication conflict. This means that the unique characteristics of the Internet can be used to reduce channel conflict. The manufacturer’s Internet business was integrated with the parent company to make it easier to communicate the same message through all marketing channels. The manufacturer distributed a base text about the posi-tioning of a product throughout all its marketing channels, to make sure that the communicated message would be the same across channels. Furthermore, the manufacturer’s web site was designed to minimise conflict among different cus-


103

tomer segments. Finally, in its advertising, the service provider refers to both the company’s direct channels and to resellers.

2.6.2 Theoretical Contributions The theoretical contributions of this study are based on only two case studies. However, as Yin (2003) claims, even a single case can confirm, challenge, or ex-tend theory.

Even though channel conflict occurred in both companies when they added the Internet channel, managers did not think that the Internet per se contributed to more conflict with its channel members.

Although goal incompatibility, domain dissensus and differing perceptions of reality are presently accepted as the major causes of channel conflict (cf. Coughlan et al., 2001), the findings from this study confirm the assumption made in the frame of reference, i.e. that differing perceptions of reality probably should be replaced with in-adequate communication as one of the primary causes of conflict, since differing per-ceptions of reality seem to be the result of inadequate communication. In fact, in-adequate communication was stated to be the major cause of channel conflict by one of the companies, and one of two major causes of conflict by the other, thus further strengthening this assumption. In addition, the channel conflict issues in-volved in this research could all be traced back to these three causes of conflict.

When assessing whether channel conflict is functional or dysfunctional, it seems that one important aspect to consider is its impact on a company’s brand. One of the companies viewed channel conflict as having only negative consequences, and it seems that this company’s brand had been negatively affected by the conflict with its resellers. On the contrary, the company that had experienced positive ef-fects on its brand, i.e. positive publicity, did not consider the channel conflict that occurred when adding the Internet a problem for the company.

This research also found that the Internet can be used to reduce channel conflict. Both companies referred to the extranet as a way to improve communication with resellers, which in turn was claimed to reduce communication-related conflict.

Furthermore, the pricing approach presently utilized by the service provider is dif-ferent from the ones identified in the literature review. Although this approach re-sembles of the identical prices in all marketing channel approach, they are not really the same, since the prices offered online sometimes are lower than prices of-fered through traditional agents, but then both through the service provider’s and resellers’ Internet channels. Therefore, it seems more correct to label this pricing approach as online specific pricing. In addition, the service provider added a service fee on its direct sales. This also contributed to reducing the conflict with its resel-

______ CHAPTER 2 ______________________________________________________________________________

104

lers, since the fee made it easier for resellers to charge for their services. Hence, based on the literature review (see Table 2.4, p. 35) and the results of this study, the following table of conflict-reducing pricing approaches when adding the Inter-net, was compiled (see Table 2.15).

Table 2.15 Conflict-Reducing Pricing Approaches When Adding the Internet

Pricing Approach Description

Higher Online Prices Higher prices on the Internet than in other channels Same basic price in all channels, but higher price on the Internet when handling and shipping fees are included

Identical Prices in All Marketing Channels Same price in all available marketing channels

Online Specific Pricing Same online price through both the producer’s and the resel-ler’s Internet channels, but not always the same as in tradi-tional channels

Auction Pricing Selling excess capacity through the Internet via auction online price uncertainty

2.6.3 Managerial ImplicationsNot all channel conflict can be avoided when producers add the Internet channel to existing marketing channels. Some of the business goals of a producer and a re-seller will always be different, which might lead to conflict, and as soon as produc-ers start to sell directly to end customers, resellers will fear disintermediation. In addition, not all channel conflict should be avoided. As Stern and Heskett (1969) maintain, without conflict, systems can become passive, non-innovative and, even-tually, non-viable. If Ducati had decided not to sell motorcycles online in order to avoid conflict with its resellers, the company would have missed the opportunity to get as much brand building publicity as it did.

However, the way a company handles the conflict when adding the Internet chan-nel seems to be crucial for the outcome of the conflict. The findings of this study indicate that it is important to handle the channel conflict proactively, i.e. to im-plement proper conflict-reducing measures before adding the Internet marketing channel so as, among other things, to reduce resellers’ concerns about disinterme-diation. One way for a company to minimise conflict with its resellers could be to involve them as much as possible in online sales. In Ducati’s case, for instance, re-sellers are even compensated for Ducati’s online sales of apparel and accessories, despite the fact that resellers are not involved in these sales at all.

Furthermore, both companies cited inadequate communication as the major cause (or one of the major causes) of conflict. Therefore, managers are advised to take measures to ensure that communication with its channel partners is working prop-erly. The use of a well-designed extranet could be a valuable tool in achieving this objective.


105

Ducati, as, for instance, Harley-Davidson, is a motorcycle brand that by itself is very niched. Therefore, by offering a motorcycle model only through the Internet, the company did not target different customer segments, since the profiles of online and offline customers were more or less the same; this, in turn, led to chan-nel conflict. However, it would probably have worked had Ducati offered the mo-torcycles sold online to dealers as well; then the competition would be level. The same pricing approach could be used, as that used when selling accessories and ap-parel online, i.e. the MSRP, plus the delivery cost, the duty, and the commission to reference dealers. Since online prices with this approach would be higher, it is likely that only customers located very far from a dealer would purchase motorcy-cles online. They still would have to pick them up at a reference dealer, since, ac-cording to Ducati, it is not possible to sell motorcycles without the assistance of a dealer. But the customer at least would not have to visit a dealer just to initiate the purchase. By using this approach, Ducati would probably not experience much conflict with dealers, since it would be clear to them that Ducati is attempting only to expand its market, not to bypass them. Another important consequence of this approach is that Ducati would be able to prevent dealers from charging too much over the MSRP. Although few motorcycles would probably be sold through the Internet, by using this approach the company would avoid reverse channel conflict (i.e. channel conflict due to resellers’ Internet sales) because, as the company stated, there is no way to avoid Internet management.

2.6.4 Suggestions for Future Research The findings of this research indicate that the way companies handle channel con-flict is crucial for the outcome of the conflict. Future research should therefore provide managers with further insights on how to manage conflict in multiple marketing channels, including the Internet.

An obvious extension of this research would be to investigate resellers’ perspectives on conflict in multiple marketing channels with a focus on the Internet. Another apparent extension would be to conduct a survey of both producers’ and resellers’ perspectives on such channel conflict, in order to evaluate the findings of this re-search.

In this study, inadequate communication was pointed out as a major cause of channel conflict. Future research, therefore, needs to address communication man-agement in multiple marketing channels, including the Internet. Since the extranet was pointed out as a valuable tool in reducing channel conflict, future research should provide managers with guidance on how an extranet can be designed in or-der to best serve its purpose as a conflict reducer. Furthermore, quantitative re-search is needed to evaluate whether one of the generally accepted causes of con-flict—differing perceptions of reality—should be replaced with inadequate com-munication, as the findings of this study indicate.

______ CHAPTER 2 ______________________________________________________________________________

106

Moreover, is channel conflict really the most serious concern for companies when adding the Internet? The results from Forrester’s research, in which 66 percent of the respondents indicated that channel conflict was the biggest issue facing manu-facturers selling online, were based on a survey of only fifty manufacturers (cf. Gil-bert & Bacheldor, 2000). Neither of the companies included in the present re-search, however, thought that channel conflict was the most serious concern when they added the Internet. Furthermore, management at both companies reported that it had not experienced more channel conflict than before that channel was added. Future research, therefore, is needed to address these issues.

__________________________________________ STUDY II: RESELLERS’ PERSPECTIVE ______

107

3 STUDY II: RESELLERS’ PERSPECTIVEThe aim of this chapter is to bridge the two studies that together form the basis for this doc-toral thesis, i.e. the producer study that was presented in the previous chapter, and the reseller study that will be presented in the forthcoming chapters. First, the research purpose of the the-sis will be reviewed, and the research questions of the reseller study will be presented. There-after, some delimitations of the study will be discussed.

3.1 Research Purpose and Research Questions n this second study, resellers’ perspectives on channel conflict will be studied. Although some of the research questions will be different from those in Study I, the research purpose remains the same. In the first chapter, the research pur-

pose of this thesis was stated as follows:

As stated in the second chapter, Rosenberg and Stern (1971) have developed a model that emphasizes the major aspects of the conflict process (see Figure 3.1).This model will serve as starting point also for the second study.

Figure 3.1 The Intrachannel Conflict Process (Rosenberg & Stern, 1971, p. 438)

The causes of channel conflict, the level of channel conflict, and the outcomes of the channel relationship are the topics to be studied. The justification for this choice is provided below with respect to each research question.

3.1.1 Research Question One Although goal conflict, domain conflict and differing perceptions of reality are presently accepted as the major causes of channel conflict (e.g., Stern & El-Ansary, 1988), the findings from Study I indicate that differing perceptions of reality probably

I

Causes of conflict

Structural and at-titudinal factors

Measurable level of conflict

Behavioralreaction(conflictresolutionstrategies)

Outcomes(behavioral and financial)


______ CHAPTER 3 ______________________________________________________________

108

should be replaced by inadequate communication, since differing perceptions of reality seem to be the result of inadequate communication. Hence, the first research ques-tion is stated as follows:

3.1.2 Research Question Two The second research question addresses the core of this thesis, namely channel con-flict. In Study I, producers stated that although channel conflict had occurred after the addition of the Internet channel, this additional channel conflict was not caused by the Internet per se. Accordingly, it seems interesting to investigate whether re-sellers think the same. Moreover, in order to investigate the impact of channel conflict, its level must be measured (e.g., Brown & Day, 1981). Hence, the second research question is formulated as follows:

3.1.3 Research Question Three As early as the 1960s, Pondy (1967) concluded that conflict is of concern to an or-ganisation only when it has implications for outcomes. Accordingly, the third re-search question is formulated as follows:

3.2 Delimitationsn Study I, producers’ perspectives were studied. In this second study, resellers of the two investigated companies (Ducati and SAS) will be researched. Al-though this approach will provide a fuller picture of the conflict within these

marketing channels, it will also restrict the study in some aspects. Since the study will be limited to these two companies, the number of respondents will also be somewhat limited. Therefore, it was decided to follow Cohen’s (1990) advice and include only a few constructs in the research, i.e. to apply the principle of “less is more”. Consequently, a number of interesting constructs will be omitted from this study (see below).

I

How can the causes of channel conflict in multiple marketing channels, including the Internet, be described?

How can the level of channel conflict in multiple marketing channels, including the Internet, be described?

How can the relationship between the level of channel conflict in multiple marketing channels, including the Internet, and the channel outcomes be described?


109

3.2.1 CommitmentSome researchers view commitment as an essential part of successful long-term re-lationships (e.g., Gundlach, Achrol & Mentzer, 1995). Although commitment has been found to be an antecedent of performance (cf. Skarmeas, Katsikeas & Schlegelmilch, 2002), the most common view among researchers is that commit-ment is the ultimate outcome of the channel relationship (cf. Geyskens, Steenkamp & Kumar, 1999). Hence, it would have been interesting to study commitment as the final outcome of the channel relationship. Nevertheless, because the intention was to limit the number of constructs included in this research, and since satisfac-tion is regarded as an antecedent of commitment (e.g., Hunt & Nevin, 1974), it was judged more important to include satisfaction and retain leave commitment for future research.

3.2.2 The Extranet In Study I, it was found that the Internet can be used to reduce channel conflict. Both companies referred to the extranet as a way to improve communication with resellers and thus reduce communication conflict. However, since the non-IATA travel agencies do not have access to SAS’ extranet (see Chapter 5.4.3, p. 136, for a further discussion), and since this was discovered only after the survey was distrib-uted (thereby making it almost impossible to distinguish between those who have access to the extranet and those who do not) the impact of the extranet could not be included in this study.

3.2.3 Functionality of Conflict and Impact on Brand Anderson and Narus (1990) included functionality of conflict in their study of the working relationship between distributors and manufacturers. Functionality of con-flict was described as follows: “in the minds of a firm’s managers, to what extent have recent disagreements with a partner firm been productively resolved?” (Anderson & Narus, 1990, p. 45) They developed two questionnaires to be as similar as possible, one aimed at distributors and the other aimed at manufacturers. They hypothesized a negative causal relationship between the functionality of con-flict and the level of conflict, since they believed that companies with the ability to resolve conflicts should have fewer recurring conflicts. However, they failed to find statistical support for their hypothesis in either study. It must be noted, though, that four items for each construct were included in their original ques-tionnaire, but after scale purification very few items remained. That is, only one item that measured functionality of conflict remained in the distributor study, and all items measuring this construct were dropped in the manufacturer study. More-over, all items used to measure conflict in the distributor study were dropped, and in the manufacturer study only one such item was retained. Other authors have suggested the opposite relationship, arguing that some conflict probably is func-tional but that there is probably a threshold level where conflict becomes dysfunc-tional (Boulding, 1965; Stern, Sternthal & Craig, 1973b). In Study I of this thesis,

______ CHAPTER 3 ______________________________________________________________

110

it was indicated that one important aspect to consider when assessing whether channel conflict is functional or dysfunctional is its impact on a company’s brand. One of the companies viewed channel conflict as having only negative conse-quences, and it seems that this company’s brand was negatively affected by the conflict with its resellers. On the contrary, the company that had experienced posi-tive effects on its brand, i.e. positive publicity, did not consider the channel con-flict that occurred when adding the Internet to be a problem for the company. Nevertheless, although it would be interesting to investigate the functionality of conflict, other concepts were judged to be more critical for this study.

3.2.4 Power and Interdependence What Etgar (1979) refers to as drive for autonomy (see Chapter 4.1), is referred to by other authors as power (e.g., Hunt & Nevin, 1974). Power can be described as the ability to get someone to do something he/she would not have done otherwise (Gaski, 1984). Power in channel relationships received considerable attention dur-ing the 1970s. Not all those who recognise a connection between power and con-flict share the same perspective. Some researchers claim that conflicts arise in a channel because superiors attempt to control the behaviour of subordinates, who resist such control, whereas other researchers view power as a result of, or a re-sponse to, conflict. But most researchers acknowledge that there is a causal rela-tionship between power and conflict, and that it can, and does, proceed in both di-rections (ibid.). Even so, unless one party is dependent on the other party, the abil-ity to influence the other party is limited (Frazier, Gill & Kale, 1989). As the de-pendence of each party upon the other party is the antecedent of power (e.g., Kumar, 2005), it would seem more interesting to study interdependence, because itrefers to the ability to replace both resellers and producers in the reseller’s trade area (Kumar, Scheer & Steenkamp, 1995).

That is, in Chapter 1.2.1 a marketing channel was defined as a system involved with the task of, either directly or indirectly, making anything of value available for use or consumption. It was also concluded that if a marketing channel cannot de-liver value on its own, it can do so indirectly by using another marketing channel. Hence, unless direct marketing channels are used, the actual delivery (i.e. the physical acquisition of the product), at least, will be handled by another marketing channel. Accordingly, in most cases, marketing channels are combinations of inter-dependent organisations that make a product available to the end user. Conse-quently, perceived interdependence is a crucial concept in channel research (Kumar et al., 1995).

Kumar et al. (1995) found that relationships with high interdependency, i.e. high symmetric dependence, displayed lower levels of conflict than asymmetrical rela-tionships. Furthermore, Frazier and Rody (1991) found that when the interde-pendence in a channel relationship is asymmetric, the more dependent party will


111

attempt to avoid conflict. A possible explanation of this could be that the more de-pendent party wishes to remain in a relationship which is as strong as possible, be-cause of the benefits it receives from the exchange (ibid.). Moreover, Frazier et al. (1989) argue that when resellers are in a highly dependent state, producers can pressure them without risking too much. Although this probably will lead to higher levels of perceived conflict, the reseller will most likely choose to remain in the relationship. So, even when conflict is high, and the reseller is relatively dissat-isfied with it, the reseller is unlikely to leave the relationship, because of its de-pendence on the producer (ibid.). This would suggest that although highly de-pendent resellers perceive conflict more frequently than resellers who are not so dependent on the producer, they former are less likely to give voice to such con-flict. Accordingly, it would have been interesting to include interdependence in the study.

However, interdependence is a formative construct (cf. Kumar et al., 1995), and the structural equation modelling software used in this study (AMOS) is not suit-able for formative analyses (cf. Chin, 1998). Because of the researcher’s non-existing knowledge on how to use other suitable software, such as the compo-nents-based approach known as partial least square (Chin, 1998), and because time did not allow the researcher to learn how to use such software, it was decided to retain interdependence for future research.

______ CHAPTER 4 ______________________________________________________________

112

4 DEVELOPING A RESEARCH MODELThe purpose of this chapter, based on the previously presented research purpose and an ex-tended literature review, is to develop relevant hypotheses and build an appropriate research model for the study of channel conflict in multiple marketing channels, including the Internet. The chapter will start with a discussion of the causes of channel conflict, followed by a discus-sion of channel conflict itself, and then the outcomes. The chapter ends with a graphical pres-entation of the research model, as well as a summary of the research hypotheses.

In order determine how channel conflict can be reduced or resolved, an analysis of its causes has to be made (Stern & Heskett, 1969). Therefore, the first part of the conceptual model focuses on the causes of channel conflict in multiple marketing channels, including the Internet (see Figure 4.1, p. 126).

4.1 Causes of Channel Conflictauses are the underlying bases of conflict that the parties may not even be aware of (Ong et al., 1990). The generally accepted causes of channel con-flict are goal incompatibility, domain dissensus and differing perceptions of reality

(e.g., Stern & El-Ansary, 1988). However, the selection of these causes is based on a conceptual paper (cf. Stern & Heskett, 1969), and only two empirical studies have been found that researched all three causes (cf. Rosenberg & Stern, 1971; Arndt & Ogaard, 1986). In addition, Rosenberg and Stern (1971) were not able to attain statistical significance for all conflict causes for each of their researched chan-nel relations.

Furthermore, although several scholars have described the causes of channel con-flict, most of their articles were written before the Internet existed, or at least be-fore it was used commercially, and the majority of the articles that discuss the Internet marketing channel are conceptual. In fact, only two empirical studies on the causes of channel conflict in multiple marketing channels, including the Inter-net, could be found: Webb and Hogan (2002), which is based on Webb’s (1997) doctoral dissertation, and the first part of this research (see Chapter 2). Webb (1997) conducted empirical research of four companies that were using multiple market-ing channels, including the Internet. It must be noted, though, that only one of these companies used the Internet as a sales channel. The case studies of these companies indicate that communication and coordination exert the most influence on conflict, and that domain similarity and goal incompatibility, instead of influ-encing conflict directly, represent issues about which conflict-related disagreements occur (Webb, 1997). However, the aspect of coordination is included in the con-cept of domain dissensus (see Section 4.1.2, p. 116 for a further discussion). In addi-tion, Webb (1997) concludes that the “properties of some of the measurement scales, most notably domain similarity and goal incompatibility were very poor” (p. 197), and therefore the results of his study should be interpreted with caution.

C

___________________________________________ DEVELOPING A RESEARCH MODEL ______

113

Moreover, in the first study of this dissertation (see Chapter 2), goal incompatibility,domain dissensus and inadequate communication were found to be the major causes of channel conflict. Therefore, it was concluded that further empirical research is needed to evaluate whether one of the generally accepted causes of channel con-flict – differing perceptions of reality – should be replaced with inadequate communi-cation, which seems to cause differing perceptions of reality (see Section 4.1.3 for a further discussion.)

Etgar (1979) instead discusses the causes of conflict in terms of attitudinal or struc-tural. Attitudinal causes are associated with disagreements about channel roles, expecta-tions, perceptions and communications, whereas structural causes consist of three factors: goal divergence, competition for scarce resources and drive for autonomy. Drive for autonomy,however, was not found to be an important source of channel conflict. The re-maining attitudinal and structural causes can all be attributed to goal incompatibility, domain dissensus or inadequate communication (cf. Etgar, 1979).

4.1.1 Goal Incompatibility A goal is the future position that an organisation wishes to attain (Schmidt & Ko-chan, 1972). Researchers claim that since the goals of one channel member very often are not compatible with the goals of another channel member, goal incompati-bility, and subsequently conflict, are very common (cf. Stern & Heskett, 1969; Schmidt & Kochan, 1972; Kochan et al., 1975; Perry & Levine, 1976; Cadotte & Stern, 1979; Etgar, 1979; Coughlan et al., 2001). Based on Reve and Stern (1979) and Cadotte and Stern (1979), the following definition of goal incompatibility is proposed:

Goal incompatibility is the extent to which different members of a marketing channel perceive that simultaneous goal accomplishment is impossible.

With the use of resellers, it becomes even more apparent that the goals of different channel members often are incompatible. Even though the use of resellers is much cheaper and offers the business greater flexibility, compared to a business-owned sales force (e.g., Friedman & Furey, 1999), coordinating relationships with resellers is a challenge, as they are independent businesses with multiple suppliers and prod-uct lines (Coughlan et al., 2001). Both the reseller and the supplier want to maxi-mise their profits. However, their ideas of how this should be accomplished often differ (Friedman & Furey, 1999). For the reseller it could mean charging customers more, while at the same time holding down expenses. The supplier, on the other hand, wants to see the reseller do almost exactly the reverse (Coughlan et al., 2001). Accordingly, conflict issues might include whether prices and service are being maintained at “reasonable” levels (Reve & Stern, 1979).

Despite the above, the goals of channel members have seldom been included in re-search on marketing channels (Achrol & Etzel, 2003). In fact, only a few empirical

______ CHAPTER 4 ______________________________________________________________

114

studies that investigated to at least some extent the relationship between goal in-compatibility and channel conflict could be found in the channel conflict literature (cf. Assael, 1969; Rosenberg & Stern, 1971; Rosenberg, 1974; Etgar, 1979; Eliash-berg & Michie, 1984; Arndt & Ogaard, 1986; Webb, 1997; Webb & Hogan, 2002). Assael (1969) conducted a content analysis of the frequency and intensity of conflict as documented in government and trade publications. This analysis led to the suggestion that goal incompatibility caused ‘friction’ between producers and re-sellers. Similarly, Rosenberg (1974), in a qualitative study of distribution managers, suggested that incompatible goals between channel members create ‘a level of ten-sion’.

Among quantitative studies on the relationship between goal incompatibility and level of channel conflict, Rosenberg and Stern (1971) conducted a case study of a manufacturer-dealer-distributor channel. They found statistical support for goal in-compatibility as a cause of conflict between manufacturers and distributors, and be-tween manufacturers and dealers, but not between distributors and dealers. On the contrary, Arndt and Ogaard (1986) found statistical support for goal incompatibility as a cause of channel conflict in all their researched relationships. However, judg-ing from the sample item that they provide, which supposedly measures goal in-compatibility (“The manufacturer often wants us to carry so many new cars that our inventory becomes too large”), it seems more likely that the item measured disagreement on stock-keeping than underlying goal incompatibility (cf. Arndt & Ogaard, 1986, p. 193). On the other hand, in their article based on Webb (1997), Webb and Hogan (2002) found statistical support for the hypothesized relationship between goal incompatibility and intensity of channel conflict, but not between goal incompatibility and frequency of conflict. In contrast, Etgar (1979) differenti-ated between perceived (cognitive/affective) conflict and behavioural/manifest conflict. He found that goal incompatibility correlated much higher with manifest conflict than with perceived conflict.

Finally, Eliashberg and Michie (1984) investigated the impact of a variety of differ-ent multiple business goal on the perception of the level of conflict in marketing channels. They studied a number of different business goals, and they asked both franchisees and franchisors to indicate the relative importance of those goals; thereby, they obtained indications of preference incongruities and perceptual dif-ferences between franchisees and franchisors. However, it should be noted that their study did not measure whether these differences in goals were incompatible. This might explain the inconclusive relationships that they found between differing business goals and the level of channel conflict.

Nevertheless, the main reason why so few researchers have included channel member goals in research on marketing channels might be that it is often assumed that resellers’ goals are an extension of producers’ goals (Achrol & Etzel, 2003), even though the previous discussion clearly indicates that this is not the case. An-


115

other reason, which Brown, Lusch and Smith (1991) refer to, is that, in a concep-tual paper, Stern and Gorman (1969) claim that “Goals are not the cause of con-flict; it is the behavior patterns pursued to achieve goals that interfere with goal at-tainment of other system members. It is true that goals produce behavior, but it is behavior, not the goals, which frustrates.” (p. 163). Yet, as previously stated, the parties in a conflict may not even be aware of its underlying causes (Ong et al., 1990). In other words, it seems that incompatible goals are the cause of conflict, whereas frustration behaviour is the symptom. It would therefore seem illogical to exclude goal incompatibility as a cause of channel conflict. Accordingly, the fol-lowing hypothesis is proposed:

Also, when an Internet channel is added, goal incompatibility can occur if the manufacturer wants to maximize its profits over the totality of its marketing chan-nels, including the online channel, since the Internet channel may offer the manu-facturer a higher margin on sales than do the traditional bricks-and-mortar chan-nels (Coughlan et al., 2001). In addition, in order to cover the costs of establishing and operating the new channel, customers may have to be pushed to buy directly from the online channel rather than through the resellers (ibid.). It seems likely that this “push” of customers will lead to domain dissensus, since the different chan-nels will compete to reach the same customers (see Section 4.1.2 for a further dis-cussion on domain dissensus). Accordingly, the following hypothesis is proposed:

4.1.2 Domain Dissensus The three critical elements of a channel domain are (1) the products to be offered; (2)the population to be served; and (3) the services to be rendered or functions to be performed(cf. Rosenberg & Stern, 1971). Webb and Lambe (2007) refer to these three ele-ments as the external sources of domain dissensus. They further discuss the internal sources of domain dissensus, which are “funding allocation, employee skills, and reliance among the channels on common manufacturing units.” (p. 33). Webb and Lambe (2007) researched internal multi-channel conflict by interviewing managers on the producer side of the producer-reseller dyad; therefore, they included both internal and external sources of domain dissensus. The present study, however, is intended to measure resellers’ perception of domain dissensus with regards to the focal producer. Hence, the study of domain dissensus will be limited to external sources. Therefore, based on Rosenberg and Stern (1971) and Foster and Shup-trine (1974), the following definition is proposed:

H1: The level of goal incompatibility among multiple marketing channels is positively associated with the level of channel conflict.

H2: The level of goal incompatibility among multiple marketing channels is positively associated with the level of domain dissensus.

______ CHAPTER 4 ______________________________________________________________

116

Domain dissensus exists when channel members’ claims are incongruent in terms of (1) the products to be offered, (2) the population to be served, and/or (3) the services to be rendered or functions to be performed.

Webb (1997) discusses coordination as one of the most influential causes of con-flict. However, coordination is included in the concept of domain dissensus, since the three critical elements of the channel domain imply coordination. Conflict is-sues with regards to domain dissensus might include who has the right to represent a particular product within a given territory (Stern & El-Ansary, 1977; Reve & Stern, 1979; Magrath & Hardy, 1987). Specifically, when manufacturers use several marketing channels simultaneously, these channels sometimes find themselves competing to reach the same customers. This might cause distributors and dealers to be bypassed by direct selling, which almost certainly will cause channel conflict (Magrath & Hardy, 1987; Bucklin et al., 1997).

Among empirical studies investigating the causes of channel conflict, domain dis-sensus, or at least some of the three elements included in the definition of domain dissensus, seems to be the most researched cause; even so, relatively few studies have been conducted (cf. Assael, 1969; Rosenberg & Stern, 1971; Foster & Shup-trine, 1974; Shuptrine & Foster, 1976; Etgar, 1979; Ross & Lusch, 1982; Arndt & Ogaard, 1986; Frazier et al., 1989; Brown & Fern, 1992; Webb, 1997; Webb & Hogan, 2002). With one exception (Ross & Lusch, 1982), these studies found at least some support for domain dissensus (or components of domain dissensus) as a cause of channel conflict. Nevertheless, although Webb and Hogan (2002), (which is based on Webb (1997)), conducted their study after the Internet started being used commercially, they did not distinguish Internet-related domain dissensus from other kinds of domain dissensus. In other words, no empirical study could be found that explicitly researched Internet-related domain dissensus.

Even so, researchers claim that, because the Internet has allowed businesses to es-tablish a direct channel to customers, the emergence of e-commerce has made do-main conflicts an even more burning issue (cf. Alba et al., 1997). A manufacturer selling online may directly cannibalize the customers of the bricks-and-mortar channel, thereby causing conflict over the population to be served (Coughlan et al., 2001). The bricks-and-mortar channel, of course, is convinced that the cus-tomers who choose to shop online at the manufacturer’s site would shop in a physical store if the web site did not exist. Furthermore, the Internet pushes old territorial borders aside, since it can take away business from any location in the market area served by the bricks-and-mortar channel, thereby causing domain con-flict over territorial rights (ibid.). Accordingly, the following hypothesis is pro-posed:

H3: The level of domain dissensus among multiple marketing channels is positively as-sociated with the level of channel conflict.


117

4.1.3 Inadequate Communication Even though the individual goals within the marketing system may correspond with one another, and the domains may be well defined, channel conflict still may occur due to differing perceptions of reality, i.e. channel members having conflicting perceptions of the same situation (Stern & Heskett, 1969). This is an important source of conflict, because it indicates that there will be different responses to the situation, which will frustrate channel members and produce conflict (Coughlan et al., 2001). One major reason for differing perceptions of reality is that different channel members are exposed to different information and influences, giving them different pieces of the overall picture (Coughlan et al., 2001). As Mohr and Nevin (1990) contend, communication is “the glue that holds together a channel of dis-tribution” (p. 36). That is, in order to achieve the needed coordination among channel members, adequate communication within the channel is necessary. For instance, information about new products, promotional campaigns, technical inno-vations, etc., has to be transferred from the producer to the resellers, whereas resel-lers have to inform the producer about market conditions, customers reactions to products and so on (Etgar, 1979). Furthermore, if the role of each channel is clearly and explicitly communicated, unwanted conflict can be effectively elimi-nated even before it occurs (Webb, 1997). In other words, frequent two-way in-terchanges are needed in order to achieve coordination among channels (Anderson & Weitz, 1992). Nonetheless, too much communication is just as bad as too little, since too much contact can overload channel members and have dysfunctional consequences (Mohr & Nevin, 1990).

When communication is working properly, it reduces dysfunctional conflict and nurtures confidence in the continuity of the relationship, i.e. commitment and loy-alty are cultivated (e.g., Anderson & Weitz, 1992). On the other hand, when communication is inadequate, misunderstandings will occur, incorrect strategies will be implemented, and mutual feelings of frustration will arise (Etgar, 1979). Accordingly, if there are communication problems within a marketing channel, different channel members will perceive the reality differently. If one uses this per-spective in analysing the table based on Rosenberg and Stern’s (1971) research on the different causes of conflict (see Appendix 1), it becomes clear that different percep-tions of reality is not really a cause; instead, it consists of issues that seem related to inadequate communication or domain dissensus, even though some of their sample is-sues do not seem to reflect conflict. Yet, some of their miscellaneous conflict issues might best be categorised as caused by inadequate communication, although this is im-possible to confirm with certainty without access to the original research data.

Nevertheless, Coughlan et al. (2001) argue that differing perceptions of reality also occur frequently between manufacturers and resellers when manufacturers add the Internet sales channel to existing bricks-and-mortar channels. Manufacturers often believe that they are just expanding their market reach, generating sales from con-

______ CHAPTER 4 ______________________________________________________________

118

sumers who can or will purchase the company’s products only through the Inter-net. Actually, 38 percent of the survey respondents who spent more online on holiday gifts in 1999 than in 1998 gave as their reason that they were able to pur-chase products online that were not accessible to them near where they lived or worked. Even so, Coughlan et al. (2001) continue, bricks-and-mortar resellers probably will perceive sales made on the manufacturer’s web site as having been stolen from them. However, when one looks closer at Coughlan et al.’s (2001) dis-cussion on differing perceptions of reality as a cause of conflict in an Internet con-text, it is apparent that their example illustrates a communication problem. In the survey they refer to, online customers responded that they were able to purchase products online that were not easily accessible to them otherwise. Thus, according to Coughlan et al. (2001), the manufacturer was not stealing sales from its bricks-and-mortar resellers, even though the bricks-and-mortar resellers probably thought it was. Hence, the goals of the companies were not in reality incompatible, even if the resellers perceived it that way. It should have been possible to pass the results from this survey on to the resellers in order to defuse the conflict situation. The failure to do so is nothing more than inadequate communication.

In their empirical study, Arndt and Ogaard (1986) researched differing perceptions of reality as a cause of channel conflict. They described perceptual divergence as follows: “Perceptions about environmental factors and bout the relative efficacy of using alternative elements of the marketing mix form the base for developing ac-tion in response to changes in the task environment. hence, incongruent percep-tions may cause incompatible behaviors causing conflict.” (p. 191). Although they found statistical support for the hypothesized relationship in all cases, their example item, “Sometimes I get the impression that the manufacturer and I are not getting through to each other” (p. 193), actually seems to measure inadequate communi-cation rather than perceptual divergence, as defined by the authors.

Ross and Lusch (1982) also examined differing perceptions of reality as a cause of conflict, but they failed to find statistical support for this relationship. Nevertheless, when describing what perceptual incongruity is, they argue that in a broker-wholesaler channel one critical issue is the extent to which brokers should provide wholesalers with assistance. They continue, “In deciding which assistances to pro-vide and emphasize, brokers must know how important the various assistances are to wholesalers. In so doing brokers will develop perceptions of which assistances are most important. At the same time, food wholesalers also have perceptions of the importance of these assistances. To the extent that these entities differ in their evaluation of the importance of these assistances, perceptual incongruities will be present.” (p. 243). However, it seems that these perceptual incongruities actually are caused by inadequate communication within the channel system.

Hence, what previous researchers (Stern & Heskett, 1969; Rosenberg & Stern, 1971; Ross & Lusch, 1982; Arndt & Ogaard, 1986; Coughlan et al., 2001) have la-


119

belled as different perceptions of reality seems to be a symptom of inadequate communication. Two studies could be found that hypothesized inadequate com-munication as a cause of channel conflict: Etgar (1979) found some evidence for the hypothesized relationship, but Anderson and Narus (1984) could not test the hypothesis because they had to drop all items designed to estimate communication.

One important aspect of communication is information sharing, which refers to the extent to which critical, often proprietary, information is communicated to one’s partner (Mohr & Spekman, 1994). Although some researchers have studied the content aspect of communication (e.g., Frazier & Summers, 1984), it will not be included in this study, since channel researchers usually treat content as an issue re-lated to influence strategies rather than to channel communication (cf. Li & Dant, 1997). However, it should be possible, without actually investigating the content of communication, to explore in more general terms whether one party perceives that the other party is keeping them well informed. In view of this consideration, the following definition is proposed:

Inadequate communication occurs when communication in marketing channels is improper in terms of information sharing, frequency, and relevancy.

Keeping in mind Mohr and Nevin’s (1990) statement that too much communica-tion can be just as bad as too little, the term improper was chosen since it implies that communication can be too excessive as well as too limited. The above discus-sion leads to the following hypotheses:

4.2 Channel Conflict n the literature, conflict has often been considered as a process composed of a series of conflict episodes (e.g., Pondy, 1967; Rosenberg & Stern, 1971). Pondy (1967), for instance, treats conflict as a dynamic process consisting of a

series of episodes that build upon one another, each episode including five stages: (1) latent conflict (the antecedent conditions); (2) perceived conflict (cognition); (3) feltconflict (affective state); (4) manifest conflict (overt behaviour); and (5) conflict aftermath(post-conflict conduct; either suppression or resolution). Pondy’s latent conflict is

I

H4: The level of inadequate communication among multiple marketing channels is posi-tively associated with the level of goal incompatibility.

H5: The level of inadequate communication among multiple marketing channels is posi-tively associated with the level of domain dissensus.

H6: The level of inadequate communication among multiple marketing channels is posi-tively associated with the level of channel conflict.

______ CHAPTER 4 ______________________________________________________________

120

similar to what in this thesis is labelled causes of channel conflict (see Section 4.1). Per-ceived conflict occurs when a channel member becomes aware of underlying incom-patibilities with another channel member, whereas felt conflict occurs when these in-compatibilities turn into feelings of hostility (Pondy, 1967). Manifest conflict refers to overt responses to perceived and/or affective conflict, i.e. when disagreements are communicated either verbally or in writing to the other channel member (Lusch, 1976b). The residual perceptions and feelings of a conflict episode form the conflictaftermath and become the latent state for future conflict episodes (Pondy, 1967). This means that each conflict situation will be interpreted by channel members based on the history of their relationship. If conflict frequently occurs in the chan-nel relationship, each new conflict situation will judged by a channel member as if its counterpart is incompetent, operates in bad faith and so on (Coughlan et al., 2001). One of the characteristics of a dysfunctional conflict is that, as time goes on, it is difficult to say what the conflict is about (Mack & Snyder, 1957). A positive relationship history, on the contrary, creates a positive future, since an incident of conflict will be de-emphasised or interpreted tolerantly (Coughlan et al., 2001).

However, as noted by Schmidt and Kochan (1972), a problem with the conflict-process approach is that it is very difficult to separate the various states of the proc-ess from one another. Brown and Day (1981) provide a possible solution to this di-lemma by treating channel conflict as a process similar to that proposed by Pondy (1967), but they examine only the manifest level of channel conflict. Even this ap-proach might pose a problem, since researchers are not always clear about what they have actually measured. Lusch (1976b), for instance, contends that although the measures used in his study were intended to measure manifest conflict, affective conflict might also have been measured, since resellers might have strong disagree-ments with, but never voice them to, the producer. In fact, he might even have measured some perceived conflict, since resellers might disagree with the producer without feeling hostility about the disagreement.

In addition, Gaski and Nevin (1985) claim that it is important to treat channel conflict according to its definition10: as a perceptually based construct. Bernard ((1949) ref. to in Gaski & Nevin, 1985) argues that this is because the conflict may exist for years in its latent form before it becomes a crisis, and that it therefore is a mistake to limit the study of conflict to its overt state. Therefore, it is channel members’ perception of channel conflict that should be studied (Gaski & Nevin, 1985). Indeed, several researchers have explicitly focused on perceived conflict in marketing channels (e.g., Rosenberg & Stern, 1971; Rosson & Ford, 1982; Eli-ashberg & Michie, 1984).

10 See Chapter 1, Section 1.4.1 for the definition of channel conflict used in this thesis.


121

Since it has been established that it is important to study perceived conflict (Ber-nard ((1949) ref. to in Gaski & Nevin, 1985)), resellers’ perceptions on conflict in marketing channels will be researched in this study.

4.2.1 Level of Channel Conflict Ong et al. (1990) conclude that, whereas causes are the underlying bases of conflict of which channel members may not ever be aware, conflict issues can be regarded as symptoms of these causes, which are prone to change over time. Consequently, the most common method to measure conflict is to first identify issues over which companies in a marketing channel might disagree, through a literature review and then pre-testing; thereafter, respondents have been asked to indicate the level of conflict on each of these issues (see Table 4.1 for an overview of such studies).

Table 4.1 Previous Measures of Marketing Channel Conflict

Conflict Dimension Studies that (1) identified pos-sible conflict issues and (2) measured level of conflict on those issues

Frequency(F)

Intensity(N)

Importance(I)

Satisfaction (S)

(Rosenberg & Stern, 1971) N (Lusch, 1976a) F (Lusch, 1976b) F (Brown, 1977) F I (Brown & Frazier, 1978) F I (Brown, 1979) F (Brown & Day, 1981) F N I (Schul, Lamb & Little, 1981) F (Brown, Lusch & Muehling, 1983) F (Eliashberg & Michie, 1984) F N (Dilts, 1985) F N (Arndt & Ogaard, 1986) F N I (Moore, 1990) F N (Brown et al., 1991) F S (Katsikeas & Piercy, 1991) F N (Katsikeas, 1992) F N (Cronin & Baker, 1993) F N I (Ganesan, 1993) N I (Price, 1993) F N (Lee, 2001) F

Measures of conflict level include the degree of conflict about an issue, i.e. intensity(Rosenberg & Stern, 1971); the frequency of disagreement about the issue (Lusch, 1976a, 1976b; Brown, 1979; Schul et al., 1981; Brown et al., 1983; Lee, 2001); the frequency and intensity of disagreement about the issue (Eliashberg & Michie, 1984; Dilts, 1985; Moore, 1990; Katsikeas & Piercy, 1991; Katsikeas, 1992; Price, 1993); the frequency and importance of the issue (Brown, 1977; Brown & Frazier, 1978); the intensity and importance of the issue (Ganesan, 1993); the frequency of dis-agreement and degree of satisfaction about the issue (Brown et al., 1991); and the

______ CHAPTER 4 ______________________________________________________________

122

frequency, intensity and importance of the issue (Arndt & Ogaard, 1986; Cronin & Baker, 1993).

Brown and Day (1981) conducted a study using the above method, where they measured conflict over a number of issues by the separate measures of frequency, in-tensity and importance. They asked resellers to recall discussions with the producer’s representative over 15 conflict issues during a certain year. Thereafter, the resellers were asked to estimate on an 11-point scale (0-10) the percentage of the discus-sions that pertained to each issue (frequency). Intensity was measured on a 7-point scale ranging from “not very intense” to “very intense”, and the importance of each issue to the reseller company’s overall profitability was measured on an 11-point scale, ranging from “very unimportant” to “very important”. When the re-searchers evaluated different combinations of these measures, they found that the measure combining frequency, intensity and normalised importance weights multiplicatively (F * N * I´) was superior to the other combinations. They also found three other measures that performed well: frequency multiplied by intensity (F * N), frequency plus intensity (F + N), and intensity alone (N).

Since (1) previous research suggests that including in the questionnaire all three conflict measures (i.e. frequency, intensity and importance) for each conflict issue might lead to respondent fatigue (e.g., Frazier et al., 1989); (2) Brown and Day (1981) conclude that the measures combining frequency and intensity perform well; and (3) Dilts (1985) argues that it is not necessary to include importance when meas-uring conflict, because the intensity of a disagreement is likely to reflect its impor-tance, the following definition is proposed for this study:

The level of channel conflict is a channel member’s perceptions of conflict with another channel member, in terms of frequency and intensity, over a representative set of issues.

4.3 Outcomesutcomes of channel exchange include economic performance and chan-nel members’ satisfaction (Brown et al., 1991). Since it is suggested that channel members’ satisfaction is related to a channel relationship’s long-

term viability (e.g., Hunt & Nevin, 1974), both performance and satisfaction are included in the conceptual model (see Figure 4.1, p. 126).

4.3.1 PerformanceA frequent assumption in the marketing literature is that conflict in marketing channels will have an impact on the performance of that channel (e.g., Brown & Day, 1981). However, among empirical studies on channel conflict (see Appendix 2for an overview) surprisingly few have actually addressed the relationship between channel conflict and performance (cf. Assael, 1969; Pearson, 1973; Lusch, 1976a;

O


123

Pearson & Monoky, 1976; Kelly & Peters, 1977; Rosson & Ford, 1980; Schul et al., 1981; Rosson & Ford, 1982; Anderson & Narus, 1984; Anderson & Narus, 1990; Cronin & Baker, 1993; Webb, 1997; Webb & Hogan, 2002; Duarte & Da-vies, 2003).

In addition, not all of the above studies researched how the level of channel con-flict impacts the performance of that channel. Instead, some have investigated the reverse relationship, i.e. how performance impacts the level of channel conflict. That is, Anderson and Narus (1984; 1990) studied how the current evaluation of past relationship outcomes impacts the level of channel conflict indirectly. Like-wise, Schul et al. (1981) hypothesized that channel members’ perceptions of past performance are an antecedent of conflict.

Duarte and Davies (2003), on the other hand, were interested only in the shape of the relationship between channel conflict and performance, the latter defined in terms of effectiveness and efficiency. They found that a linear model predicted ef-fectiveness best, whereas an S curve better described the relationship between con-flict and efficiency. However, they did not attempt to investigate the direction of the conflict-performance relationship.

Studies that investigated how the level of channel conflict impacts performance have yielded ambiguous results. Pearson (1973) failed to find a significant differ-ence in performance between channels characterised by conflict and channels char-acterised by cooperation. Webb (1997), on the other hand, did not research the di-rect relationship between the level of channel conflict and its impact on perform-ance; instead, he included a products’ life cycle stage (emergence, growth, and ma-turity-decline) as a moderator, but he failed to find statistical support for the rela-tionship. Nevertheless, when analysing the same data more thoroughly, Webb and Hogan (2002) found that the negative effects of frequency of conflict on channel performance increased in the maturity-decline stage of the product life cycle, rela-tive to the growth stage. They also studied the direct relationship between per-formance and conflict, and they found that the greater the frequency of channel conflict, the lower the level of channel performance. A few other studies also found at least some evidence that higher levels of conflict is negatively related to performance (cf. Lusch, 1976a; Pearson & Monoky, 1976; Kelly & Peters, 1977; Cronin & Baker, 1993). Similarly, Rosson and Ford (1980; 1982) found that the lower the level of channel conflict, the higher the level of performance.

As can be seen, the findings regarding the relationship between channel conflict and performance are rather indefinite. Reasons for this could be that the studies used very different, and sometimes even inappropriate, measures of performance. Pearson (1973), for instance, used three different performance measures, i.e. service level, turnover and adjusted turnover. However, as previously stated, he failed to

______ CHAPTER 4 ______________________________________________________________

124

find support for the conflict-performance relationship, and he concluded that two of the performance measures, service level and turnover, actually were contradic-tory. Lusch (1976a) used dealers’ return on asset and dealers’ asset turnover to measure operating performance. However, he concluded that these two measures were subject to error, since different dealerships use different accounting methods. Other measures of performance that have been used include the level of service provided by the wholesaler (Pearson & Monoky, 1976), distributors’/franchisees’ perceptions of their profitability compared to that of similar operations (Kelly & Peters, 1977), manufacturers’ real and perceived results from export relationship, i.e. trends in foreign market sales, satisfaction with foreign market sales, coopera-tiveness of the relationship, and satisfaction with the relationship (Rosson & Ford, 1980), payoff from using a target manufacturer (Cronin & Baker, 1993), and man-agers’ perception of channel system performance (Webb, 1997; Webb & Hogan, 2002).

A company’s performance is undoubtedly of greatest concern to managers. There-fore, this study will investigate how the level of channel conflict affects perform-ance. Day and Nedungadi (1994) found that most managers base their decisions on how they perceive the financial performance of their business, rather than on more objective measures of performance. Accordingly, perceptual measures of perform-ance will be used in this research. Based on the definition provided by Li and Dant (1997, p. 202), the following definition is proposed:

Perceived performance is a channel member’s overall evaluation and assessment of the out-comes achieved within an exchange relationship with another channel member.

Hence, the following hypothesis:

Cronin and Baker (1993) found that as the performance of a working relationship increases, the level of satisfaction increases. Hence, the following hypothesis is pro-posed:

4.3.2 SatisfactionChannel member satisfaction is important for several reasons. Perhaps most impor-tant is the belief that satisfied channel members will cooperate much better than dissatisfied ones, and thus be less likely to terminate the relationship (Hunt &

H7: The level of channel conflict among multiple marketing channels is negatively asso-ciated with the level of perceived performance.

H8: The level of perceived performance is positively associated with the level of perceived satisfaction with the channel system.


125

Nevin, 1974). Nevertheless, just like performance, satisfaction in marketing chan-nels has been measured in numerous ways.

Measures of performance and satisfaction have not always been differentiated. As noted in the previous section, Rosson and Ford (1980) included satisfaction with foreign market sales as one of several performance measures. But even when satis-faction has been measured as a separate construct, there is no consensus regarding its conceptualisation. For instance, satisfaction has been measured by how satisfied the parties were with the performance of the other party (Rosenberg & Stern, 1971; Wilkinson, 1981; Cronin & Baker, 1993), and by how satisfied channel members’ were with a negotiation with the other party (Ganesan, 1993). Several researchers have measured channel member’s satisfaction, or dissatisfaction, with the channel relationship or arrangement (Brown & Frazier, 1978; Schul et al., 1981; Anderson & Narus, 1984; Gaski & Nevin, 1985; Arndt & Ogaard, 1986; Frazier et al., 1989; Anderson & Narus, 1990; Price, 1993; Rawwas, Vitell & Bar-nes, 1997; Webb, 1997; Lee, 2001; Webb & Hogan, 2002; Leonidou, Barnes & Talias, 2006). Others have focused explicitly on channel members’ economic satis-faction, such as the parties’ satisfaction with their rewards (Dwyer, 1980), or the difference between channel members’ expectations and actual profits (Brown et al., 1991).

Since the main focus of this study is the addition of the Internet direct sales chan-nel, satisfaction with the channel arrangement will be measured in general terms. Accordingly, for this study the following definition, adapted from Gaski and Nevin (1985, p. 131), will be used:

Satisfaction is a channel member’s overall approval of the channel arrangement.

It is true that both channel conflict and satisfaction were included in Price’s (1993) and Gaski and Nevin’s (1985) studies, but the relationship between these constructs was never investigated. Brown and Frazier (1978) did not hypothesize a relation-ship between satisfaction and level of channel conflict directly, but they nonethe-less found that higher levels of conflict were related with lower levels of satisfaction in the channel relationship. Anderson and Narus (1984), on the other hand, hy-pothesized that the level of channel conflict would be negatively associated with the level of satisfaction. However, the measures that they used to measure the level of conflict and the level of satisfaction were not sufficiently distinct. Instead, they combined those two constructs to form a cooperation/satisfaction construct, which they used for model testing. Thus, their hypothesized relationship between level of channel conflict and satisfaction could not be tested.

Among those who actually have researched the relationship between channel con-flict and satisfaction, the direction of the relationship has varied. Some researchers

______ CHAPTER 4 ______________________________________________________________

126

have found that higher level of satisfaction leads to lower levels of channel conflict (Schul et al., 1981; Brown et al., 1991). But the main stream of channel conflict research has investigated the opposite direction of the relationship. That is, several researchers have found that as the level of conflict increases, the level of satisfaction decreases (cf. Dwyer, 1980; Wilkinson, 1981; Arndt & Ogaard, 1986; Frazier et al., 1989; Anderson & Narus, 1990; Cronin & Baker, 1993; Ganesan, 1993; Rawwas et al., 1997; Webb, 1997; Lee, 2001; Webb & Hogan, 2002; Leonidou et al., 2006). Thus, following the main stream of research on channel conflict, the fol-lowing hypothesis is proposed:

4.4 Research Model & Hypothesis Overview graphical presentation of the frame of reference helps to define what will, and will not, be studied (Miles & Huberman, 1994). Accordingly, the re-search model representing the above hypothesized relationships is provided

in Figure 4.1. In addition, a summary of the hypotheses is presented in Table 4.2.

Figure 4.1 Research Model

AGOAL

INCOMPATIBILITY

INADEQUATE COMMUNICATION

DOMAIN DISSENSUS

LEVEL OFCHANNEL CONFLICT

PERFORMANCE

SATISFACTION

H1+

H2+

H3+

H4+

H5+

H6+

H9-

H7-

H8+

H9: The level of channel conflict is negatively associated with perceived satisfaction with the channel system.


127

Table 4.2 Overview of Hypotheses

H.no.

Hypothesized Relationship

H1 The level of goal incompatibility among multiple marketing channels is positively associ-ated with the level of channel conflict.

H2 The level of goal incompatibility among multiple marketing channels is positively associ-ated with the level of domain dissensus.

H3 The level of domain dissensus among multiple marketing channels is positively associated with the level of channel conflict.

H4 The level of inadequate communication among multiple marketing channels is positively associated with the level of goal incompatibility.

H5 The level of inadequate communication among multiple marketing channels is positively associated with the level of domain dissensus.

H6 The level of inadequate communication among multiple marketing channels is positively associated with the level of channel conflict.

H7 The level of channel conflict among multiple marketing channels is negatively associated with the level of perceived performance.

H8 The level of perceived performance is positively associated with the level of perceived sat-isfaction with the channel system

H9 The level of channel conflict is negatively associated with the level of perceived satisfac-tion with the channel system.

______ CHAPTER 5 ______________________________________________________________

128

5 METHODOLOGYIn this chapter, a description of the methodology used to answer the research questions will be presented. First, the research purpose will be discussed, followed by the research approach, the research strategy, data collection, data analysis and, finally, issues of validity and reliability.

This chapter addresses the methodological issues of Study II. The methodological issues pertaining to the first study are presented in Chapter 2.3, p. 37.

5.1 Research Purpose air, Money, Samouel and Page (2007) discuss research purposes as ex-ploratory, descriptive, or causal. Exploratory research is used to study a little-understood phenomenon, to identify or discover central categories of

meaning (Marshall & Rossman, 1999) and/or when the goal is to develop hy-potheses and propositions for further research (Marshall & Rossman, 1999; Yin, 2003). Descriptive research is used when the purpose is to document and describe the phenomenon under consideration (Marshall & Rossman, 1999). Most surveys are descriptive (Hair et al., 2007). Causal research is used to test whether or not one event causes another. The cause must occur before the effect, the two con-structs must be related to each other, a theoretical explanation is needed, and other potential causes must be controlled or eliminated. When trying to establish cause and effect, an experiment is appropriate (ibid.). Since few empirical studies of con-flict in multiple marketing channels, including the Internet, have been conducted (see Chapter 4.1 for a further discussion on previous research), this research was to some extent exploratory. However, since existing theories within the channel con-flict area were used to describe resellers’ perspectives on conflict in multiple mar-keting channels including the Internet, the research purpose was also descriptive. In other words, the purpose was both exploratory and descriptive.

5.2 Research Approach he research approach can be either quantitative, qualitative, or a combination of the two (Sullivan, 2001). Quantitative research involves numbers, counts and measures of things, whereas qualitative research involves mainly words,

pictures, descriptions, or narratives (ibid.). Which approach to choose depends mainly on two factors: the state of knowledge of the research problem, and the re-searcher’s assessment of the nature of the phenomenon to be studied (Sullivan, 2001). When there is enough previous research on the phenomenon, constructs and hy-potheses can be developed. As for knowledge of the research problem, the first study (see Chapter 2) provides a good preliminary understanding of the research problem, which along with existing theories made it possible to formulate hypotheses. As for the phenomenon to be studied, i.e. resellers’ perspectives on conflict in multiple mar-keting channels with a focus on the Internet, it is valuable to include the views of

H

T

__________________________________________________________ METHODOLOGY ______

129

as many resellers as possible, since quantification provides much more objective measures of the phenomenon (Sullivan, 2001). Accordingly, the quantitative research approach was chosen for this study.

5.3 Research Strategyive major research strategies exist: experiment, survey, archival analysis, history and case study (Yin, 2003). All of them have different advantages, and they are chosen depending on the type of research question(s) asked, whether

control over behavioural events is needed, and/or whether the focus is on con-temporary events. It should be noted that the contrast between quantitative and qualitative evidence does not distinguish the various research strategies. In other words, case studies should not be confused with qualitative research, since they can include, or even be limited to, quantitative research (ibid.). The research questions of this study were formulated as how-questions. When research questions focus mainly on ‘why’- or ‘how’-questions, the best research strategy will probably be case studies, histories, or experiments (Yin, 2003). Since the focus was on channel conflict when companies added the Internet marketing channel, the research prob-lem of this thesis was contemporary, thus ruling out histories. Among the remain-ing strategies, experiments require control over the behaviour of the subjects, whereas case studies are preferred when the relevant behaviour cannot be manipu-lated (cf. Yin, 2003). Since the researcher had no influence over resellers’ perspec-tives on conflict in multiple marketing channels, and since this study aimed at ex-ploring and describing resellers’ perspectives on this phenomenon, the case study re-search strategy was judged to be most appropriate for this research.

A case study can be described as a research strategy that focuses on the understand-ing of the dynamics present in a single setting (Eisenhardt, 1989). The researcher can decide to conduct research on one single case or on multiple cases. A single-case study can be used, e.g., when the case is critical for testing a well formulated theory, or when the case is extreme or unique (Yin, 2003). Multiple-case studies involve, as the name implies, more than one case. The disadvantage of multiple-case studies is that they require more time and resources than single-case studies. The advantage, though, is that they provide more compelling evidence and are more robust (ibid.). Since this research was not dependent on a single unique case, and since multiple-case studies are judged to be more robust, a multiple-case study was chosen for this research.

5.4 The Empirical Investigation ata reduction can be described as an analysis that sharpens, sorts, focuses, discards, and organises data so that conclusions can be drawn (Miles & Huberman, 1994). It is a process that starts even before data is collected,

since data collection is guided by the researcher’s decisions on which cases to in-

F

D

______ CHAPTER 5 ______________________________________________________________

130

clude, the research questions, the conceptual frame of reference, approaches to data collection, and so on (ibid.). In the following sections, the sample selection proce-dure, questionnaire development, data collection, and response rate will be pre-sented, followed by a discussion of how the data were analysed.

5.4.1 SampleEven if one would like to, one cannot study everyone, everywhere, doing every-thing. Choices regarding whom to study, where, when, about what and why, place limits on the conclusions one can draw (Miles & Huberman, 1994). The sub-jects one chooses to look at, or talk with, form the sample (Chisnall, 1997). In the first study, the producers’ perspective was studied, whereas the present study aims at capturing the resellers’ perspective. Since few studies exist where both producers’ and resellers’ perspectives on channel conflict have been empirically researched (see Appendix 2 for an overview), it was decided to study resellers of the previously studied companies Ducati and SAS. These two companies are, as already stated, of polar type as suggested by Eisenhardt (1989). Ducati is a manufacturing company that makes motorcycles, whereas Scandinavian Airlines is a service business. In ad-dition, before adding the Internet sales channel, Ducati sold its products solely through resellers, whereas SAS sold direct-to-the-customer even before the addi-tion of the Internet sales channel, via its telephone sales channel (see Figure 2.7, p. 58).

5.4.2 Questionnaire Development The first study of this thesis provided a good preliminary understanding of the working relationships between the producers and their resellers. Therefore, the first study, along with the procedure suggested by Churchill (1979), guided the ques-tionnaire development. First, the domain of each construct was clearly defined in terms of what would be included or excluded. That is, since only resellers’ per-spective will be studied in the survey, the conceptual definitions provided in the previous chapter were transformed into operational definitions, see Table 5.1.Thereafter, the literature was searched to locate relevant scales, which provided a large pool of items for each construct. Based on the definition of the construct, items were selected from this pool to cover its domain as well as possible. If scales were not available or inappropriate, new ones were developed.

Before pre-testing, eight academics individually reviewed the questionnaire. This review led to some modifications of the instructions, the scaling of the items, as well as the items themselves. The questionnaire was pre-tested by a series of inter-views on a convenience sample of managers at companies that were resellers of SAS or Ducati. Managers at these companies were visited, and as they filled out the questionnaires they were asked to provide suggestions for improving the clarity of the instructions, the appropriateness of the scales, and the relevance of the items.

__________________________________________________________ METHODOLOGY ______

131

Table 5.1 Conceptual and Operational Definitions of the Constructs

Conceptual Area

Concept/ Construct

Conceptual Definition Operational Definition

GoalIncompatibility

The extent to which dif-ferent members of a mar-keting channel perceive that simultaneous goal ac-complishment is impossi-ble.

The extent to which resel-lers perceive that their goals and the focal producer’s goals cannot be accom-plished simultaneously.

DomainDissensus

Domain dissensus exists when channel members’ claims are incongruent in terms of (1) products to be offered, (2) the population to be served, and/or (3) the services to be rendered or functions to be per-formed.

The extent to which resel-lers perceive that their claims and the focal producer’s claims are incongruent in terms of (1) products to be offered, (2) the population to be served, and/or (3) the services to be rendered or functions to be performed.

Causes of ChannelConflict


Inadequate communica-tion occurs when com-munication in marketing channels is improper in terms of information shar-ing, frequency, and rele-vancy.

The extent to which resel-lers perceive that communi-cation with the focal pro-ducer is improper in terms of information sharing, fre-quency, and relevancy.

ChannelConflict

Level ofChannelConflict

Channel member’s per-ceptions of conflict with another channel member, in terms of frequency and intensity, over a represen-tative set of issues.

Reseller’s perceptions of dis-agreements, in terms of fre-quency and intensity, be-tween the reseller and the focal producer over a repre-sentative set of issues.

Performance

Channel member’s overall evaluation and assessment of the outcomes achieved within an exchange rela-tionship with another channel member.

Reseller’s overall evaluation and assessment of the out-comes achieved within the exchange relationship with the focal producer. Outcomes

SatisfactionA channel member’s over-all approval of the channel arrangement.

Reseller’s overall approval of the focal channel arrange-ment.

Before each additional interview, insights from the previous interview were incor-porated into the questionnaires; some items were dropped, others were changed, and others were added, until there appeared to be no significant ambiguities or problems with the questionnaire. Since the questionnaire would be sent to several countries, it was pre-tested with resellers in each of these countries: Norway, UK, US, Canada, and Italy. Translations and back-translations of the questionnaire into Norwegian and Italian were conducted by professional translators.

______ CHAPTER 5 ______________________________________________________________

132

The scale development for each construct included in this thesis (goal incompatibil-ity, domain similarity, inadequate communication, level of channel conflict, performance, andsatisfaction) is described in detail below. Unless otherwise stated, a 7-point Likert-type scale anchored by ’strongly disagree’ and ’strongly agree’ was used. In order to increase reliability, multiple items were developed for each construct. Some addi-tional items/constructs were also included in the questionnaire; they are beyond the scope of this thesis but were retained for future research. There were very few differences between the Ducati version and the SAS version of the questionnaire, except that in a few cases the wording of the items and the explanatory text dif-fered because of company/industry-related differences. As for the constructs in-cluded in this research, there were no differences between the two versions of the questionnaire. Accordingly, only the English version of the SAS questionnaire is presented (see Appendix 4). The items included in this research are presented in Table 6.1, p. 145.

Goal Incompatibility Even though goal incompatibility has been widely recognized as a cause of con-flict, no relevant scales could be found. Brown and Day (1981), for instance, in-cluded goal incompatibility in their study but used only one item to measure it. Furthermore, although John and Reve (1982) thoroughly described how their scale items were developed, following the procedure suggested by Churchill (1979), their actual research instrument was not included in their paper. The only information available is that the scale consisted of 3 items, and that Cronbach’s was 0.55. Because of the low , no attempts were made to obtain the original scale. Instead, items were adapted from Brown and Day (1981), Arndt and Ogaard (1986), Mohr, et al. (1996), Webb and Hogan (2002), and Webb (1997). During field interviews respondents expressed a lack of precise knowledge about the pro-ducers’ goals, so ‘we perceive’ was added to three of the items, e.g., ‘we perceive that the goals of (producer’s name) are compatible with the goals of our company.’

Domain Dissensus Items were adapted from various sources, i.e. Etgar (1979), Rueckert and Orville (1987), Anderson and Weitz (1992), Webb (1997), and one item was of the re-searchers’ own construction. Some of the items were rephrased to capture the do-main dissensus derived from the producer’s direct-to-the-customer Internet sales.

Inadequate Communication Items were adapted from McQuiston (2001), Lusch and Brown (1996), and Anderson and Weitz (1992). Several were judged to be double-barrelled, and were therefore divided into two separate items.

__________________________________________________________ METHODOLOGY ______

133

Level of Channel Conflict In the first study, it was found that the term ‘conflict’ was interpreted by the re-spondents as very strong and negatively. Thus, in line with previous research (e.g., Katsikeas, 1992), the term ‘disagreement’ was used in the questionnaire.

Several studies in which possible conflict issues have been identified to measure the level of channel conflict have been conducted with resellers of manufactured prod-ucts (e.g., Lusch, 1976b; Brown & Day, 1981). Although these conflict issues per-haps could be adapted for the Ducati reseller survey, it would be impossible to use those items in the SAS reseller survey, since SAS is a service company and most conflict issues in those studies were goods related. Instead, the more general con-flict issues provided by Katsikeas and Piercy (1991) and Katsikeas (1992) were used as a starting point. In addition to these items, items on sales commissions, pro-ducer’s current direct-to-the-customer Internet sales, and producer’s direct-to-the-customer Internet sales when producer introduced that sales channel, were in-cluded. During pre-testing, one item regarding the producer’s sales organisation was added. Finally, an open-ended item (‘Other, please specify _____’) gave re-spondents the opportunity to add an optional conflict issue. The frequency and in-tensity of disagreements were measured separately for all items, using 7-point Likert-type scales. The frequency scale options ranged from ‘very infrequently’ to ‘very frequently’, whereas the intensity scale options ranged from ‘very weak’ to ‘very strong’.

Performance As could be seen in the previous chapter, performance has been measured in many different ways in studies of channel conflict. None of these measures were judged to be appropriate for this research. Instead, perceived performance was measured in terms of relationship performance and sales performance. The relationship performance scale was adapted in full from Skarmeas, Katsikeas and Schlegelmilch (2002) and refers to the extent to which the reseller views the focal relationship as successful. The scale has a reported Cronbach’s of .80 (cf. Skarmeas et al., 2002).

Some researchers stress the importance of measuring the performance of a com-pany relative to other companies (e.g., Achrol & Etzel, 2003). However, previous research has found that people make comparative judgments based on expectations, i.e. outcomes that are worse than expected are rated below the expectation refer-ence point, and those that are better than expected are rated above the reference point (cf. Oliver, 1980). In addition, Day and Nedungadi (1994) found that most managers use their perceptions of the financial performance of their business as the basis for decisions, rather than more objective measures. Considerations about competitors, as well as other important factors, are reasonably included in manag-ers’ expectations of sales performance. Accordingly, the items in the sales perform-

______ CHAPTER 5 ______________________________________________________________

134

ance scale were developed with regards to different aspects of the sales resulting from the focal relationship, compared to expectations.

SatisfactionThe intention was to capture the reseller’s perception of the whole channel system: the system between the reseller and the producer, between the reseller and the end customer, and between the producer and the end customer. Since no appropriate scales could be found, one item was adapted from Brown and Frazier (1978), and the remaining items were the researchers’ own.

5.4.3 Data Collection and Response Rate The data collection procedure for each case will be described in detail below. In both cases the questionnaire (see Appendix 4) was delivered by mail, accompanied by a signed cover letter and a pre-paid response envelope. Although an identifica-tion number was put on the questionnaire, the cover letter guaranteed respon-dents’ anonymity (see Appendix 5 for a sample cover letter).

DucatiFirst, the questionnaire was translated and back-translated into Italian and Norwe-gian by professional translators. Thereafter, addresses to resellers in Canada, Great Britain, Ireland, Italy, Norway, Sweden and United States were obtained through Ducati’s “Dealer Locator” service at the ducati.com web site. In total, 347 ad-dresses to resellers of Ducati were obtained. However, some of these resellers had been selling Ducati for only a few months and were therefore deleted. Two resel-lers in the French part of Canada provided only French versions of their web sites, and since the questionnaire intended for Canada was in English, these were also deleted from the list. These deletions left 332 Ducati resellers to whom question-naires with cover letters were mailed. The letter was addressed to the Presi-dent/CEO. Two weeks later, follow-up cards to non-respondents were sent. Only a few resellers responded, and therefore personal telephone calls were made to all resellers in Sweden, Norway, Great Britain and Ireland, followed by a second mail-ing of the questionnaire to these resellers. Because of the limited time available, re-sellers in Canada and the US received only a second mailing of the questionnaire and no telephone call.

A total of 44 questionnaires was returned, but one of those questionnaires was not filled in and had to be discarded. Since 5 questionnaires were returned as undeliv-erable, the eligible sample was 327, rendering an effective response rate of 13.1%. The major reason for this rather low response rate is that the questionnaire was dis-tributed in April-May 2007, the peak season for the motorcycle industry (a fact commented upon by several resellers). Another reason is that the response rate in Italy was very low; although the questionnaire was translated (and back-translated) by professionals into Italian, only 5.1% of the 137 Italian resellers returned the

__________________________________________________________ METHODOLOGY ______

135

questionnaire. If Italy had been excluded from the sample, the response rate would have increased to 18.9%. Nevertheless, in a recent review of research in industrial marketing channels, Woodside and Wilson (2003, p. 494) found that typical re-sponse rates for mail surveys range from 8 to 30 percent, so although the effective response rate of Ducati is rather low (13.1%), it is within the typical range.

SASFirst, the questionnaire was translated and back-translated into Norwegian by pro-fessional translators. It proved to be a challenge to obtain addresses of SAS resellers. Since the two persons who were involved in the first study had left the company, their successor, i.e. the director of sales, was contacted. He, however, was not will-ing to participate in the study. Therefore, addresses of Swedish travel agencies were purchased from PAR, and addresses for Norwegian travel agencies were purchased from CDG Sandberg A/S. In both cases, the addresses seemed somewhat dubious; several were of private persons with no company name, some company names in-dicated that the companies were active in an industry other than the travel industry (e.g., Bay Horse & Equipment), some were duplicates, and some even represented SAS competitors (e.g., Air France). It was therefore decided that in order to qualify as a serious travel agency, a company address at least had to appear in the respective country’s “yellow pages” online service (i.e. gulasidorna.eniro.se and www.kvasir.no/firmasok). The company also had to be labelled as a travel agency, or something similar, in the yellow pages. Even if a company was listed as a travel agency, it was sometimes obvious that it did not sell airline tickets (e.g., Härjedals Buss, a bus travel company), and such addresses were also excluded. After the ad-dresses were subjected to the above criteria, their number decreased from 1237 to 667 for the Swedish travel agencies, and from 688 to 329 for the Norwegian.

Even after all this, several addresses appeared ambiguous. Therefore, in an attempt to improve the response rate, respondents were asked in the cover letter to send back the questionnaire even if they had never sold SAS, with a note of this fact, since they then would avoid receiving remainders. When contact names were not available, the letter was addressed to the President of the company (“till VD” or “til daglig leder”). Two weeks after the mailing, reminder cards were sent to non-respondents, which yielded a total of 54 Norwegian and 165 Swedish responses. Out of these, 47 (7 Norwegian and 40 Swedish) were travel agencies reporting that they did not sell SAS. In addition, 16 Norwegian and 5 Swedish questionnaires were returned as undeliverable. Hence, the final sample for the Norwegian travel agencies was 306. 47 usable responses rendered an effective response rate of 15.4%. Two Swedish questionnaires were discarded as incomplete. With a final sample of 622 travel agencies and 123 usable responses, the effective response rate was 19.8% for the Swedish travel agencies. For Norway and Sweden combined, the effective response rate for travel agencies was 18.3%.

http://www.kvasir.no/firmasok

______ CHAPTER 5 ______________________________________________________________

136

Although a lot of effort was spent during the process of refining addresses to travel agencies, this approach also had its advantages. Quite soon after the mailing, several travel agencies got in touch with the researchers by phone, e-mail, or letter. A number of those travel agencies told the researchers that they were not accredited by IATA11 and explained the difference. Non-IATA travel agencies did not have access to SAS’ extranet, sassalesinfo.com, since the IATA number functions as the entrance code to sassalesinfo.com. According to the non-IATA travel agencies, some years ago they were treated the same as the IATA accredited agencies. For instance, when IATA travel agencies still received commission on sales, so did non-IATA travel agencies. Today, when non-IATA travel agencies book tickets with SAS they have to pay the same fee for each ticket as do private persons, which make them quite angry with SAS. As one respondent who had had his own travel agency for forty years put it, “I am doing SAS a favour, but instead of re-warding me, they are charging me.”

There were several reasons why the non-IATA travel agencies did not want to join IATA. For instance, a quite long IATA education is required. Since several man-agers of non-IATA travel agencies were close to retirement, they found this unfea-sible. Although it is possible to circumvent this requirement by employing person-nel with an IATA education, there are additional requirements for IATA member-ship, such as premises according to certain requirements. Since several non-IATA travel agencies were home-based, they did not fulfil these requirements. A third is-sue is that IATA accreditation requires a banker’s guarantee, which of course costs money. Since pre-testing did not reveal that these different types of travel agencies existed, the questionnaire did not capture whether travel agencies were IATA ac-credited or not. Thus, non-IATA travel agencies were included in the sample, which would not have been the case had the addresses been obtained from SAS. This research, therefore, provides a more complete picture of travel agencies’ views of SAS than was expected.

5.4.4 Data Analysis First, selected items were reverse coded (see Table 6.1, p. 145 for an overview of reverse coded items). Thereafter, a t-test of the two SAS samples (Norway and Sweden) was conducted, which led to the conclusion that the whole SAS sample could be used in the subsequent analyses. Thereafter, a t-test between the two samples (SAS and Ducati) was conducted. Since the t-test revealed that the two samples were significantly different on more than half of the items, they were ana-lysed separately.

11 For further information on IATA accreditation, see http://www.iata.org/ps/certification/accreditation-travel/index.htm

http://www.iata.org/ps/certification/accreditation-travel/index.htm

__________________________________________________________ METHODOLOGY ______

137

Normality Analysis The mean is one of the most commonly used measures of central tendency for metric scales (Hair et al., 2007, p. 317). But in order to get a better picture of the sample distributions, histograms of each item, the range, minimum and maximum, standard deviation, skewness, kurtosis and Kolmogorov-Smirnov’s Z test of nor-mality were also looked at. Since seven point Likert-type scales were used, standard deviations between 1.0 and 3.0 were considered medium (cf. Hair et al., 2007, p. 320). Skewness values larger than + 1 and kurtosis values larger than + 3 were de-fined as indicators of a non-normal distribution (cf. Hair et al., 2007, p. 321). This information is provided for each sample separately (see Table 6.1, p. 145).

Missing Data Little’s MCAR test was used to test whether data were missing completely at ran-dom (MCAR) (Little, 1988). The required minimum of p > .05 for data to be MCAR (cf. Hair, Black, Babin, Anderson & Tatham, 2006, p. 68), was used as the reference value. The Expectation-Maximization (EM) method (cf. de Vaus, 2002, p. 69) was used to substitute missing data.

Measure Validation As previously stated, several new scales were developed. Since exploratory factor analysis is useful as a preliminary analysis when the theory about the relations of in-dicators is not detailed enough (Gerbing & Anderson, 1988), exploratory factor analyses, using principal components with Varimax rotation, were conducted (see Chapter 6.4). Since a sample size of 50 is needed to conduct a factor analysis (cf. Hair, Anderson, Tatham & Black, 1998, p. 98), the exploratory factor analysis was conducted only with the SAS data.

In order to decide the number of factors that should be extracted in a factor analy-sis, a combination of several techniques is usually used (Chatterjee, Jamieson & Wiseman, 1991). The latent root criterion, i.e. considering only factors with eigen-values greater than one, is the most commonly used technique (Hair et al., 1998, p. 103). However, when the number of items is below 20, the latent root criterion might lead to too few factors (ibid.). An additional technique is therefore to iden-tify the amount of variance explained, i.e. the factors extracted should account for some appreciable percentage of the total variability in the original data (Chatterjee et al., 1991). In the social sciences, solutions that explain 60% of the total variance are usually considered satisfactory (Hair et al., 1998). This procedure was used as a starting point in each exploratory factor analysis, besides for the level of channel con-flict items (see Chapter 6.4.2 for a further discussion).

Other criteria also needed to be met. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy should exceed the required value of .50 for a factor analysis, Bartlett’s test of sphericity should be significant, the communalities should be above the

______ CHAPTER 5 ______________________________________________________________

138

recommended .50, and the measure of sampling adequacy (MSA) should be accept-able, i.e. .50 or above (Hair et al., 1998). In all cases, KMO was between .84 and .91, i.e. values considered “meritorious” or “marvellous” (cf. de Vaus, 2002, p. 137), and MSA was at least “middling”, i.e. .70 or above, and in several cases “meritorious”, i.e. .80 or above (cf. Hair et al., 1998, p. 99).

After an exploratory factor analysis has been conducted, a confirmatory factor analysis is needed to evaluate and likely refine the scales with regards to unidimen-sionality (Gerbing & Anderson, 1988). Scales with more than three retained indi-cators were therefore submitted to a confirmatory factor analysis. This procedure led to the identification of two scales that were multidimensional (see Chapter 6.4.4).

Hypothesis Testing The original intention was to use Structural Equation Modelling (SEM) with AMOS 5.0 software to test the research model, using data from both cases. How-ever, since a minimum of 5 observations per estimated parameter is needed (Hair et al. 1998, p. 604), the research model requires a sample size of at least 115 re-sponses, even if summated scales are used. Accordingly, only the SAS data could be analysed with SEM. Due to the limited sample size for Ducati (n = 43), multiple regression analysis was used to test the hypotheses for this case.

Even so, it was judged important to be able to compare the main hypothesized re-lationships, i.e. the relationships including the overall level of channel conflict, and the overall performance, across the two samples. Accordingly, for the Ducati data, the three level of channel conflict scales were summated and averaged, into one scale measuring overall level of channel conflict. This procedure was also followed for the two performance scales, because comparability of the two samples was con-sidered more important than the dimensionality of the scales. As Iacobucci (2001, pp. 61-62) states, “unidimensionality itself is sometimes less important than other theoretical goals that may motivate the use of multifaceted scales”.

First, SEM was used to test the outcomes part of the research model. The model fit the data very well. However, when testing the whole research model, including the antecedents, the fit statistics suggested a poor fit to the data. Anderson and Gerbing (1988) state that research models often need some kind of re-specification. Byrne (2001, p. 88) notes that AMOS produces two types of information that can be useful in detecting model misspecification, the standardised residual values and the modification indices. When standardized residual values exceed 2.58, misfit is possible. But since the residuals are not independent of one another, one should not attempt to test them. Because the modification indices reflect the extent to which the model is appropriately described (ibid.), the magnitude of the standard-ized residuals was used to control for possible model misspecification, whereas the

__________________________________________________________ METHODOLOGY ______

139

modification indices were used to re-specify the model. Since it is important that re-specifications are based on theoretical and content considerations, and not just statistical ones (Anderson & Gerbing, 1988), all re-specifications were justified by theory or content. As a result, two post hoc hypotheses were added.

The re-specified research model was used to test the hypotheses in both cases, i.e. by SEM for SAS and by multiple regression for Ducati. In order to assess the strength of association between variables, the reference values provided by de Vaus (2002, p. 272) were used (see Table 5.2).

Table 5.2 Strength of Association between Variables (adapted from de Vaus, 2002, p. 272)

Coefficient Range Strength of Association 0.01 - 0.09 Trivial 0.10 – 0.29 Low to moderate 0.30 – 0.49 Moderate to substantial 0.50 – 0.69 Substantial to very strong 0.70 – 0.89 Very strong 0.90 – 0.99 Near perfect

AMOS, as well as SPSS, reports only the results of two-tailed significance tests. Since the direction of all hypotheses was specified, the significance values were converted to one-tailed. The AMOS output provides the critical ratio (C.R.), which is the same as a t-value (cf. Byrne, 2001, p. 241 and Hair et al., 1998, p. 206). To be consistent, only t-values were used in this research.

Test of Model Fit A number of indices exist to test how well data fit a model. These can be classified into three different types: (1) absolute fit measures, (2) incremental fit measures, and (3)parsimonious fit measures (Hair et al., 1998). Researchers are advised to employ one or more of each type to test the model. Absolute fit measures assess the degree to which the overall model predicts the observed covariance or correlation matrix, whereas incremental fit measures check that the proposed model exceeds the null model. Finally, parsimonious fit measures assess whether a good fit has been achieved by over-fitting the data with too many parameters (ibid.). Absolute fit measures included in this research were the likelihood-ratio chi-square ( 2), including degrees of freedom (df) and significance value (p), the goodness-of-fit index (GFI), and theroot mean square of approximation (RMSEA). The incremental fit measure was thecomparative fit index (CFI), and the parsimonious fit measure was the normed chi-square value ( 2/df).

Other commonly used indices include the adjusted good-ness-of-fit index (AGFI) andthe normed fit index (NFI), but since they may underestimate the fit for small sam-

______ CHAPTER 5 ______________________________________________________________

140

ples (cf. Byrne, 2001, p. 82-83) they were deemed inappropriate. On the contrary, CFI and RMSEA are among the measures least affected by sample size (ibid).

There is no threshold value for GFI, but most researchers suggest that .90 or greater should be used to indicate good fit (Hair et al., 2006); 2/df should be greater than 1.0 and smaller than 3.0; p should be non-significant at least at the > .05 level (Hair et al., 1998); CFI should be .90 or higher (Byrne, 2001); and RMSEA should preferably be .08 or less, with an upper threshold of .10 (Hair et al., 2006). The reference values used in this research are summarized in Table 5.3.

Table 5.3 Reference Values to Test Model Fit

Index Term Type of Measure Reference value Normed chi-square value 2/df Parsimonious fit 1 < x < 3

Goodness-of-fit index GFI Absolute fit x > .90

Comparative fit index CFI Incremental fit x > .90

Root mean square of approximation RMSEA Absolute fit x < .10

5.4.5 Issues of Validity and Reliability The research design is the logical sequence that connects the research questions, the empirical data, and the conclusions (Yin, 2003). Therefore, the case study in-vestigator must maximise the quality of the design, i.e. the validity and reliability (ibid.). Validity refers to how well a specific research method measures what it is supposed to measure (Chisnall, 1997). There are three kinds of measurement valid-ity: content validity, construct validity, and criterion validity (Hair et al., 2007). Reliabil-ity, on the other hand, refers to how consistent the measures are (Potter, 1996).

Content or Face Validity Content or face validity refers to whether the scale measures what it is supposed to measure (Hair et al., 2007). The extended literature review, along with the first study, provides a good understanding of the constructs to be included in this re-search. This understanding provided the basis for the operational definitions that were developed for each construct (see Section 5.4.2), which in turn were used as benchmarks during scale development. As described in Section 5.4.2, several items used by previous researchers were judged to be double-barrelled, and were there-fore split into separate items. The pre-test indicated that the content validity of the scales was satisfactory. The exploratory factor analysis revealed that one item, which originally was thought to measure goal incompatibility, instead loaded on the domain dissensus factor. Since the item conformed to the definition of domain dissensus, it was concluded that face validity was further increased by retaining the item (see Chapter 6.4.1 for a further discussion).

__________________________________________________________ METHODOLOGY ______

141

Construct Validity and Common Method Variance Construct validity refers to establishing correct operational measures for the concepts being studied (Yin, 2003). There are two kinds of construct validity: convergent va-lidity and discriminant validity (Hair et al., 2007). Convergent validity refers to the extent to which one construct is positively correlated with other measures of the same construct, whereas discriminant validity refers to the extent to which the construct does not correlate with other measures that are different from it. Common method variance, on the other hand, refers to bias caused by the measurement method rather than by the measures themselves (Podsakoff, MacKenzie, Lee & Podsakoff, 2003).

First, Harman’s single-factor test (Podsakoff & Organ, 1986) was applied to the SAS sample to test for discriminant validity and common method vari-ance/common method bias (see Table 6.7, p. 158). That is, all items were entered into an exploratory factor analysis. Since the sample size was above 150 but below 200, factor loadings below .45 were suppressed (cf. Hair et al., 1998, p. 112). All cross-loadings were below the suppressed value of .45, which indicated that the discriminant validity of the scales was satisfactory. With this technique, it is assumed that common method variance is present if one factor in the un-rotated solution accounts for the majority of the total variance (Podsakoff & Organ, 1986). Total variance explained was almost 69%. The largest factor in the un-rotated solution explained just above 31% of the total variance. (In the rotated solution the largest factor explained just above 18% of the total variance). It was therefore concluded that the common method variance was low.

Second, as suggested by de Vaus (2002, p. 30), a bivariate Pearson correlation ma-trix was created to test both discriminant and convergent validity (see Table 6.8, p. 159). In all cases, the correlations were highest among items intended to measure the same construct. In all cases, correlations were lower (or at least not higher) with items that were intended to measure other constructs. Hence, both discriminantand convergent validity were judged to be satisfactory.

Criterion Validity Criterion validity refers to whether a construct performs as expected relative to other constructs. It can be determined by testing for predictive validity and concurrent valid-ity (Hair et al., 2007). Since this study was conducted at one point in time, only concurrent validity is relevant. Concurrent validity refers to theoretically estab-lished, pre-specified relationships between independent and dependent variables or constructs. The scores on these constructs should be highly correlated (ibid.). All relationships between constructs in the research model were based on theory. Ac-cordingly, concurrent validity was established by the structural equation modelling of the SAS data and the multiple regression analyses of the Ducati data.

______ CHAPTER 5 ______________________________________________________________

142

ReliabilityReliability refers to whether the findings and conclusions would be the same if an-other researcher conducted the same study12 (Yin, 2003). Reliability also concerns minimising errors and biases (ibid.). Chronbach’s was calculated for all resulting scales, for each sample separately (see Chapter 6.4.5). In order to assess the reliabil-ity of two-item scales, the Spearman Brown formula, i.e. split-half reliability, was also used (cf. Hulin, 2001). The Spearman Brown calculations changed the reliabil-ity from .68 to .69 for the Internet-related conflict scale used with SAS; otherwise, the figures were exactly the same as the Cronbach’s ’s calculated using SPSS. Most of the scales were highly reliable. However, in the case of SAS, three two-item scales had reliabilities below the generally agreed upon .70 limit, but well above the .60 limit for moderate strength of association (cf. Hair et al., 2007, p. 244).

Multicollinearity In regression analysis, researchers must check for multicollinearity, since highly col-linear items can distort the results substantially or make them unstable, and thereby impossible to generalize (Hair et al., 1998, p. 208). Multicollinearity refers to the correlation among independent variables (Hair et al., 2007). One test for multicol-linearity is to check the variance inflation factor (VIF) (Hair et al., 1998, p. 208). Accordingly, VIF values were determined for each multiple regression. Since all VIF values were rather close to 1.0, and nowhere near the threshold level of 10.0 (cf. Hair et al., 1998), it was concluded that multicollinearity did not adversely af-fect the regression analyses.

Non-response BiasIf respondents differ substantially from non-repondents, non-response bias occurs (Armstrong & Overton, 1977). One way to estimate non-response bias is to sample non-respondents (ibid.). In the case of Ducati, telephone calls were made to non-respondents in Sweden, Norway, Ireland and Great Britain. As previously stated, these telephone calls revealed that resellers were extremely busy, since the ques-tionnaire was distributed during spring, the time of the year when motorcycle dealers sell the majority of their motorcycles. Hence, resellers said that they had to prioritize business activities but would respond to the questionnaire if they were able to find time. Several of them did. Although a t-test can be used to compare late respondents with early respondents (Armstrong & Overton, 1977), this was judged as meaningless, because of unknown differences in mail delivery time to, within and from the various countries, and because the sample size for each coun-try was so limited.

12 In order to facilitate replication of this research, a goal of this thesis was to describe how each step of the research process was carried out. All material related to the case studies is organised, docu-mented and stored in a manner that enables others to retrieve it efficiently at some later date.

__________________________________________________________ METHODOLOGY ______

143

Since the mail delivery time was also different between the Norwegian SAS sample and the Swedish SAS sample, it was decided to conduct separate non-response bias analyses for each. Early and late respondents were compared, since early respon-dents are likely to be more interested in the subject, whereas late respondents are more likely to be similar to non-respondents (cf. Armstrong & Overton, 1977). Accordingly, for the Swedish sample it was decided to compare the first quarter re-spondents (n = 30) with the last quarter respondents (n = 30) by an independent samples t-test on the retained items (see Table 6.6, p. 155 for an overview of re-tained items). Of the 25 compared items, only 1 item (x6q4) was found to be sig-nificantly different between early and late respondents at the p < .05 level; early re-spondents reported a higher mean (3.73 vs. 2.50, p = .019). For the Norwegian sample, the total number of usable responses was 47. Given the limited sample size, it was decided to compare the first 23 responses with the last 24 responses. Again, only 1 item (x8g) was significantly different between early and late respondents at the p < .05 level; early respondents reported a lower mean (2.43 vs. 3.29, p = .026). As only one item out of 25 was significantly different in each sample, and since that item differed between the samples, it was concluded that non-response bias did not extensively affect the results.

______ CHAPTER 6 ______________________________________________________________

144

6 RESULTS & ANALYSISIn this chapter, the results of the survey among resellers of Ducati and SAS are analyzed and presented. Development of the questionnaire and a more detailed explanation of the data analysis procedures were presented in the previous chapter. First, descriptive statistics are pre-sented, followed by measure purification and validation. Thereafter, results from the testing of the research model and the hypotheses are reported for each sample.

6.1 Descriptive Statistics or reasons explained in the methodology chapter, see Chapter 5.4.3, p. 134, the Ducati sample was relatively small, i.e. it consisted of 43 cases. Out of these, 6 were Swedish, 1 was Norwegian, 3 came from Great Britain, 23

came from the US, 3 were Canadian, and 7 were Italian. Since the sample was rather small, descriptive statistics for the whole Ducati sample will be presented, rather than for each country separately (see Table 6.1).

The SAS sample consisted of 123 Swedish cases and 47 Norwegian cases. A t-testrevealed nine out of 48 items with significant differences at the p < .05 level be-tween SAS Sweden and SAS Norway. Since none of these items was retained after scale purification (see Table 6.6, p. 155 for an overview of retained items), it was decided to use the whole SAS sample in the forthcoming analyses. Hence, only de-scriptive statistics for the whole SAS sample are provided (see Table 6.1).

Since this research consisted of two samples, an independent samples t-test of the equality of means was conducted between the Ducati and SAS samples, for the items that were used to measure the constructs included in this research. The t-values and the two-tailed levels of significance are presented in Table 6.1. Since Levene’s test was significant at the p < .05 for 9 out of 48 items, and since the sample sizes were unequal, the reported t-values do not assume equal variances (cf. Green & Salkind, 2003, p. 155). Non-significant (n.s.) means p > .05.

Sample mean (M), standard deviation (S.D.), skewness and kurtosis values are also reported in Table 6.1. Items that were substantially skewed, or too peaked, are highlighted with grey background. For a further discussion of skewness and kurto-sis, see Section 6.2, p. 148. Items that are marked with ® were reverse coded. Note that the original items were not necessarily worded in exactly the same way as in the tables. For the original items, see Appendix 4.

F

______________________________________________________ RESULTS & ANALYSIS ______

145

Table 6.1 Descriptive Statistics – Ducati and SAS*

Scale/Items Sample n Min

Max

M S.D. t(p,2-tail)

Skewness

Kur-tosis

Inadequate Communication:Ducati 43 2 7 3.86 1.49 .70 -.20 x3a Producer has a good

understanding of the type of information reseller needs ® SAS 169 1 7 4.26 1.53

-1.56 (n.s.) -.00 -.55

Ducati 43 2 7 3.98 1.58 .57 -.62 x3b Producer has a good understanding of the frequency of information reseller needs ® SAS 169 1 7 4.43 1.55

-1.69 (n.s.) -.09 -.58

Ducati 43 1 7 3.67 1.81 .39 -.66 x3c Producer keeps reseller well informed about relevant events and changes ® SAS 169 1 7 4.50 1.63

-2.73 .008 -.07 -.99

Ducati 43 1 7 3.09 1.62 .91 .54 x3d Producer keeps reseller well informed about their products ® SAS 169 1 7 4.07 1.66

-3.53 .001 -.03 -.79

Ducati 43 1 7 4.12 1.80 .23 -.96 x3e Producer keeps reseller well informed about what is going on in producer’s company ® SAS 170 1 7 4.79 1.49

-2.26 .028 -.18 -.81

Domain Dissensus:Ducati 42 1 7 3.40 2.15 .43 -1.28 x4a Producer has given reseller an

exclusive territory for their products ® SAS 169 1 7 6.53 1.27

-9.03 .000 -3.02 8.78

Ducati 43 1 7 3.19 1.59 .50 -.14 x4b Producer’s and reseller’s respective duties are well defined ® SAS 169 1 7 4.50 1.91

-4.64 .000 -.23 -1.05

Ducati 43 1 7 3.02 1.42 .63 .11 x4c Sometimes producer tries to take over reseller’s activities SAS 168 1 7 3.88 2.16

-3.12 .002 .02 -1.42

Ducati 40 1 7 5.05 1.65 -.59 -.11 x4d Producer’s direct-to-the- customer Internet sales serves the same customers as reseller SAS 170 1 7 5.85 1.69

-2.76 .008 -1.44 .97

Ducati 40 1 7 5.15 1.73 -.83 .00 x4e Producer’s direct-to-the- customer Internet sales markets identical products as reseller SAS 170 1 7 5.67 1.75

-1.71 (n.s.) -1.34 .79

Ducati 41 1 7 4.71 2.05 -.50 -1.09 x4f Reseller competes with producer’s direct-to-the- customer Internet sales SAS 170 1 7 6.21 1.53

-4.39 .000 -2.19 4.00

Goal Incompatibility: Ducati 43 1 7 3.09 1.58 .89 .32 x5a Reseller’s and producer’s goals

are compatible ® SAS 168 1 7 5.07 1.53

-7.33 .000 -.45 -.62

Ducati 43 1 7 3.65 1.66 .52 -.70 x5b Reseller and producer share same view on how to compete in reseller’s local market ® SAS 167 1 7 5.37 1.50

-6.17 .000 -.73 -.26

Ducati 43 1 7 3.37 1.65 .64 -.55 x5c It is feasible for both reseller and producer to exceed their respective goals ® SAS 169 1 7 4.10 1.80

-2.54 .013 .09 -.95

Ducati 43 1 7 3.86 1.55 .16 -1.03 x5d Reseller disagrees with producer on how to achieve their respective goals SAS 168 1 7 3.98 1.98

-.41 (n.s.) .04 -1.14

______ CHAPTER 6 ______________________________________________________________

146


Max

M S.D. t(p,2-tail)

Skewness

Kur-tosis

Ducati 40 1 7 4.72 1.78 -.53 -.70 x5e Producer’s direct-to-the- customer Internet sales comes at the expense of reseller SAS 169 1 7 5.64 1.71

-2.94 .005 -1.28 .79

Frequency of Channel Conflict:Ducati 43 1 7 3.07 1.80 .53 -.83 x6q1 Frequency of disagreements

on information exchanges SAS 169 1 7 2.91 1.69

.52(n.s.) .60 -.50

Ducati 43 1 6 3.00 1.66 .46 -.91 x6q2 Frequency of disagreements on product aspects SAS 170 1 7 3.07 1.72

-.25 (n.s.) .46 -.80

Ducati 43 1 7 3.23 1.99 .48 -1.22 x6q3 Frequency of disagreements on pricing issues SAS 169 1 7 3.78 1.90

-1.63 (n.s.) .11 -1.04

Ducati 43 1 7 2.77 1.69 .76 -.29 x6q4 Frequency of disagreements on payment terms SAS 170 1 7 2.86 1.82

-.31 (n.s.) .75 -.48

Ducati 42 1 7 3.02 1.77 .49 -.67 x6q5 Frequency of disagreements on commission on sales SAS 166 1 7 3.81 2.34

-2.40 .019 .13 -1.51

Ducati 43 1 7 3.72 1.88 .22 -.82 x6q6 Frequency of disagreements on producer’s sales organisation SAS 169 1 7 3.47 1.92

.79(n.s.) .26 -1.11

Ducati 43 1 7 3.33 1.95 .41 -1.12 x6q7 Frequency of disagreements on advertising and other promotional activities SAS 167 1 7 3.11 1.79

.66(n.s.) .37 -.95

Ducati 41 1 7 2.61 1.70 .78 -.31 x6q8 Frequency of disagreements on producer’s Internet sales today SAS 169 1 7 4.50 2.10

-6.07 .000 -.35 -1.12

Ducati 41 1 7 2.61 1.84 1.12 .39 x6q9 Frequency of disagreements on producer’s Internet sales when channel was introduced SAS 169 1 7 4.64 2.08

-6.18 .000 -.42 -1.09

Ducati 43 1 7 3.23 1.88 .53 -.60 x6q10 Frequency of disagreements on reseller’s sales effort SAS 168 1 7 2.85 1.54

1.23(n.s) .46 -.48

Intensity of Channel Conflict: Ducati 43 1 6 2.58 1.53 .67 -.61 x6r1 Intensity of disagreements on

information exchanges SAS 167 1 7 2.63 1.40

-.21 (n.s) .54 -.22

Ducati 42 1 5 2.57 1.38 .43 -1.08 x6r2 Intensity of disagreements on product aspects SAS 167 1 7 2.90 1.60

-1.35 (n.s.) .58 -.30

Ducati 42 1 7 2.93 1.98 .84 -.41 x6r3 Intensity of disagreements on pricing issues SAS 168 1 7 3.35 1.75

-1.25 (n.s.) .46 -.63

Ducati 42 1 7 2.52 1.52 1.05 .76 x6r4 Intensity of disagreements on payment terms SAS 166 1 7 2.75 1.74

-.85 (n.s.) .90 -.00

Ducati 42 1 7 2.93 1.76 .65 -.38 x6r5 Intensity of disagreements on commission on sales SAS 166 1 7 3.58 2.27

-2.01 .048 .34 -1.37


147


Max

M S.D. t(p,2-tail)

Skewness

Kur-tosis

Ducati 42 1 7 3.33 1.83 .23 -.94 x6r6 Intensity of disagreements on producer’s sales organisation SAS 163 1 7 3.18 1.74

.50(n.s.) .41 -.74

Ducati 42 1 7 3.17 1.85 .40 -.92 x6r7 Intensity of disagreements on advertising and other promotional activities SAS 166 1 7 2.87 1.71

.95(n.s.) .58 -.64

Ducati 39 1 7 2.59 1.80 .88 -.17 x6r8 Intensity of disagreements on producer’s Internet sales today SAS 167 1 7 4.28 2.08

-5.12 .000 -.15 -1.25

Ducati 39 1 7 2.72 1.96 .84 -.47 x6r9 Intensity of disagreements on producer’s Internet sales when channel was introduced SAS 167 1 7 4.66 2.12

-5.50 .000 -.43 -1.18

Ducati 41 1 7 3.07 1.95 .74 -.53 x6r10 Intensity of disagreements on reseller’s sales effort SAS 165 1 7 2.72 1.49

1.08(n.s.) .58 -.12

Relationship Performance: Ducati 43 2 7 5.23 1.41 -.65 -.36 x8a Relationship with producer

has been very productive SAS 169 1 7 3.85 1.68

5.50.000 -.09 -.76

Ducati 43 2 7 5.30 1.32 -.79 .19 x8b Time and energy spent on relationship with producer has been very worthwhile SAS 170 1 7 3.70 1.72

6.66.000 .16 -.86

Ducati 43 2 7 5.23 1.31 -.59 .20 x8c Relationship has been very effective SAS 169 1 7 3.60 1.63

6.94.000 .13 -.75

Ducati 43 2 7 5.21 1.46 -.48 -.90 x8d Relationship with producer is very rewarding SAS 170 1 7 3.74 1.74

5.69.000 .13 -.84

Sales Performance: Ducati 43 1 6 3.84 1.40 -.19 -.82 x8e Return on investment from

selling producer’s products has exceeded expectations SAS 169 1 7 3.16 1.54

2.78.007 .25 -.60

Ducati 43 1 7 3.65 1.46 .21 -.45 x8f Growth in sales over time from selling producer’s products has exceeded expectations SAS 169 1 7 3.24 1.54

1.62(n.s.) .19 -.69

Ducati 43 1 7 3.56 1.42 .27 -.23 x8g Growth in profit over time from selling producer’s products has exceeded expectations SAS 166 1 7 2.86 1.52

2.83.006 .55 -.34

Ducati 43 1 7 3.63 1.48 .36 -.03 x8h Financial results from selling producer’s products has exceeded expectations SAS 166 1 7 2.82 1.44

3.21.002 .52 -.32

Satisfaction: SAS 41 1 7 4.02 1.46 -.20 -.13 x8o Reseller is very satisfied with

arrangement of overall channel system Ducati 167 1 7 3.80 1.58

.86(n.s.) -.11 -.76

SAS 39 1 5 2.28 1.15 .62 -.20 x8p Reseller is very satisfied with producer’s decision to sell products online Ducati 168 1 7 1.90 1.40

1.77(n.s.) 2.01 3.97

______ CHAPTER 6 ______________________________________________________________

148


Max

M S.D. t(p,2-tail)

Skewness

Kur-tosis

SAS 42 1 7 3.93 1.50 -.32 -.55 x8q Reseller does not think that current channel system needs to be changed Ducati 165 1 7 3.64 1.74

1.06(n.s.) .16 -.76

SAS 42 1 6 3.38 1.56 .01 -1.05 x8r Reseller feels that given the prerequisites, current channel system is the best possible Ducati 165 1 7 3.22 1.71

.57(n.s.) .45 -.67

*Note. For the original wording of items, see Appendix 4.Items that are marked with ® have been reverse coded. Reported t-test values and levels of significance do not assume equal variances. (n.s.) = Non-significant (p > .05). Grey highlighting refers to substantially skewed or excessively peaked items.

The cases that were selected for this research, i.e. Ducati and SAS, were chosen because in several respects they were of polar types (see Chapter 5.4.1). In addition, the above t-tests revealed several significant differences at the p < .05 level. Hence, it was decided to conduct separate analyses of the Ducati and SAS samples.

6.2 Normality Analysesll items were tested individually for normality. As suggested by Hair et al. (1998, p. 73), both graphical analyses and statistical tests were conducted to assess the degree of departure.

6.2.1 Normality Analysis - Ducati The standard deviations for all items were within the range of 1.15 and 2.15 (see Table 6.1), which is considered medium when using a 7-point Likert-type scale (cf. Hair et al., 2007, p. 320). Histograms of each item indicated that some possibly di-verged from normal distribution. Two items, x6q9 and x6r4, had skewness values just above the +1 limit: 1.12 and 1.05, respectively. However, since kurtosis for those items was well below + 3, and since Kolmogorov-Smirnov’s Z test of nor-mality indicated that the distribution was normal, it was concluded that the distri-bution of the Ducati sample could be considered as normal.

6.2.2 Normality Analysis - SAS As can be seen in Table 6.1, all items had estimated standard deviations within the range of 1.27 and 2.34, which is considered medium (cf. Hair et al., 2007, p. 320). However, histograms of each item indicated that some items diverged from normal distribution. The Kolmogorov-Smirnov test yielded significance levels below .05 for several items, another possible indication of ‘non-normal’ distributions (cf. de Vaus, 2002, p. 77). Six items had skewness values smaller than -1 (x4a, x4d, x4e, x4f, x5e, and x6h), whereas item x8p had a skewness value larger than +1 (see Ta-ble 6.1). Items x4a, x4f and x8p had kurtosis values exceeding +3. However, since

A


149

non-normality in practice does not have a severe effect on results (cf. de Vaus, 2002, p. 78), it was decided to ignore these non-normal distributions.

6.3 Missing Dataefore any missing data correction can be implemented, it is necessary to understand why data is missing (Hair et al., 1998, p. 48). Hence, the data were examined with regards to the extent and randomness of missing val-

ues.

6.3.1 Missing Data Analysis - Ducati As already stated, the Ducati sample consisted of 43 cases, 35 of which had no missing data at all. The highest number of missing values for any item was 4. Be-cause of the limited sample size, this means that the three items with four missing values (x6r8, x6r9 and x8p), had a missing value percentage of 9.3 each. Two of these items were among the last ten items measuring intensity of conflict, and since the respondent by then already had responded to ten items measuring the fre-quency of conflict, this could be an indication of fatigue. However, since Little’s MCAR test showed a significance level of p = .953, it was concluded that the data were MCAR. The EM method was used to substitute for missing data.

6.3.2 Missing Data Analysis - SAS Two Swedish SAS questionnaires/cases were eliminated because the proportion of missing data in these cases was too high. Of the remaining 168 cases, 142 had no missing data. Generally, among the items with missing data, only one or two cases were missing. In fact, there were no items with 5% or more missing values. The missing patterns (cases with missing values) table revealed no consistent patterns for missing data. Missing values were computed using the EM method, since no statis-tically reliable deviations from randomness were found using Little's MCAR test (p= .439).

6.4 Measure Validation ince several new scales were used to measure the constructs included in this research, the scales were exposed to vigorous tests of validity and reliability. As a first step, an exploratory factor analysis, i.e. principal component analysis

with varimax rotation, was conducted on the SAS sample. Factor loadings below the minimum recommended value of .45 for a sample size of at least 150 (cf. Hair et al., 1998, p. 112) were suppressed. First, the items used to measure the causes of channel conflict were analysed, followed by the level of channel conflict items, and, fi-nally, the outcomes items.

B

S

______ CHAPTER 6 ______________________________________________________________

150

6.4.1 Exploratory Factor Analysis of the Causes of Channel Conflict Sixteen items (x3a-x5e) were used to measure the causes of channel conflict (seeTable 6.1). Some of these items did not load high enough on any factor, others had too low communality, and one item formed a factor of its own that did not make any theoretical sense. Accordingly, 4 items were dropped from further analysis. Three factors with eigenvalues above 1 were extracted. The Kaiser-Meyer-Olkin(KMO) measure of sampling adequacy was .84, Bartlett’s Test of Sphericity was sig-nificant, and the measure of sampling adequacy (MSA) varied between .74 and .89. Cumulatively, the three factors explained 70.62% of the total variance, and the communalities for the 12 retained items ranged between .56 and .82 (see Table6.2).

Table 6.2 Factor Analysis of the Causes of Channel Conflict*

FactorItemCom-

munica-tion

Domain GoalCom-

munali-ties

x3d Producer keeps reseller well informed about their products ®

.89 .82

x3c Producer keeps reseller well informed about relevant events and changes ®

.86 .80

x3e Producer keeps reseller well informed about what is going on in producer’s company ®

.81 .70

x3b Producer has a good understanding of the frequency of information reseller needs ®

.81 .78

x3a Producer has a good understanding of the type of information needed by reseller ®

.77 .74

x4d Producer’s direct-to-the-customer Internet sales serves the same customers as does reseller

.85 .76

x4f Reseller competes with producer’s direct-to- the-customer Internet sales

.80 .66

x4e Producer’s direct-to-the-customer Internet sales markets identical products as reseller

.78 .62

x5e Producer’s direct-to-the-customer Internet sales comes at the expense of reseller’s channel .72 .56

x5c It is feasible for both reseller and producer to exceed their respective goals ® .83 .71

x5b Reseller and producer share same view on how to compete in reseller’s local market ®

.76 .71

x5a Reseller’s and producer’s goals are compatible ®

.70 .62

Eigenvalue** 3.71 2.60 2.16 % of Variance** 30.93 21.69 18.01 70.62

Extraction Method: Principal Component Analysis Rotation Method: Varimax with Kaiser Normalization * Rotation converged in 5 iterations. ** Values refer to the rotated solution Note. Factor loadings below .45 were suppressed (cf. Hair et al., 1998, p. 112).


151

The three retained factors were labelled inadequate communication, domain dissensus and goal incompatibility. However, item x5e, which originally was thought to meas-ure goal incompatibility, instead loaded on domain dissensus. This seems logical: if the producer’s direct-to-the-customer sales come at the expense of the reseller channel, the parties claims are incongruent in terms of population to be served and the services to be rendered/functions to be performed, thereby leading to domain dissensus (see Chapter 4.1.2 for the definition of domain dissensus). It was therefore concluded that the item should be retained in the domain dissensus factor, because it conformed to the definition of that factor and increased its face validity.

6.4.2 Exploratory Factor Analysis of the Level of Channel Conflict The original intention was to use both frequency and intensity to represent the level of channel conflict. However, although several different combinations of these measures were investigated, none yielded attractive properties for use in Structural Equation Modelling (SEM). In a study of, among other constructs, con-flict, Lee (2001) used a Partial Least Square (PLS) Latent Path Model to estimate the path diagrams of his model. Since the PLS approach is in certain respects simi-lar to SEM, it was decided to follow the example of Lee (2001), who used only fre-quency to capture the level of channel conflict. Nine items were used to measure the present frequency of conflict. These items all loaded on one factor (see Table6.3).

Table 6.3 Factor Analysis of the Level of Channel Conflict*

Level of Channel Conflict

Commu-nalities

x6q1 Frequency of disagreements on information exchanges .77 .60 x6q2 Frequency of disagreements on product aspects .77 .60 x6q7 Frequency of disagreements on advertising, etc. .77 .59 x6q6 Frequency of disagreements on producer’s sales organisation .75 .57 x6q3 Frequency of disagreements on pricing issues .70 .49 x6q8 Frequency of disagreements on producer’s Internet sales today .68 .47 x6q4 Frequency of disagreements on payment terms .68 .46 x6q10 Frequency of disagreements on reseller’s sales effort .67 .44 x6q5 Frequency of disagreements on commission on sales .63 .40 Eigenvalue 4.61 % of Variance 51.23 51.23

Extraction Method: Principal Component Analysis * The solution cannot be rotated. Note. Factor loadings below .45 were suppressed (cf. Hair et al., 1998, p. 112).

KMO was .87, Bartlett’s Test of Sphericity was significant, and the MSA ranged from .84 to .93. However, the factor only explained 51.23% of the total variance (see Table 6.3). Nevertheless, since Brown and Day (1981), in their comparison of dif-ferent measures of conflict in marketing channels, reported that their highest vari-

______ CHAPTER 6 ______________________________________________________________

152

ance was 42.8% (for their frequency measure), it was decided to proceed with the one factor solution.

6.4.3 Exploratory Factor Analysis of Outcomes The 12 items representing performance and satisfaction (x8a-x8h, and x8o-x8r) (see Table 6.1) were submitted to exploratory factor analysis. One item was dropped, since it did not load on the intended factors. Thereafter, two factors with eigenvalues above 1 were extracted. KMO was .91, Bartlett’s Test of Sphericity was significant, and MSA ranged between .77 and .95. The communalities for the 11 retained items were between .52 and .82, and together the two factors explained 70.44% of the total variance (see Table 6.4).

Table 6.4 Rotated Component Matrix*

FactorItemPerform-

anceSatisfac-

tion

Commu-nalities

x8f Growth in sales over time from selling producer’s products has exceeded reseller’s expectations

.89 .82

x8g Growth in profit over time from selling producer’s products has exceeded reseller’s expectations .87 .78

x8e Return on investment from selling producer’s products has exceeded reseller’s expectations .85 .75

x8h Financial results from selling producer’s products exceeded reseller’s expectations .84 .74

x8b Time and energy spent on relationship with producer has been very worthwhile .82 .72

x8d Relationship with producer is very rewarding .82 .73

x8a Relationship with producer has been very productive

.80 .72

x8c Relationship has been very effective .80 .70

x8r Reseller feels that given the prerequisites the current channel system is the best possible

.83 .73

x8q Reseller does not think that current channel system needs to be changed

.72 .54

x8o Reseller is very satisfied with arrangement of overall channel system

.69 .52

Eigenvalue** 5.72 2.03 % of Variance** 51.97 18.48 70.44

Extraction Method: Principal Component Analysis Rotation Method: Varimax with Kaiser Normalization * Rotation converged in 3 iterations ** Values refer to the rotated solution Note. Factor loadings below .45 were suppressed (cf. Hair et al., 1998, p. 112).


153

6.4.4 Confirmatory Factor Analysis After an exploratory factor analysis has been conducted, a confirmatory factor analysis (CFA) is needed to evaluate, and likely refine, the resulting scales with re-gards to unidimensionality (Gerbing & Anderson, 1988). In the present study, a CFA was conducted with the SAS data. As it turned out, some of the scales were multidimensional. For that reason, two indicators in the inadequate communicationscale had to be dropped. Moreover, the eight performance indicators as a first order construct represented a poor fit to the data. For instance, 2/df was 14.95, GFI was .611, and RMSEA was .289. Performance was therefore re-specified as a second order construct, consisting of sales performance and relationship performance (see Figure 6.1). This model fit the data very well: 2/df for the re-specified model was 1.13, GFI was .983, CFI was .999, and RMSEA was .028 (see Table 6.5).

Figure 6.1 Performance

As a first order construct, the nine level of channel conflict indicators also fit the data poorly; e.g., GFI was .874 and RMSEA was .133. Instead, level of channel conflictwas found to be a second order construct, consisting of six indicators (see Figure6.2). That is, level of channel conflict was found to be a three factor model, formed by indicators related to communication-related conflict issues, monetary-related conflict issues, and Internet-related conflict issues. This model fit the data very well. The estimation yielded a 2/df value of 1.74, a GFI of .980, a CFI of .987, and a RMSEA of .067 (see Table 6.5).

______ CHAPTER 6 ______________________________________________________________

154

Figure 6.2 Level of Channel Conflict

The final scale with more than three indicators was domain dissensus. The scale fit the data well; e.g., GFI was .986 (see Table 6.5). Although the RMSEA for the domain dissensus scale was slightly high, it was below the generally accepted .10 value (cf. Hair et al., 2006, p. 748). The goodness-of-fit measures for the scales with more than three indicators are summarized in Table 6.5.

Table 6.5 Fit Indices for Scales

Scale 2 df 2/df p GFI CFI RMSEA

Performance 9.05 8 1.13 .338 .983 .999 .028

Level of Channel Conflict 10.47 6 1.74 .106 .980 .987 .067

Domain Dissensus 4.98 2 2.49 .086 .986 .987 .094

6.4.5 Reliability Analyses of Resulting Scales Reliability analyses were conducted for each of the resulting scales. (For an over-view of the reliability of each scale, see Table 6.6). The reliabilities for SAS and Ducati are presented separately, i.e. the columns marked D represent the results for the Ducati sample, whereas the columns marked S represent SAS. Reliabilities in brackets are provided just as reference values, i.e. these scales were not used in the forthcoming analyses (see Section 6.4.7 for a further discussion).


155

Table 6.6 Reliability Analysis of Scales

Corrected Item-Total Correlation

if Item Deleted

Cronbach’s*

Scale/Retained Items

D S D S D S

Inadequate Communication: .85 .88

x3c Producer keeps reseller well informed about relevant events and changes ®

.72 .78 .79 .82

x3d Producer keeps reseller well informed about their products ®

.70 .82 .81 .78

x3e Producer keeps reseller well informed about what is going on in producer’s company ®

.75 .71 .76 .88

Domain Dissensus: .82 .81

x4d Producer’s direct-to-the-customer Internet sales serves the same customers as the reseller does

.69 .74 .76 .70

x4e Producer’s direct-to-the-customer Internet sales markets identical products as the reseller does

.60 .60 .80 .77

x4f Reseller competes with producer’s direct-to- the-customer Internet sales

.76 .60 .72 .77

x5e Producer’s direct-to-the-customer Internet sales comes at the expense of reseller’s channel

.55 .57 .82 .79

Goal Incompatibility: .75 .77

x5a Reseller’s and producer’s goals are compatible ® .58 .59 .65 .69

x5b Reseller and producer share the same view on how to compete in reseller’s local market ®

.53 .66 .71 .62

x5c It is feasible for both reseller and producer to exceed their respective goals ®

.61 .56 .62 .75

Level of Channel Conflict: .85 (.82)

x6q3 Frequency of disagreements on pricing issues .75 .54 .80 .80

x6q4 Frequency of disagreements on payment terms .63 .56 .83 .80

x6q5 Frequency of disagreements on commission on sales .72 .54 .81 .81

x6q6 Frequency of disagreements on producer’s sales organisation .63 .62 .83 .79

x6q7 Frequency of disagreements on advertising and other promotional activities

.70 .68 .82 .78

x6q8 Frequency of disagreements on producer’s Internet sales today

.41 .62 .85 .79

______ CHAPTER 6 ______________________________________________________________

156


if Item Deleted

Cronbach’s*


D S D S D S

Level of Communication-Related Conflict: (.79) .81

x6q6 Frequency of disagreements on producer’s sales organisation

.66 .68 - -

x6q7 Frequency of disagreements on advertising and other promotional activities

.66 .68 - -

Level of Monetary-Related Conflict: (.78) .65

x6q3 Frequency of disagreements on pricing issues .64 .48 - -

x6q4 Frequency of disagreements on payment terms .64 .48 - -

Level of Internet-Related Conflict: (.55) .69

x6q5 Frequency of disagreements on commission on sales

.46 .52 - -

x6q8 Frequency of disagreements on producer’s Internet sales today .46 .52 - -

Performance: .91 (.93)


.73 .80 .89 .92

x8b Time and energy spent on relationship with producer has been very worthwhile

.71 .80 .90 .92

x8d Relationship with producer is very rewarding .72 .81 .89 .92

x8e Return on investment from selling producer’s products has exceeded reseller’s expectations

.74 .80 .89 .92

x8f Growth in sales over time from selling producer’s products has exceeded expectations

.81 .86 .88 .92

x8g Growth in profit over time from selling producer’s products has exceeded expectations

.75 .79 .89 .92

Relationship Performance: (.92) .91


.85 .80 .87 .89

x8b Time and energy spent on relationship with producer has been very worthwhile

.83 .85 .89 .85

x8d Relationship with producer is very rewarding .83 .82 .89 .88


157


if Item Deleted

Cronbach’s*


D S D S D S

Sales Performance: (.95) .93

x8e Return on investment from selling producer’s products has exceeded reseller’s expectations

.88 .85 .95 .91

x8f Growth in sales over time from selling producer’s products has exceeded expectations

.90 .88 .93 .89

x8g Growth in profit over time from selling producer’s products has exceeded expectations .92 .85 .92 .91

Satisfaction with Channel System: .74 .66

x8o Reseller is very satisfied with arrangement of overall channel system

.57 .42 .64 .63

x8q Reseller does not think that current channel system needs to be changed

.51 .41 .71 .65

x8r Reseller feels that given the prerequisites the current channel system is the best possible

.61 .60 .59 .38

*For two-item scales, the Spearman Brown formula was applied to estimate the reliability (cf. Hu-lin, 2001, p. 55) Values within brackets refer to scales that were not used in the forthcoming analyses (see Section 6.4.7 for a further discussion).

As can be seen, most scales were highly reliable. Although for the Ducati sample the level of Internet-related conflict scale had a split-half reliability of only .55, it must be noted that this scale was not used in the forthcoming analyses; only the summarized level of conflict scale, with Cronbach’s of .85, was used for analysing the Ducati data (see Section 6.4.7 for a further discussion). For the SAS sample, the satisfaction scale had a split-half reliability of .66, the level of monetary-related conflict scale had a split-half reliability of .65, and the Internet-related conflict scale had a split-half reliability of .69. Although the reliability for these three scales is be-low the generally agreed upon .70 limit, it is well above the .60 limit for moderate strength of association (cf. Hair et al., 2007, p. 244).

6.4.6 Validity & Common Method Variance First, an exploratory factor analysis of all resulting scales, Harman’s single-factor test (Podsakoff & Organ, 1986) was used to estimate discriminant validity and common method variance. Factor loadings below the minimum recommended value of .45 for a sample size of at least 150 (cf. Hair et al., 1998, p. 112) were suppressed. KMOwas .86, Bartlett’s Test of Sphericity was significant, and MSA ranged between .71 and .93. Six factors with eigenvalues above one were extracted; together, the six factors explained 68.67% of the total variance. The communalities varied between .49 and .86, and the largest factor explained just above 18% of the total variance (see Table 6.7). In the un-rotated solution, the largest factor explained 31.32% of the

______ CHAPTER 6 ______________________________________________________________

158

total variance, and it was therefore concluded that the common method variance was low (cf. Podsakoff & Organ, 1986). All cross-loadings were below the sup-pressed value of .45, which indicates that the discriminant validity of the scales was satisfactory.

Table 6.7 Factor Analysis of Retained Indicators - SAS*

Indicator/Scale Per-form-ance

Level of ChannelConflict

Domain Com-munica-

tion

Goal Satisfac-tion

Com-munali-

tiesx8f PERF .89 .86 x8g PERF .88 .82 x8e PERF .87 .81 x8b PERF .73 .73 x8a PERF .72 .72 x8d PERF .72 .76 x6q7 CONFL .80 .68 x6q6 CONFL .76 .64 x6q4 CONFL .69 .56 x6q8 CONFL .68 .66 x6q3 CONFL .68 .55 x6q5 CONFL .65 .49 x4d DOMAIN .85 .76 x4f DOMAIN .79 .67 x4e DOMAIN .75 .59 x5e DOMAIN .73 .57 x3e COMM .83 .78 x3d COMM .83 .84 x3c COMM .77 .79 x5c GOAL .78 .74 x5b GOAL .70 .71 x5a GOAL .63 .60 x8r SATISF .79 .70 x8q SATISF .76 .62 x8o SATISF .58 .52 Eigenvalue** 4.55 3.27 2.79 2.56 2.12 1.88 % of Variance** 18.19 13.09 11.17 10.24 8.46 7.51 68.67

Extraction Method: Principal Component Analysis Rotation Method: Varimax with Kaiser Normalization * Rotation converged in 6 iterations ** Values refer to the rotated solution Note. Factor loadings below .45 were suppressed (cf. Hair et al., 1998, p. 112).

A correlation matrix of the retained items provided further evidence of discrimi-nant validity, as well as convergent validity (see Table 6.8). As can be seen, all items loaded primarily on the construct they were designated to measure. It was there-fore concluded that both discriminant validity and convergent validity were satisfactory.


159

Table 6.8 Discriminant and Convergent Validity - Bivariate Correlations

x3c

x3d

x3e

x4d

x4e

x4f

x5e

x5a

x5b

x5c

x6q3

x6q4

x6q5

x6q6

x6q7

x6q8

x8a

x8b

x8d

x8e

x8f

x8g

x8o

x8q

x8r

x8r

1

x8q 1

.46*

*

x8o 1

.23*

*

.48*

*

x8g 1

.25*

*

.21*

*

.31*

*

x8f

1

.83*

*

.31*

*

.23*

*

.36*

*

x8e 1

.84*

*

.80*

*

.33*

*

.22*

*

.31*

*

x8d 1

.65*

*

.70*

*

.63*

*

.36*

*

.25*

*

.32*

*

x8b 1

.80*

*

.62*

*

.67*

*

.60*

*

.30*

*

.24*

*

.26*

*

x8a

1

.78*

*

.74*

*

.62*

*

.70*

*

.64*

*

.36*

*

.26*

*

.34*

*

x6q8 1

-.14

-.12

-.14

-.10

-.12

-.10

-.10 .07

-.10

x6q7 1

.56*

*

-.18

*

-.16

*

-.24

**

-.15

-.13

-.08

-.15

*

.00

-.01

x6q6 1

.68*

*

.46*

*

-.16

*

-.11

-.12

-.17

*

-.13

-.14

-.08 .02

-.04

x6q5 1

.33*

*

.38*

*

.52*

*

-.27

**

-.17

*

-.18

*

-.18

*

-.17

*

-.24

**

-.12

-.01

-.13

x6q4 1

.41*

*

.41*

*

.44*

*

.35*

*

-.24

**

-.18

*

-.26

**

-.15

*

-.16

*

-.13

-.20

*

-.12

-.20

**

x6q3 1

.48*

*

.38*

*

.42*

*

.42*

*

.36*

*

-.17

*

-.10

-.12

-.06

-.04

-.03

-.12

-.06

-.15

x3c

x3d

x3e

x4d

x4e

x4f

x5e

x5a

x5b

x5c

x6q3

x6q4

x6q5

x6q6

x6q7

x6q8

x8a

x8b

x8d

x8e

x8f

x8g

x8o

x8q

x8r

______ CHAPTER 6 ______________________________________________________________

160

x3c

x3d

x3e

x4d

x4e

x4f

x5e

x5a

x5b

x5c

x6q3

x6q4

x6q5

x6q6

x6q7

x6q8

x8a

x8b

x8d

x8e

x8f

x8g

x8o

x8q

x8r

x5c 1

.18*

.18*

.18* .09

.18*

.21*

*

-.43

**

-.47

**

-.48

**

-.31

**

-.42

**

-.34

**

-.32

**

-.21

**

-.24

**

x5b 1

.54*

*

.14

.25*

*

.26*

*

.21*

*

.30*

*

-.49

**

-.46

**

-.49

**

-.42

**

-.46

**

-.42

**

-.41

**

-.34

**

-.03

-.26

**

x5a

1

.60*

*

.46*

*

.08

.19* .14

.18*

.19*

.21*

*

-.44

**

-.50

**

-.48

**

-.42

**

-.44

**

-.40

**

-.20

*

-.16

*

-.21

*

x5e 1 .12

.22*

*

.25*

*

.07

.07

.19* .13

.06

.30*

*

-.15

-.09

-.06

-.18

*

-.22

**

-.18

*

-.19

*

-.00

-.12

x4f

1

.46*

*

.15

.14

.04

.08

.07

.12

.12

.05

.32*

*

-.04

-.02

-.06

-.12

-.11

-.20

*

-.14

-.05

-.11

x4e 1

.49*

*

.39*

*

.07

.14

.17* .10

.13

.13

.08

.14

.32*

*

-.12

-.07

-.08

-.09

-.12

-.07

-.15 .05

-.08

x4d 1

.63*

*

.55*

*

.58*

*

.18*

.24*

*

.21*

*

.11

.03

.18*

.20*

*

.14

.40*

*

-.14

-.08

-.10

-.19

*

-.19

*

-.18

*

-.14 .02

-.10

x3e 1

.16* .09

.26*

*

.13

.38*

*

.38*

*

.24*

*

.25*

*

.16* .13

.15*

.17*

.19*

-.38

**

-.35

**

-.40

**

-.36

**

-.33

**

-.33

**

-.32

**

-.11

-.25

**

x3d 1

.70*

*

.10

.03

.08

.09

.40*

*

.42*

*

.26*

*

.28*

*

.25*

*

.19*

.23*

*

.26*

*

.16*

-.53

**

-.51

**

-.56

**

-.35

**

-.38

**

-.34

**

-.39

**

-.15

-.29

**

x3c 1

.78*

*

.64*

*

.06

-.00 .03

.06

.42*

*

.45*

*

.33*

*

.27*

*

.30*

*

.17*

.27*

*

.27*

*

.17*

-.50

**

-.44

**

-.52

**

-.40

**

-.41

**

-.42

**

-.34

**

-.11

-.30

**

x3c

x3d

x3e

x4d

x4e

x4f

x5e

x5a

x5b

x5c

x6q3

x6q4

x6q5

x6q6

x6q7

x6q8

x8a

x8b

x8d

x8e

x8f

x8g

x8o

x8q

x8r

** Correlation is significant at the p < .001 level (2-tailed) * Correlation is significant at the p < .05 level (2-tailed)

6.4.7 Summated Averaged Scales Given the reliability and validity of the scales, corresponding indicators (see Table6.6) were summated and averaged. The means and standard deviations for the summated scales are presented separately for each sample (see Table 6.9 and Table6.10). For Ducati, the six items that formed the three dimensions of the second order construct labelled as level of channel conflict (see Table 6.6), were summated and averaged to form the variable CONF_FRE. Also, the items on the perform-ance scale, which were related to relationship performance as well as sales perform-ance (see Table 6.6), were summated and averaged to form the variable


161

PERF_SUM. For a justification of the use of summated scales that are multidimen-sional, see Chapter 5.4.4, p. 138.

Table 6.9 Summated Averaged Scales - SAS

Scale/Construct Label n M S.D.

Goal Incompatibility (GOAL_SUM) 168 4.83 1.34 Domain Dissensus (DOM_SUM) 168 5.84 1.34 Inadequate Communication (COMM_SUM) 168 4.45 1.44 Communication-Related Conflict Issues (CF_C_SUM) 168 3.29 1.70 Monetary-Related Conflict Issues (CF_M_SUM) 168 3.33 1.60 Internet-Related Conflict Issues (CF_I_SUM) 168 4.15 1.93 Sales Performance (PERS_SUM) 168 3.08 1.44 Relationship Performance (PERR_SUM) 168 3.73 1.56 Satisfaction with Channel System (SAT_SUM) 168 3.53 1.29

Table 6.10 Summated Averaged Scales - Ducati

Scale/Construct Label n M S.D.

Goal Incompatibility (GOAL_SUM) 43 3.37 1.33 Domain Dissensus (DOM_SUM) 43 4.87 1.46 Inadequate Communication (COMM_SUM) 43 3.63 1.53 Level of Channel Conflict (CONF_FRE) 43 3.11 1.38 Performance (PERF_SUM) 43 4.46 1.17 Satisfaction with Channel System (SAT_SUM) 43 3.75 1.22

6.5 Hypothesis Testing he hypothesis testing of the SAS data was conducted by Structural Equa-tion Modelling and is presented in the following section. The hypothesis testing of the Ducati-data is presented in Section 6.5.3.

6.5.1 Structural Equation Modelling - SAS First, the outcomes part of the research model was tested (see Figure 6.3). The model fit the data very well. The 2 was 9.94 with 7 degrees of freedom, p was .192, GFI was .981, CFI was .990, and RMSEA was .050 (see Table 6.16, p. 166). The coefficients of determination (R2) are summarized in Table 6.11.

Table 6.11 R2-Correlations for Outcomes Part of Research Model

Construct R2

PERF .09 SAT_SUM .21 CF_I_SUM .51 CF_M_SUM .51 CF_C_SUM .56 PERS_SUM .66 PERR_SUM .84

T

______ CHAPTER 6 ______________________________________________________________

162

Figure 6.3 Outcomes Part of Research Model

When the estimates were reviewed, it was found that, with one exception, the re-lationships were both reasonable and statistically significant (see Table 6.12). In other words, the relationship between conflict and satisfaction was not statistically significant. The highest standardized residual covariance, between satisfaction and monetary-related conflict issues, was -1.25, which means that no statistically sig-nificant discrepancy in covariance was found between the constructs (cf. Byrne, 2001, p. 89).

Table 6.12 Maximum Likelihood Estimates - Outcomes Part of Research Model

IndependentVariable

DependentVariable

Unstandard-ized

Loadings

S.E. t p1-

tailed

Standard-ized

LoadingsCONFLICT PERF -0.38 .12 -3.10 .001 -.30 PERF PERR_SUM 1.00 .92 PERF PERS_SUM 0.82 .10 7.76 .000 .81 PERF SAT_SUM 0.41 .08 5.07 .000 .45 CONFLICT CF_C_SUM 1.12 .16 7.09 .000 .75 CONFLICT CF_M_SUM 1.00 .71 CONFLICT SAT_SUM -0.03 .10 -0.29 .386 -.02 CONFLICT CF_I_SUM 1.21 .17 7.05 .000 .71

Nevertheless, when the whole research model was tested, including the antece-dents (see Figure 6.4), the statistics suggested a poor fit to the data. For instance, GFI was .873, CFI was .795 and RMSEA was .166 (see Table 6.16, p. 166).


163

Figure 6.4 Research Model

Several standardized residuals with values exceeding 2.58 were found. The modifi-cation indices indicated a substantial drop in the overall 2 value if a relationship between inadequate communication and performance was added, and between goal incompatibility and performance (see Table 6.13).

Table 6.13 Modification Indices and Parameter Change Statistics

Independent Variable Dependent Variable M.I. Par. Change COMM_SUM PERF 30.98 -.43 GOAL_SUM PERF 41.77 -.54 COMM_SUM CF_I_SUM 5.06 -.19 DOM_SUM CF_I_SUM 7.06 .24 PERS_SUM CF_M_SUM 4.84 .15 SAT_SUM CF_C_SUM 5.42 .19 COMM_SUM PERR_SUM 11.78 -.18 GOAL_SUM PERR_SUM 12.53 -.20

______ CHAPTER 6 ______________________________________________________________

164

In a recent study, Kim, Cavusgil and Calantone (2006) found that high quality in-formation exchange had a positive impact on companies’ performance. They con-tend that “information exchange seems to be a cornerstone for other channel capa-bilities and market performance.” (p. 51). Since high-quality information exchange was found to have an impact on performance, it seems plausible that inadequate communication also would impact performance. Accordingly, the relationship be-tween inadequate communication and performance was included in the re-specified research model (see Figure 6.5). Hence, the following ad hoc hypothesis is proposed:

This quote, from SAS’ 2003 annual report, seems to support the inclusion of a re-lationship between goal incompatibility and performance, since it clearly indicates the desire from SAS’ side to dis-intermediate:

“Scandinavian Airlines’ goal is to achieve 40% of sales via the Internet. For a variety of rea-sons, a majority of big customers want distribution to continue via travel agencies. This limits the potential for Scandinavian Airlines’ share of Internet sales.” (SAS, 2004, p. 17)

In fact, SAS’ Internet sales seem to have affected travel agencies. For instance, in 2004, 21% of SAS Sveriges’ tickets were booked via the Internet, whereas 65% were booked via travel agencies (SAS, 2005, p. 38). In 2006, the share of Internet bookings increased to 31%, whereas the share of tickets booked via travel agencies decreased to 62% (SAS, 2007, p. 38). The decrease in bookings via travel agencies would reasonably have impacted travel agencies’ performance. Hence, it seems logical to also include a relationship between goal incompatibility and performance in the re-specified research model (see Figure 6.5). Accordingly, the following ad hoc hypothesis is proposed:

The estimation of the re-specified research model yielded a 2/df value of 1.67, a GFI of .957, a CFI of .973, and a RMSEA of .063, which indicates that the model fits the data very well.

H11: The level of goal incompatibility among multiple marketing channels is negatively associated with the level of perceived performance.

H10: The level of inadequate communication among multiple marketing channels is negatively associated with the level of perceived performance.


165

Figure 6.5 Re-Specified Research Model

The R2 values for the re-specified research model are shown in Table 6.14.

Table 6.14 R2 Correlations for the Re-Specified Research Model

Construct R2

GOAL_SUM .24 DOM_SUM .06 CONFLICT .22 PERF .54 SAT_SUM .22 CF_I_SUM .54 CF_M_SUM .50 CF_C_SUM .54 PERS_SUM .63 PERR_SUM .88

Most of the relationships in the re-specified research model were reasonable and statistically significant well beyond the p < .05 level. However, a few relationships

______ CHAPTER 6 ______________________________________________________________

166

were not statistically significant, i.e. those between inadequate communication and domain dissensus, level of channel conflict and performance, and level of channel conflict and satisfaction with channel system (see Table 6.15).

Table 6.15 Maximum Likelihood Estimates - Re-Specified Model

IndependentVariable

DependentVariable

Unstan-dardizedLoadings

S.E. t p1-tailed

StandardizedLoadings

COMM_SUM GOAL_SUM 0.45 .06 7.15 .000 .48 COMM_SUM DOM_SUM 0.00 .08 0.05 .482 .00 GOAL_SUM DOM_SUM 0.24 .09 2.82 .002 .24 COMM_SUM CONFLICT 0.26 .09 2.84 .002 .27 DOM_SUM CONFLICT 0.23 .09 2.61 .004 .22 GOAL_SUM CONFLICT 0.18 .10 1.78 .038 .17 CONFLICT PERF 0.01 .08 0.16 .438 .01 COMM_SUM PERF -0.38 .07 -5.26 .000 -.37 GOAL_SUM PERF -0.52 .08 -6.92 .000 -.48 CONFLICT CF_I_SUM 1.00 .73 CONFLICT CF_M_SUM 0.81 .11 7.28 .000 .71 CONFLICT CF_C_SUM 0.88 .12 7.39 .000 .73 PERF PERR_SUM 1.00 .94 PERF SAT_SUM 0.40 .07 5.59 .000 .45 PERF PERS_SUM 0.78 .07 11.69 .000 .79 CONFLICT SAT_SUM -0.04 .08 -0.47 .345 -.04

The highest standardized residual covariance (.33) was found between domain dis-sensus and Internet-related conflict issues. In other words, no statistically significant discrepancy with regards to covariance was found between the constructs (cf. Hair et al., 1998, p. 641). Even so, the modification indices suggest the inclusion of a relationship between domain dissensus and Internet-related channel conflict, al-though the expected parameter change is very small. Since the goodness-of-fit measures of the re-specified model indicated a substantially good fit (see Table6.16), and since there is a risk of over-fitting a model when including additional parameters (Byrne, 2001, p. 93), it was decided to stop fitting the model.

Table 6.16 Fit Indices for the Tested Models

Model 2 df 2/df p GFI CFI RMSEA

Outcomes Part of Research Model

9.94 7 1.42 .192 .981 .990 .050

ResearchModel

123.06 22 5.59 .000 .873 .795 .166

Re-specifiedResearch Model

33.33 20 1.67 .031 .957 .973 .063


167

6.5.2 Hypothesis Testing – SASThe results from the SEM of the re-specified research model were used to test the hypotheses for SAS (see Table 6.17). The results of hypothesis testing for the Ducati data are provided in Section 6.5.3.

Table 6.17 Hypothesis Testing – SAS (AMOS Estimates)

H.no.

Independent Variable

DependentVariable

Standard-ized

Loadings

t p1-

tailed

Result

1 GOAL_SUM CONFLICT .17 1.78 .038 Supported 2 GOAL_SUM DOM_SUM .24 2.82 .002 Supported 3 DOM_SUM CONFLICT .22 2.61 .004 Supported 4 COMM_SUM GOAL_SUM .48 7.15 .000 Supported 5 COMM_SUM DOM_SUM .00 0.05 .482 Not Supported 6 COMM_SUM CONFLICT .27 2.84 .002 Supported 7 CONFLICT PERF .01 0.16 .438 Not Supported 8 PERF SAT_SUM .45 5.59 .000 Supported 9 CONFLICT SAT_SUM -.04 -0.47 .345 Not Supported 10 COMM_SUM PERF -.37 -5.26 .000 Supported 11 GOAL_SUM PERF -.48 -6.92 .000 Supported

Hypothesis one suggested that the level of goal incompatibility among multiple mar-keting channels is positively associated with the level of channel conflict. There was statistical support for this relationship at the p < .05 level. There was a low to moderate, positive relationship between incompatible goals of companies and the level of channel conflict ( = .17), which means that hypothesis one was sup-ported.

Hypothesis two posited that the level of goal incompatibility among multiple mar-keting channels is positively associated with the level of domain dissensus. There was a low to moderate, positive, statistically significant relationship at the p < .005 level between incompatible goals and domain dissensus ( = .24). Thus, statistical support was found for hypothesis two.

Hypothesis three suggested that the level of domain dissensus among multiple mar-keting channels is positively associated with the level of channel conflict. There was statistical support at the p < .005 level for a low to moderate, positive relation-ship ( = .22), which means that hypothesis three was supported.

Hypothesis four proposed that inadequate communication among multiple market-ing channels is positively associated with goal incompatibility. Statistical support was found for a moderate to substantial, positive relationship at the p < .001 level ( = .48). Thus, hypothesis four was supported.

______ CHAPTER 6 ______________________________________________________________

168

Hypothesis five suggested that inadequate communication among multiple market-ing channels is positively associated with domain dissensus. However, this relation-ship was statistically non-significant (p = .482). Thus, hypothesis five was not sup-ported.

Hypothesis six proposed that inadequate communication among multiple marketing channels is positively associated with higher levels of channel conflict. Statistical support at the p < .005 level was found for a low to moderate, positive relationship ( = .27). Thus, hypothesis six was supported.

Hypothesis seven posited that the level of channel conflict among multiple marketing channels is negatively associated with perceived performance. However, this rela-tionship was statistically non-significant (p = .438), so statistical support was not found for a relationship between conflict and performance.

Hypothesis eight suggested that perceived performance is positively associated with perceived satisfaction with channel system. A moderate to substantial positive, sta-tistically significant relationship at the p < .001 level was found ( = .45), which means that hypothesis eight was supported.

Hypothesis nine proposed that the level of channel conflict is negatively associated with perceived satisfaction with the channel system. The relationship was non-significant (p = .345). Thus, hypothesis nine was not statistically supported.

Hypothesis ten, one of two ad hoc hypotheses, proposed that the level of inadequate communication among multiple marketing channels is negatively associated with the level of perceived performance. Statistical significance at the p < .001 level in-dicated support for a moderate to substantial, negative relationship ( = -.37), which means that hypothesis ten was statistically supported.

Hypothesis eleven, the second ad hoc hypothesis, proposed that the level of goal in-compatibility among multiple marketing channels is negatively associated with the level of perceived performance. A moderate to substantial, negative, statistically significant relationship at the p < .001 level was found ( = -.48). Therefore, hy-pothesis eleven was supported.

6.5.3 Hypothesis Testing - Ducati The regression analyses of the Ducati sample were conducted with the re-specified research model as the basis (see Figure 6.5, p. 165). A summary of the hypothesis testing is presented in Table 6.24, p. 172.


169

Level of Channel Conflict as Dependent Variable Hypothesis one suggested that the level of goal incompatibility among multiple mar-keting channels is positively associated with the level of channel conflict; hypothesis three posited that the level of domain dissensus among multiple marketing channels is positively associated with the level of channel conflict; hypothesis six proposed that inadequate communication among multiple marketing channels is positively associated with higher levels of channel conflict. A standard multiple regression was performed with level of channel conflict as the dependent variable, and goal in-compatibility, domain dissensus, and inadequate communication as independent variables. The regression statistics are presented in Table 6.18.

Table 6.18 Multiple Regression on Level of Channel Conflict

UnstandardizedCoefficients

StandardizedCoefficients

t p1-

VIF R2Model

B S.E. tailed

1 (Constant) 0.41 0.73 0.56 .288

GOAL_SUM 0.13 0.16 .13 0.84 .202 1.35

DOM_SUM 0.21 0.14 .22 1.56 .064 1.19

COMM_SUM 0.34 0.15 .37 2.26 .015 1.56

.33

Predictors: (Constant), COMM_SUM, DOM_SUM, GOAL_SUM Dependent Variable: CONF_FRE F(3, 27) = 6.30, p < .001

R was significantly different from zero, F(3, 27) = 6.30, p < .001, with R2 at .33, indicating that 33% of the variability in level of channel conflict is predicted by goal incompatibility, inadequate communication, and domain dissensus. However, the relationship between goal incompatibility and level of channel conflict was sta-tistically non-significant (p = .202), as was the relationship between domain dissen-sus and level of channel conflict (p = .064). Therefore, statistical support was not found for hypotheses one and three. Nevertheless, a moderate to substantial posi-tive relationship, statistically significant at the p < .05 level, was found between in-adequate communication and level of channel conflict ( = .37). Hypothesis six was therefore statistically supported.

Goal Incompatibility on Domain Dissensus Hypothesis two posited that the level of goal incompatibility among multiple mar-keting channels is positively associated with the level of domain dissensus. As can be seen in Table 6.19, the relationship between goal incompatibility and domain dissensus was statistically non-significant (p = .135); hence, support was not found for hypothesis two.

______ CHAPTER 6 ______________________________________________________________

170

Table 6.19 Simple Regression of Goal Incompatibility on Domain Dissensus

Unstandardized Coefficients

Standardized Coefficients

t p1-tailed

R2Model

B S.E.

1 (Constant) 4.24 0.61 6.91 .000

GOAL_SUM 0.19 0.17 .17 1.12 .135 .03

Predictors: (Constant), GOAL_SUM Dependent Variable: DOM_SUM F(1, 29) = 1.25

Inadequate Communication on Goal Incompatibility Hypothesis four proposed that inadequate communication among multiple market-ing channels is positively associated with goal incompatibility. The results of the regression analysis are presented in Table 6.20. The R2 indicates that 26% of the variability in level of channel conflict is predicted by inadequate communication. As statistical support was found for a substantial to very strong, positive relationship at the p < .001 level ( = .51), hypothesis four was supported.

Table 6.20 Simple Regression of Inadequate Communication on Goal Incomp.



t p1-tailed

R2Model

B S.E.

1 (Constant) 1.77 0.46 3.85 .000

COMM_SUM 0.44 0.12 .51 3.80 .000 .26

Predictors: (Constant), COMM_SUM Dependent Variable: GOAL_SUM F(1, 29) = 14.42

Inadequate Communication on Domain Dissensus Hypothesis five proposed that inadequate communication among multiple marketing channels is positively associated with domain dissensus. The results of the regres-sion analysis are presented in Table 6.21.

Table 6.21 Simple Regression of Inadequate Communication on Domain Diss.



t p1-tailed

R2Model

B S.E.

1 (Constant) 3.51 0.54 6.51 .000

COMM_SUM 0.38 0.14 .39 2.75 .004 .16

Predictors: (Constant), COMM_SUM Dependent Variable: DOM_SUM F(1, 29) = 7.54


171

The R2 indicates that 16% of the variability in domain dissensus is predicted by in-adequate communication. Statistical significance at the p < .005 level indicated support for a moderate to substantial, positive relationship ( = .39), which means that hypothesis five was statistically supported.

Performance as Dependent Variable Hypothesis seven posited that the level of channel conflict among multiple marketing channels is negatively associated with perceived performance; hypothesis ten pro-posed that the level of inadequate communication among multiple marketing channels is negatively associated with the level of perceived performance; hypothesis eleven suggested that the level of goal incompatibility among multiple marketing channels is negatively associated with the level of perceived performance. Accord-ingly, a standard multiple regression was performed with performance as the de-pendent variable and goal incompatibility, inadequate communication and level of channel conflict as independent variables. The regression statistics are presented in Table 6.22.

Table 6.22 Multiple Regression on Performance



t p1-

VIF R2Model

B S.E. tailed

1 (Constant) 6.58 0.46 14.37 .000

CONF_FRE -0.02 0.12 -.02 -0.16 .437 1.40

GOAL_SUM -0.34 0.13 -.39 -2.69 .005 1.37

COMM_SUM -0.25 0.12 -.32 -2.02 .025 1.65

.40

Predictors: (Constant), COMM_SUM, GOAL_SUM, CONF_FRE Dependent Variable: PERF_SUM F(3, 27) = 8.66, p < 0.001

R was significantly different from zero, F(3, 27) = 8.66, p < .001, with R2 at .40, which indicates that 40% of the variability in performance is predicted by goal in-compatibility, inadequate communication, and level of channel conflict. However, the relationship between level of channel conflict and performance was statistically non-significant (p = .437), and therefore statistical support was not found for hy-pothesis seven. On the other hand, moderate to substantial, negative, statistically significant relationships at the p < .05 level were found between goal incompatibil-ity and performance ( = -.39), and between inadequate communication and per-formance ( = -.32). Hypotheses ten and eleven were therefore statistically sup-ported.

Satisfaction as Dependent Variable Hypothesis eight suggested that perceived performance is positively associated with perceived satisfaction with channel system, whereas hypothesis nine proposed that

______ CHAPTER 6 ______________________________________________________________

172

the level of channel conflict is negatively associated with perceived satisfaction with the channel system. A standard multiple regression was performed with satisfaction as the dependent variable, and performance and level of channel conflict as inde-pendent variables. The regression statistics are presented in Table 6.23.

Table 6.23 Multiple Regression on Satisfaction



t p1-

VIF R2Model

B S.E. tailed

1 (Constant) 1.18 0.83 1.43 .080

CONF_FRE -0.06 0.12 -.07 -0.52 .302 1.12

PERF_SUM 0.62 0.14 .59 4.46 .000 1.12 .38

Predictors: (Constant), PERF_SUM, CONF_FRE Dependent Variable: SAT_SUM F(2, 28) = 12.18, p < .001

R was significantly different from zero, F(2, 28) = 12.18, p < .001, with R2 at .38, which indicates that 38% of the variability in satisfaction is predicted by perceived performance and level of channel conflict. However, the relationship between level of channel conflict and satisfaction was statistically non-significant (p = .302), and therefore statistical support was not found for hypothesis nine. On the other hand, a substantial to very strong, positive, statistically significant relationships at the p < .001 level was found between perceived performance and satisfaction with channel system ( = .59). Hypotheses eight was therefore statistically supported.

Summary of Hypothesis Testing - Ducati The results of hypothesis testing with the Ducati sample are summarized in Table 6.24.

Table 6.24 Hypothesis Testing – Ducati (Regression Analysis)

H.no.

IndependentVariable

DependentVariable

Standard-ized

Loadings

t p1-tailed

Result

1 GOAL_SUM CONFLICT .13 0.84 .202 Not Supported 2 GOAL_SUM DOM_SUM .17 1.12 .135 Not Supported 3 DOM_SUM CONFLICT .22 1.56 .064 Not Supported 4 COMM_SUM GOAL_SUM .51 3.80 .000 Supported5 COMM_SUM DOM_SUM .39 2.75 .004 Supported6 COMM_SUM CONFLICT .37 2.26 .015 Supported7 CONFLICT PERF -.02 -0.16 .437 Not Supported 8 PERF SAT_SUM .59 4.46 .000 Supported9 CONFLICT SAT_SUM -.07 -0.52 .302 Not Supported 10 COMM_SUM PERF -.32 -2.02 .025 Supported11 GOAL_SUM PERF -.39 -2.69 .005 Supported

________________________________ DISCUSSION, CONCLUSIONS & CONTRIBUTIONS ______

173

7 DISCUSSION, CONCLUSIONS & CONTRIBUTIONS In this chapter, the results of the data analyses will be further discussed. Thereafter, the limi-tations of the study will be presented, followed by theoretical contributions of the study includ-ing an empirically derived model, managerial implications, and suggestions for future research.

7.1 Discussion n this section the main findings from the analyses of results from the two resel-ler surveys, i.e. Ducati and SAS, will be discussed. Recall that the research purpose of this thesis is:

Based on a literature review and the first study of this thesis, three research ques-tions were formulated:

How can the causes of channel conflict in multiple marketing channels, including the Internet be described? How can the level of channel conflict in multiple marketing channels, including the Inter-net be described? How can the relationship between the level of channel conflict in multiple marketing channels, including the Internet and the channel outcomes be described?

These questions were further developed into a research model and research hy-potheses (see Chapter 4.4, p. 127 for an overview). In the previous chapter, the hypothesis testing indicated that some differences existed between the two samples, i.e. Ducati and SAS. Hence, a comparison of the results of the hypothesis tests is provided in Table 7.1. Differences are highlighted in grey.

Table 7.1 Hypothesis Tests – Comparison of Ducati and SAS

H.no.

IndependentVariable

DependentVariable

Result for Ducati N = 43

Result for SAS N = 168

1 GOAL_SUM CONFLICT Not Supported Supported2 GOAL_SUM DOM_SUM Not Supported Supported3 DOM_SUM CONFLICT Not Supported Supported4 COMM_SUM GOAL_SUM Supported Supported 5 COMM_SUM DOM_SUM Supported Not Supported 6 COMM_SUM CONFLICT Supported Supported 7 CONFLICT PERF Not Supported Not Supported 8 PERF SAT_SUM Supported Supported 9 CONFLICT SAT_SUM Not Supported Not Supported 10 COMM_SUM PERF Supported Supported 11 GOAL_SUM PERF Supported Supported

ITo explore and describe managers’ perspectives on conflict in multiple marketing channels,

with a focus on the Internet

______ CHAPTER 7 ______________________________________________________________

174

As can be seen, Hypotheses one, two and three were statistically supported for SAS, but not for Ducati. On the contrary, Hypothesis five was supported for Ducati but not for SAS. The results of the hypothesis testing, along with the differences be-tween the cases, will be discussed with respect to the research questions, i.e. the causes of channel conflict, the level of channel conflict, and the relationship between the level of channel conflict and outcomes.

7.1.1 Causes of Channel Conflict The hypothesised causes of channel conflict are goal incompatibility, domain dissensus,and inadequate communication. Relationships among these causes were also hypothe-sized (see Table 4.2, p. 127, for an overview of the hypothesized relationships).

Goal Incompatibility As can be seen in Table 7.1, for Ducati, statistical support could not be found for goal incompatibility being a cause of channel conflict, whereas comparable statisti-cal support was found for SAS. There were also differences between the cases re-garding the relationship between goal incompatibility and domain dissensus. That is, for Ducati statistical support could not be found for goal incompatibility as a cause of domain dissensus, whereas statistical support for this relationship was found for SAS.

A probable explanation for these contradictory results is that SAS’s expressed goal is to increase its direct-to-the-customer Internet sales (as discussed in Chapter 6.5, p. 164), which in the end will lead to dis-intermediation. This goal obviously is in-compatible with the goal of resellers, who, in order to make money, wish to re-main the link between end customers and SAS. SAS has already tried to exclude non-IATA travel agencies from the business relationship, a fact which probably is reflected in the responses from these travel agencies. Non-IATA travel agencies do not have access to SAS’s extranet (sassalesinfo.com), and they have to pay the same fee as private persons in order to purchase tickets for their customers (see Chapter 5.4.3 for a further discussion). (That is not to say that IATA-accredited travel agencies object to these actions from SAS; rather the reverse, since IATA accredi-tation is connected with costs, among other things). On the contrary, since Ducati’s resellers were involved in Ducati’s online sales, most of those resellers probably perceive that Ducati wishes to maintain its relationship with them, and that Ducati’s main goal in doing so is to help Ducati increase its sales, a goal which is compatible with the resellers’ goal. The descriptive statistics seem to support this assumption, as SAS’s resellers reported a fairly high mean (5.37 and 5.07) on two of the most pronounced items measuring goal incompatibility, whereas Ducati’s resel-lers reported a significantly lower mean on the same items (3.65 and 3.09; see Table6.1, p. 145). Besides, Ducati’s resellers reported quite low levels of conflict on items measuring Internet-related conflict (M = 2.61 and 2.61), whereas SAS’s re-sellers reported fairly high levels of conflict on the same items (M = 4.50 and 4.64).


175

These results further indicate that goal incompatibility for Ducati was not closely associated with Ducati’s Internet sales.

Moreover, SAS removed the sales commission to all travel agencies (see Chapter 2.4.2, p. 61). In addition, non-IATA travel agencies have to order tickets in the same way as private persons, which means that they actually have to pay a fee to purchase tickets for their customers (see Chapter 5.4.3). These actions are incom-patible with agency theory13, which is based on the premise that principals delegate duties to agents to act on their behalf. Since it is assumed that agents prefer to pur-sue their self-interest over the interests of the principals, goal conflicts are resolved through compensation based on performance (Eisenhardt, 1988). Consequently, SAS’s removal of compensation to all travel agencies, along with their effort to ex-clude non-IATA travel agencies from the business relationship, are probably other reasons for the fairly high level of reported goal incompatibility among travel agen-cies. As one non-IATA reseller put it, “I am doing SAS a favour, but instead of rewarding me, they are charging me.” (see Chapter 5.4.3). Ducati, on the other hand, granted its resellers a reduced commission on their online sales of motorcy-cles (see Chapter 2.4.1, p. 53). That is, in order to purchase a motorcycle online, customers had to choose a reference dealer, who delivered the motorcycle to the end customer. The commission was reduced, since the resellers no longer were performing all the traditional services, such as stock holding and marketing. Ac-cordingly, resellers were compensated based on performance, i.e. for delivering the motorcycles to customers, handling customer expectations, etc. However, in con-trast to agency theory, Ducati also compensated resellers for online sales of accesso-ries and apparel, although their resellers were not involved at all in these sales. Hence, it seems as if Ducati compensates resellers even for the potential loss of sales. As previously stated, according to agency theory, compensation resolves goal conflicts with agents (cf. Eisenhardt, 1988). This seems to be true also in the Ducati case, since resellers of Ducati reported a significantly lower level of goal in-compatibility compared to SAS’s resellers.

The reasons discussed above can also likely explain why statistical support was found for the relationship between goal incompatibility and domain dissensus for SAS, but not for Ducati. Namely, SAS’s resellers seem to perceive that SAS’s ex-pressed goal of increasing its direct-to-the-customer Internet sales has brought about domain dissensus. In fact, from 2004 to 2006, the share of Internet bookings increased 10%, whereas the share of tickets booked via travel agencies decreased 3%. Accordingly, travel agencies yielded rather high means on the items measuring Internet-related domain dissensus, such as SAS’s Internet channel serves the same customers as they do, markets identical products, competes with the reseller, and

13 The agency theory has been applied in a number of settings, including inter-organisational rela-tionships (Eisenhardt, 1988).

______ CHAPTER 7 ______________________________________________________________

176

operates at the expense of the reseller’s channel; the means on those items were in general significantly higher for SAS than for Ducati (see Table 6.1, p. 145).

To sum up, SAS’s expressed goal of increasing its online sales, along with their de-cisions to remove the resellers’ commission on sales and to exclude non-IATA travel agencies from the relationship, are probably the main reasons for goal con-flict with their resellers. Ducati’s decisions to include resellers as much as possible in online sales, and to compensate them, are probably the main reasons for the rather low levels of goal incompatibility reported by their resellers.

Accordingly, in line with the results of previous research (Rosenberg & Stern, 1971), the findings of this research indicate partial support for the hypothesis that the level of incompatible goals among channel members is positively associated with the frequency of con-flicts. The findings further suggest partial support for the hypothesis that the level of goal in-compatibility among channel members is positively associated with the level of Internet-related domain dissensus.

Domain Dissensus Regarding domain dissensus as a cause of channel conflict, only items associated with Internet-related domain dissensus could be retained. Statistical support for the hypothesized relationship could not be found for Ducati, whereas such statistical support was found for SAS. The probable reason for the contradictory results is the same as the one discussed in the previous section, i.e. that Ducati’s resellers are in-volved in the transactions, whereas SAS’s resellers are totally bypassed by SAS’s Internet sales. As a result, as previously stated, SAS’s resellers in general reported a significantly higher level of domain dissensus than did Ducati’s resellers (see Table6.1, p. 145). In other words, Internet sales do not seem to give rise to much do-main dissensus for Ducati, whereas they seem to cause rather high levels of domain dissensus for SAS.

In the first study of this thesis, Ducati’s management did not think that their Inter-net sales caused much domain dissensus with resellers. Consequently, they did not think that domain dissensus caused much conflict with resellers. SAS’s manage-ment, on the other hand, claimed that domain dissensus was one of the major causes of channel conflict. It can therefore be concluded that the results of the re-seller study support the results of the producer study. The results of the reseller study are similar to those of Rosenberg and Stern (1971), although for obvious rea-sons their study did not address Internet-related domain dissensus.

Accordingly, the results of this research indicate partial support for the hypothesis that Inter-net-related domain dissensus is a cause of channel conflict.


177

Inadequate Communication In line with the findings of Etgar (1979), for both Ducati and SAS statistical sup-port was found for the hypothesized positive relationship between inadequate communication and level of channel conflict.

In fact, inadequate communication was the only statistically supported cause of channel conflict for Ducati. Nevertheless, the means for items measuring inade-quate communication were in general significantly higher for SAS than for Ducati. The fact that non-IATA travel agencies do not have access to SAS’s extranet (www.sassalesinfo.com) has likely contributed to the high mean. However, since it is impossible within the scope of this thesis to distinguish between non-IATA and IATA accredited travel agencies, (see Chapter 5.4.3, p. 136 for a further discussion) this factor could not be controlled for.

In both cases, statistical support was found for the hypothesized positive relation-ship between inadequate communication and goal incompatibility. In other words, the findings indicate that inadequate communication increases the likelihood of in-compatible goals among channel members.

It was also hypothesized that inadequate communication is positively associated with domain dissensus. Statistical support for this relationship was found for Ducati but not for SAS. The probable reason for this result is that for Ducati, the ex-pressed goal of Internet sales of motorcycles was to create publicity in order to build the brand (see Chapter 2, p. 49), even if some resellers interpreted it differ-ently. Some resellers got angry and claimed that they would have been able to sell motorcycles online just as well as Ducati did, if the company would just have given them the opportunity (see Chapter 2, p. 51). Hence, some of Ducati’s resellers ex-perienced domain-related conflict, such as, Ducati’s direct-to-the-customer Inter-net sales competed with, and came at the expense of, the resellers, since Ducati probably did not communicate sufficiently the real intention of the Internet sales, which was to build the brand rather than dis-intermediate. Accordingly, for Ducati, it seems as if the perception of domain dissensus derived from inadequate communication – not incompatible goals. For SAS, on the other hand, the hy-pothesized relationship between inadequate communication and domain dissensus was not statistically supported. Still, travel agencies reported significantly higher levels of perceived domain dissensus on most items, as compared to Ducati’s resel-lers. As previously discussed, the relationship between incompatible goals and do-main dissensus was statistically supported for SAS, because, among other reasons, the goal of SAS’s Internet sales is to compete with travel agencies. It is therefore likely that it does not matter how well SAS communicates with its resellers – better communication would not lead to any changes in the underlying goal incompati-bility, which seems to be the major cause of SAS’s domain dissensus.

http://www.sassalesinfo.com

______ CHAPTER 7 ______________________________________________________________

178

In other words, in both cases support was found for a positive relationship between inadequate communication and level of channel conflict. Moreover, in both cases it was found that inade-quate communication increases the likelihood of incompatible goals between channel members. Furthermore, the findings provide partial support for a positive relationship between inade-quate communication and domain dissensus, i.e. for Ducati the majority of the domain dis-sensus was likely caused by misunderstandings. For SAS, statistical support was found in-stead for a positive relationship between channel members’ incompatible goals and domain dissensus. In such instances, improved communication will probably not lead to major changes in channel members’ perceptions of incongruent claims on the domain.

7.1.2 Level of Channel Conflict General items were used to measure level of channel conflict, since it was judged important to be able to use the same items for both channel settings (see Section5.4.2, p. 133). Resellers’ perceptions of both frequency (F) and intensity (N) across a representative set of issues were measured (see Table 6.1, p. 145). Overall, it was found that in both cases reseller-producer relationships were governed by some conflict between the parties, although the perceived level of conflict was rather low. More specifically, resellers of both companies perceived that, on average, dis-agreements were rather infrequent on information exchanges, product aspects, pricing is-sues, payment terms, sales commission, producer’s sales organisation, advertising and other promotional activities, and resellers’ sales effort. Likewise, the disagreements reported by resellers of both companies on these items were rather weak.

On the other hand, SAS’s resellers reported significantly higher levels of disagree-ment on producer’s Internet sales when channel was introduced, and producer’s present Internet sales, than did Ducati’s reseller. In fact, for SAS the means for frequency and intensity of the producer’s Internet sales when the channel was introduced were even higher than the midpoint of the scale (F = 4.64; N = 4.66), compared to the rather low levels of disagreement reported by Ducati’s resellers (F = 2.61; N = 2.72). Similar results were reported for producer’s present Internet sales (F = 4.50, N = 4.28; F = 2.61, N = 2.59 for SAS and Ducati, respectively).

Nevertheless, although, as previously noted, Brown and Day (1981) used several different combinations of frequency and intensity in their comparisons of different measures of channel conflict, none of their proposed measures rendered attractive properties for Structural Equation Modelling. Hence, it was decided to use only the frequency of channel conflict to test the research model (see Section 6.4.2, p. 151 for a further discussion).

Brown and Day (1981) conducted an un-rotated, exploratory factor analysis on each of their proposed measures of channel conflict. They concluded that the measures were unidimensional, although all the items did not load on a single fac-tor. In this study, an un-rotated exploratory factor analysis yielded a one-factor so-


179

lution. As the variance extracted for the frequency measure was much higher than that reported by Brown and Day (1981) (51.2% vs. 42.8%), it was decided to pro-ceed with the one-factor solution. The confirmatory factor analysis, however, re-vealed that the first order construct fit the data poorly; instead, level of channel conflict was found to consist of three dimensions: communication-related conflict,monetary-related conflict, and Internet-related conflict.

A respectable amount of the channel conflict variance was explained by its causes. For Ducati, the causes of channel conflict explained 33 percent of the variance in the level of channel conflict, and for SAS they explained 22 percent.

7.1.3 Relationship between Level of Channel Conflict and Outcomes Regarding the relationship between level of channel conflict and channel out-comes, it was decided to test first the outcomes part of the research model (i.e. the hypothesized relationships between the level of channel conflict, performance, and satisfaction) for SAS (see Figure 6.3, p. 162). A moderate to substantial negative re-lationship was found between level of conflict and performance, whereas the rela-tionship between level of channel conflict and satisfaction was non-significant. A substantial positive relationship was found between performance and satisfaction. The level of channel conflict explained 9 percent of the performance variance.

Nevertheless, when the causes of channel conflict were included in the model, moderate to substantial negative relationships were found between goal incompatibil-ity and performance and between inadequate communication and performance, whereas the relationship between level of channel conflict and performance was found to be non-significant. It is noteworthy that these results apply to both cases. For Ducati, 40 percent of the variance in performance is explained, compared to 53 percent for SAS (see Figure 6.5, p. 165). Moreover, in both cases, a substantial positive rela-tionship was found between performance and satisfaction.

The results of this study therefore indicate that goal incompatibility and inadequate communi-cation, rather than level of conflict, are negatively associated with performance. The results further suggest that performance is positively associated with satisfaction.

7.2 Limitations, Contributions and Conclusions his research contributes in several ways to the existing body of knowledge. However, since some of the limitations of the study might have affected the results, they are discussed in this section. Thereafter, some theoretical

contributions based on the results are presented, including an empirically derived model. Finally, the managerial implications of the study are presented.

T

______ CHAPTER 7 ______________________________________________________________

180

7.2.1 Limitations of the Study The first limitation stems from the fact that only two companies (Ducati and SAS) were involved in the study. The companies were selected because, in addition to using resellers, they sell directly to end customers via the Internet. Another reason for selecting these companies is they are different types; i.e. Ducati is a manufac-turer and SAS is a service provider. In addition, both the producers’ and the resel-lers’ perspectives were studied; this has rarely been done in previous research (see Appendix 2). Nevertheless, as a result of the research design, the sample size for Ducati is rather limited (N = 43). Although comparable to sample sizes used in some previous studies of channel conflict (e.g., Brown & Frazier, 1978; Rosson & Ford, 1980), the limited size of this sample decreases the probability of obtaining statistical significance for real relationships (cf. Cohen, 1990). This low power could partly explain the differences between the results of the hypothesis tests for the two cases. It should be noted, however, that one of the hypotheses, i.e. that in-adequate communication is positively associated with domain dissensus, was sup-ported in the case with the limited sample size but not in the other case.

Second, one reason for the lower levels of domain dissensus, goal incompatibility and Internet-related conflict for Ducati could be that they did not sell motorcycles online during the last few years. However, the average duration of resellers’ rela-tionships with Ducati was close to ten years, and almost all the resellers had a rela-tionship with Ducati when Ducati sold motorcycles online. Consequently, it is reasonable to assume that resellers would have reported higher levels of conflict on the item disagreements on producer’s direct-to-the-customer Internet sales when channel was introduced, provided that they considered the producer’s online sales of motorcycles to be a genuine problem.

Finally, in the reseller study, the key-informant approach (Sullivan, 2001) was used to collect data, i.e. the president/CEO was addressed. Although some researchers have questioned the ability of single informants to provide reliable data in self-report studies (e.g., Seidler, 1974), more recent research has found that single in-formants from senior management provide reliable and valid data (cf. Zahra & Covin, 1993).

7.2.2 Theoretical Contributions Although goal incompatibility, domain dissensus, and differing perceptions of real-ity are presently accepted as the major causes of channel conflict (cf. Coughlan et al., 2001), the results of this study confirm the results from the producer study, i.e. that inadequate communication probably should replace differing perceptions of reality as one of the primary causes of conflict, since differing perceptions of reality seem to be the result of inadequate communication.


181

Furthermore, to the researcher’s knowledge, the reseller study constitutes the first attempt to investigate the relationships among the causes of channel conflict. The results of this study suggest that the causes in fact are inter-related. For both cases, substantial statistical support was found for the hypothesized positive relationship between inadequate communication and goal incompatibility. Moreover, for SAS, statistical support was found for the hypothesized positive relationship between goal incompatibility and domain dissensus, whereas for Ducati, statistical support was found for the hypothesized positive relationship between inadequate commu-nication and domain dissensus. The probable reason for the contradictory results is that, whenever inadequate communication is related to domain dissensus, most of the domain dissensus is likely caused by misunderstandings. On the contrary, when channel members’ incompatible goals lead to domain dissensus, the latter is most likely caused by channel members’ incompatible claims on the domain. In the lat-ter case, improved communication is not likely to lead to any major changes in the channel members’ claims.

Although it has been argued in conceptual papers that the addition of an Internet channel leads to channel conflict (Smith et al., 1999; Gilbert & Bacheldor, 2000), the present study is probably the first to test this assumption empirically. Indeed, for SAS, significant support was found for the hypothesized positive relationships between Internet-related domain dissensus and level of channel conflict, and be-tween goal incompatibility and level of channel conflict. However, Ducati, statisti-cal support could not be found for either of these relationships. The probable rea-son for the contradictory results is that SAS’s resellers were totally bypassed by the producer’s Internet sales, which threatened the very existence of the reseller chan-nel; thus, resellers perceived conflict. Ducati’s resellers were involved in the pro-ducer’s Internet sales, in that they received a reduced commission on the pro-ducer’s Internet sales as well as contact information on the end customers. Hence, the results indicate that if producers handle channels pro-actively, and they involve resellers in their Internet sales from the beginning, i.e. apply channel integration, conflict can be minimized.

Concerning the measurement of channel conflict, it was found that the different combinations of frequency and intensity suggested by Brown and Day (1981) did not perform well in the confirmatory factor analysis of the present study. Thus, only frequency was used to measure the level of channel conflict. Moreover, previ-ous research has treated channel conflict as a unidimensional construct (e.g., Brown & Day, 1981). The present research, however, revealed that channel conflict is a multi-dimensional construct. That is, three dimensions were identified: communica-tion-related conflict, monetary-related conflict, and Internet-related conflict.

Regarding the hypothesized negative relationship between channel conflict and performance, when the underlying causes of conflict were not included in the model, there was a moderate to substantial negative relationship between channel

______ CHAPTER 7 ______________________________________________________________

182

conflict and performance. However, when the causes of conflict were included in the model, moderate to substantial negative relationships were found between two underlying causes of conflict, (goal incompatibility and inadequate communication), and resellers’ perceived performance, whereas the relationship between channel conflict and performance became insignificant. In other words, unlike previous research (Lusch, 1976a; Pearson & Monoky, 1976; Kelly & Peters, 1977; Cronin & Baker, 1993; Webb & Hogan, 2002), the reseller study failed to find statistical support for the hypothesized negative relationship between level of channel conflict and perceived performance. Likewise contrary to previous research (Dwyer, 1980; Arndt & Ogaard, 1986; Frazier et al., 1989; Anderson & Narus, 1990; Cronin & Baker, 1993; Ganesan, 1993; Rawwas et al., 1997; Webb, 1997; Lee, 2001; Webb & Ho-gan, 2002; Leonidou et al., 2006), the present study failed to find statistical support for the hypothesized negative relationship between level of channel conflict and satisfaction. It should be noted that similar results were found for both cases. It can therefore be concluded that conflict per se does not significantly affect resellers’ per-ceived performance or satisfaction. Since causes are the underlying bases of con-flict, whereas disagreements are symptoms of the causes (Ong et al., 1990), unless disagreements lead to changes in the underlying cause or causes of the conflict, re-sellers’ perceived performance will continue to be negatively affected by these causes. To the researcher’s knowledge, the reseller study constitutes the first at-tempt to investigate the direct relationships between inadequate communication and perceived performance, and between goal incompatibility and perceived per-formance, in multiple marketing channels.

Empirically Derived Model An empirically derived model that incorporates what was learned from the results of the reseller study and the previous discussion is presented in Figure 7.1. A sum-mary of the proposed relationships is presented in Table 7.2. To facilitate compari-son with the original research model (see Figure 4.1, p. 126) the propositions have the same numbers as the corresponding original hypotheses. Therefore, the most noteworthy changes to the original model are discussed at the end of this section.

Similar with the findings of Rosenberg and Stern (1971), the results of the reseller study suggest partial support for goal incompatibility as a cause of channel conflict. Rosenberg and Stern (1971) concluded that they found partial support concerning the causes of conflict because the issues “may have been resolved through the par-ticular dynamics of each relationship” (p. 441). This seems to be true also in the present study. As discussed previously, for Ducati the relationship was insignificant, probably since Ducati’s resellers were involved in Ducati’s Internet sales. Hence, the first proposition:

P1: The level of goal incompatibility among multiple marketing channels is positively as-sociated with the level of channel conflict.


183

Figure 7.1 Empirically Derived Model

Table 7.2 Overview of Propositions

P.no.

Proposed Relationship

P1 The level of goal incompatibility among multiple marketing channels is positively associ-ated with the level of channel conflict.

P2 The level of goal incompatibility among multiple marketing channels is positively associ-ated with the level of domain dissensus.

P3 The level of domain dissensus among multiple marketing channels is positively associated with the level of channel conflict.

P4 The level of inadequate communication among multiple marketing channels is positively associated with the level of goal incompatibility.

P5 The level of inadequate communication among multiple marketing channels is positively associated with the level of domain dissensus.

P6 The level of inadequate communication among multiple marketing channels is positively associated with the level of channel conflict.

P7 The level of goal incompatibility among multiple marketing channels is negatively associ-ated with the level of perceived performance.

P8 The level of inadequate communication among multiple marketing channels is negatively associated with the level of perceived performance.

P9 The level of perceived performance among multiple marketing channels is positively asso-ciated with the level of perceived satisfaction with the channel system.

The results from the reseller study also indicate partial support for the positive rela-tionship between goal incompatibility and the level of domain dissensus. The probable reason for this outcome is the same as discussed in relation to Proposition one above. Accordingly, the second proposition:

GOALINCOMPATIBILITY

INADEQUATE COMMUNICATION

DOMAIN DISSENSUS

LEVEL OFCHANNEL CONFLICT

PERFORMANCE

SATISFACTION

P1+

P2+ P3+

P4+

P5+

P6+

P8-

P7-

P9+

______ CHAPTER 7 ______________________________________________________________

184

Both the producer study and the reseller study provide support for Proposition three, i.e. domain dissensus was found to be a cause of channel conflict, although the reseller study only provided partial support for this relationship. The probable reason for this outcome is the same as discussed in relation to Proposition one. Ac-cordingly:

Concerning Proposition four, statistical support was found in both cases for the hypothesized positive relationship between inadequate communication and goal incompatibility. In other words, the findings indicate that inadequate communica-tion increases the likelihood of incompatible goals among channel members. The probable reason for this is that inadequate communication leads to misunderstand-ings, which leads resellers to adopt faulty goals that are incompatible with the pro-ducers’ goals, which in turn leads to goal-related conflict. This reasoning is in line with Etgar (1979), who argues that inadequate communications “often lead to misunderstandings, incorrect strategies, and mutual feelings of frustration” (p. 65). Following from this discussion, the fourth proposition:

Regarding Proposition five, in the reseller study support was found for Ducati for the hypothesized positive relationship between inadequate communication and Internet-related domain dissensus. Even though the hypothesized relationship was non-significant for SAS it is plausible that some domain dissensus occurred there as well because of inadequate communication, although the measures used in this re-search did not capture it. Hence, the following proposition:

In line with the findings of Etgar (1979), the hypothesized positive relationship be-tween inadequate communication and level of channel conflict was supported in both cases by the reseller study. In addition, the results support the findings from the producer study. Accordingly, the following proposition:

P2: The level of goal incompatibility among multiple marketing channels is positively as-sociated with the level of domain dissensus.

P3: The level of domain dissensus among multiple marketing channels is positively asso-ciated with the level of channel conflict.

P4: The level of inadequate communication among multiple marketing channels is posi-tively associated with the level of goal incompatibility.

P5: The level of inadequate communication among multiple marketing channels is posi-tively associated with the level of domain dissensus.

P6: The level of inadequate communication among multiple marketing channels is posi-tively associated with the level of channel conflict.


185

The most significant changes to the model involve the removal of the relationships between level of channel conflict and performance, level of channel conflict and satisfaction, and the proposed negative relationships between goal incompatibility and performance, and between inadequate communication and performance. Sub-stantial statistical support was found for goal incompatibility and inadequate com-munication, rather than level of channel conflict, having a negative influence on resellers’ perceived performance. It is noteworthy that similar results were found for both cases.

As previously noted, the same measures were used in both channel settings in order to capture level of channel conflict. Thus, one possible explanation for the non-significant relationships between level of channel conflict and performance, and be-tween level of channel conflict and satisfaction, is that the measures were too im-precise and therefore did not capture the conflict. However, since the causes of conflict explained considerable portions of the variance in the level of channel conflict (22 percent and 33 percent, respectively) the measures do not seem to be the reason for the non-significant result. Instead, since disagreements are just symp-toms of the underlying causes (Ong et al., 1990), a more plausible explanation is that unless disagreements lead to a change in the underlying cause or causes, the basis for conflict will not disappear. Resellers have a number of options for dealing with this problem, e.g., end the business relationship, argue even more, behave in their own interest, or simply comply with the situation. Today, the airline industry is distinguished by its refusal to give a commission to travel agencies and by its di-rect sales of tickets to end customers via the Internet, issues that are related to travel agencies’ performance. Although travel agencies reported rather high levels of conflict on these issues, such conflicts are unlikely to lead to any changes in the underlying causes. So, even though travel agencies are not pleased with the situa-tion (e.g., they reported rather high levels of goal incompatibility), if they want to remain in the travel business they simply have to adjust to the fact that SAS is competing with them and no longer gives them a commission on sales. In other words, disagreements about the eliminated commission on sales and SAS’s Internet sales have not led to any changes in the underlying causes of conflict, (goal incom-patibility and domain dissensus), and therefore the conflict remains, which in turn affects resellers’ perceived performance.

As regards Proposition seven, moderate to substantial statistical support was found for both cases for the hypothesized negative relationship between goal incompati-bility and perceived performance. Although the study investigated whether pro-ducer’s and reseller’s goals are compatible, it did not investigate the nature of these goals. Achrol and Etzel (2003) argue that over time a number of goals (cost mini-mization, survival, growth, customer service, meeting competition, market share, cash flow, supplier relations, and so on) can be important to companies. They found that resellers’ performance is affected by the goals they adopt in relation to

______ CHAPTER 7 ______________________________________________________________

186

the market environment. It can therefore be assumed that, even though resellers are expected to act on producers’ behalf, the producer and resellers are separate businesses that wish to maximize their own performance and frequently function in different market environments; hence, it is likely that the goals of the parties are at least to a certain degree incompatible, and that resellers’ performance is thereby to some extent affected. Agency theory postulates that as long as agents receive com-pensation based on performance it is likely that goal incompatibility will be kept at low levels and, therefore, not give rise to much conflict (cf. Eisenhardt, 1988). However, most resellers carry multiple brands, so if they perceive that the rewards from selling these other brands exceed those from selling the brand of the focal producer, resellers may reject the products of the focal producer (Gilliland, 2004). Consequently, if resellers perceive that goal incompatibility with the focal producer makes it harder for them to achieve the expected performance on the focal pro-ducer's brand, as compared to competitors’ brands, they are likely to reject the former, which in turn will lead to decreased performance for the focal producer. For SAS, it was found that when it removed the commission on sales and intro-duced its online sales, some travel agencies tried to reject SAS and redirect their travel bookings to airlines that were still offering a commission (see Chapter 2.4.2, p. 62), an action that probably affected SAS’s performance. In other words, goal incompatibility between channel members might affect performance of both par-ties. Accordingly, the following proposition:

For both cases, moderate to substantial statistical support was found for the hy-pothesized negative relationship between inadequate communication and resellers’ perceived performance. As Kim et al. (2006) contend, “high-quality information exchange not only helps coordinate with channel partners but also improves re-sponsiveness of the partnership and market performance” (p. 52). For instance, in-formation about new products, promotional campaigns, and so on has to be trans-mitted from the producer to the resellers (Etgar, 1979). Inadequate communication can therefore be expected to lead to a lack of coordination. Since lack of coordina-tion leads to lost opportunities and low levels of customer satisfaction (Gilliland, 2004), it is hardly surprising that inadequate communication was found to nega-tively affect resellers’ performance. However, lost opportunities and low levels of customer satisfaction are also likely to affect the producers’ performance. Hence, Proposition eight:

P7: The level of goal incompatibility among multiple marketing channels is negatively as-sociated with the level of perceived performance.

P8: The level of inadequate communication among multiple marketing channels is nega-tively associated with the level of perceived performance.


187

Regarding the ninth proposition, in line with results from previous research (Cronin & Baker, 1993), statistical support was found in both cases for the hy-pothesized positive relationship between performance and satisfaction. Accord-ingly, the following proposition:

7.2.3 Managerial Implications From the findings of this study, a number of managerial implications can be identi-fied. The implications apply to companies using resellers in general, and in particu-lar to firms that complement their reseller networks with direct-to-the-customer Internet sales.

Goal Incompatibility & Domain Dissensus Channel members’ goals have seldom been included in research on marketing channels, since it often has been assumed that resellers’ goals are an extension of producers’ goals (Achrol & Etzel, 2003). However, in this research it was found that resellers perceived their goals and the focal producer’s goals to be relatively or even considerably incompatible. It was further found that incompatible goals had significant negative performance implications for resellers. So what implications do resellers’ perceptions of their own performance have for the producer? Since resel-lers often represent multiple brands, they are likely to reject a brand if they per-ceive that it is offers lesser rewards compared to other brands (Gilliland, 2004). Re-seller rejection, in turn, will have negative performance implications for the pro-ducer. Accordingly, goal incompatibility should not be taken lightly by producers.

Already from the beginning, Ducati’s resellers were involved in Ducati’s online sales. When customers ordered motorcycles online they had to choose a reference dealer, who delivered the motorcycle to the customer. Reference dealers received a reduced commission on these sales, since they did not have any marketing costs, etc. Reference dealers also received commissions on the producer’s Internet sales of apparel or accessories, although these goods were delivered directly to the custom-ers. It therefore seems as if Ducati compensates resellers for the potential loss of sales. In addition, the reference dealer received contact information on these cus-tomers. Travel agencies, on the other hand, were totally bypassed by SAS with re-gard to online sales. Moreover, travel agencies did not receive a commission on tickets that they sold. As could be expected, Ducati’s resellers reported much lower levels of goal incompatibility, domain dissensus, and disagreements regarding the producer’s Internet sales. It seems clear that when incompatible goals refer to the mere existence of a channel per se, channel members of the threatened channel will perceive conflict. Consequently, producers who are planning to sell directly to end customers in addition to using resellers can probably decrease the levels of goal in-

P9: The level of perceived performance among multiple marketing channels is positivelyassociated with the level of perceived satisfaction with the channel system.

______ CHAPTER 7 ______________________________________________________________

188

compatibility and domain dissensus by involving resellers in online sales from the beginning, i.e. channel integration. An example of channel integration is to use reference resellers, which receive both the customer contact information and a commission on the producer’s sales. They might even deliver the products to the end customers.

As the airline industry is distinguished by its lack of commission on sales and by producers who bypass resellers by using the Internet, it is hardly surprising that travel agencies reported rather high levels of goal incompatibility, domain dissen-sus, and disagreements concerning the producer’s Internet sales. So, if there were an option that offered travel agencies better rewards, the agencies would likely try to influence customers to choose that option. Depending on whose perspective one takes, this tactic offers either an opportunity or a threat to companies. For in-stance, the National Hockey League star Peter Forsberg and associates are planning to start an airline, Höga Kusten Flyg, in January, 2008. The airline will operate be-tween the Swedish cities of Stockholm and Örnsköldsvik. SAS is already serving this route. Even if Höga Kusten Flyg offered travel agencies just a small commis-sion on sales, it would probably influence them to advise customers to fly with Höga Kusten Flyg. Swedish law would probably prevent SAS from offering travel agencies a commission on sales, unless they offered it for all its Swedish routes. Similarly, if the Swedish train operator SJ started to offer travel agencies a commis-sion on sales, the agencies would probably advise customers to choose the high ve-locity train X2000 instead of flying for relatively short routes, such as Stockholm-Gothenburg14, especially since the time saved by flying is minimal. Travel agencies would probably raise additional arguments in favour of train travel, such as envi-ronmental impact.

Nevertheless, resellers and producers are separate businesses that wish to maximize their performance. Therefore, all goal incompatibility cannot, and should not, be avoided. But, producers likely can benefit from avoiding unnecessary goal incom-patibility. It was found in the present study that one major reason for incompatible goals among channel members is inadequate communication. The probable reason for this is that when communication is inadequate, misunderstandings occur. This means that resellers needlessly adopt faulty goals that are incompatible with the producer’s goals. One way for producers to avoid unnecessary goal incompatibility is to improve their communication with resellers.

Inadequate Communication George Bernard Shaw once said, “The greatest problem with communication is the illusion that it has been accomplished”. This seems to be true also in the com-

14 The distance between Stockholm and Gothenburg is 478 km, see www.lundberg-lagerstedt.se/distance/

http://www.lundberg-lagerstedt.se/distance




189

panies studied in this research, since resellers of both companies reported that the focal producer did not communicate well enough with them. It was further found that inadequate communication had a substantial effect on resellers’ perceived per-formance. Moreover, inadequate communication probably leads to a lack of coor-dination. Lack of coordination, in turn, leads to lost opportunities and low levels of customer satisfaction (Gilliland, 2004), and probably affects the producer’s per-formance as well. Inadequate communication is therefore likely to have negative performance implications for both resellers and producers. Although numerous re-searchers have stressed the importance of well-functioning communication in mar-keting channels (e.g., Anderson & Weitz, 1992; Kim et al., 2006), it seems that it cannot be overemphasized. The producer study indicates that a well-designed ex-tranet can support producers in providing resellers with good information.

Level of Channel ConflictAs pointed out by Anderson and Narus (1990), “Even firms in successful partner-ships would readily acknowledge that disagreements are inevitable” (p. 56). None-theless, the traditional view of managers is that conflict should be reduced, or even avoided (Menon et al., 1996). Since the present research failed to find statistical support for the hypothesized negative relationship between level of channel con-flict and perceived performance, one could argue that managers do not need to consider channel conflict. But rather than ignore conflicts, managers should view them as symptoms of their underlying causes. With this perspective, conflicts actu-ally offer managers the opportunity to identify the core elements of the conflict and diminish them before they have significant performance implications for both parties in the channel relationship.

7.3 Suggestions for Future Research The results of this research question some of the premises in the channel conflict literature. Specifically, goal incompatibility and inadequate communication were found to have a direct and negative effect on resellers’ performance in both cases. In contrast, channel conflict per se was not detrimental to firms. Although the channels included in this research were of polar types (i.e. manufacturing company vs. service provider), both channel systems comprised multiple marketing channels, including the Internet. Future research should investigate whether the results of this research apply to other channel types and/or systems.

Since in both cases goal incompatibility and inadequate communication were found to negatively influence resellers’ perceived performance, future research should include in-depth interviews with resellers in order to deepen the under-standing of these relationships. In addition, future research should address the im-pact of these relationships on producers’ performance. Moreover, it was investi-gated in this research whether resellers perceive their goals and the producer’s goals

______ CHAPTER 7 ______________________________________________________________

190

to be compatible, but the nature of these goals was not explored. This is an issue for future research. Research is also needed on how communication within mar-keting channels can be improved. For instance, the results of the producer study indicate that an extranet can support producers in providing resellers with adequate information, and that conflict with resellers thereby is reduced. Hence, future re-search should address the effect of extranets on marketing channels, as well as the structure of well-designed extranets.

This research also sheds new light on the constitution of the channel conflict con-struct. Namely, channel conflict was found to be a multi-dimensional construct, the dimensions being communication-related conflict, monetary-related conflict, and Internet-related conflict. Although a substantial portion of the variance in channel conflict was explained (33 and 22 percent for Ducati and SAS, respec-tively), future research should develop additional dimensions of channel conflict. It should also develop additional items for each dimension, since conflict issues can likely be used to identify the causes of conflict that in turn influence performance negatively. One way to facilitate such work could be to focus on conflict within one channel setting.

Several other important constructs might be added to the model. For example, commitment is viewed by some researchers as essential to successful long-term re-lationships (e.g., Gundlach et al., 1995). Perceived interdependence is another im-portant construct in channel research (e.g., Kumar et al., 1995), since in most cases marketing channels are combinations of interdependent organisations. Finally, al-though some scholars argue that some conflict in marketing channels is needed to avoid passivity and lack of innovation (e.g., Boulding, 1965; Stern & Heskett, 1969), functionality of conflict is a construct that has been more or less neglected in the channel conflict literature. In fact, only one study could be found that in-cluded this concept (i.e. Anderson & Narus, 1990). Future research is therefore en-couraged to investigate whether conflict in marketing channels can be functional and, if so, in which situations.

_____________________________________________________________ REFERENCES ______

REFERENCES

Achrol, R. S., & Etzel, M. J. (2003). The Structure of Reseller Goals and Perform-ance in Marketing Channels. Journal of the Academy of Marketing Science, 31(2), 146-163.

Achrol, R. S., & Stern, L. W. (1988). Environmental Determinants of Decision-Making Uncertainty in Marketing Channels. Journal of Marketing Research, 25(February), 36-50.

Alba, J., Lynch, J., Weitz, B. A., Janiszewski, C., Lutz, R., Sawyer, A., et al. (1997). Interactive Home Shopping: Consumer, Retailer, and Manufac-turer Incentives to Participate in Electronic Marketplaces. Journal of Market-ing, 61(July), 38-53.

Ancarani, F. (2002). Pricing and the Internet: Frictionless Commerce or Pricer's Paradise? European Management Journal, 20(6), 680-687.

Ancarani, F., & Shankar, V. (2004). Price Levels and Price Dispersion Within and Across Multiple Retailer Types: Further Evidence and Extension. Journal of the Academy of Marketing Science, 32(2), 176-187.

Anderson, E., Day, G. S., & Rangan, V. K. (1997). Strategic Channel Design. Sloan Management Review, 38(4), 59-69.

Anderson, E., & Weitz, B. (1992). The Use of Pledges to Build and Sustain Commitment in Distribution Channels. Journal of Marketing Research, 29(1),18-34.

Anderson, J. C., & Gerbing, D. W. (1988). Structural Equation Modeling in Prac-tice: A Review and Recommended Two-Step Approach. Psychological Bul-letin, 103(3), 411-423.

Anderson, J. C., & Narus, J. A. (1984). A Model of the Distributor's Perspective of Distributor-Manufacturer Working Relationships. Journal of Marketing, 48(Fall), 62-74.

Anderson, J. C., & Narus, J. A. (1990). A Model of Distributor Firm and Manu-facturer Firm Working Partnerships. Journal of Marketing, 54, 42-58.

Armstrong, J. S., & Overton, T. S. (1977). Estimating Nonresponse Bias in Mail Surveys. Journal of Marketing Research, 14(August), 396-402.

Arndt, J., & Ogaard, T. (1986). A Comparative Study of Channel Conflict: Preliminary Findings. Paper presented at the 1986 Educators' Conference, T. A. Shimp, S. Sharma, W. Dillon, R. Dyer, M. Gardner, G. John, J. Lindgren & J. Quelch (ed.), 191-194. American Marketing Association.

Ashcraft, D. (2001). The multichannel pricing challenge. DSN Retailing Today, 40(1), 16.

Assael, H. (1969). Constructive Role of Interorganizational Conflict. AdministrativeScience Quarterly, 13(December), 537-582.

Baker, W., Marn, M., & Zawada, C. (2001). Price Smarter on the Net. HarvardBusiness Review, 79(2), 122-127.

______ REFERENCES _____________________________________________________________

Balasubramanian, S. (1998). Mail versus Mall: A Strategic Analysis of Competition Between Direct Marketers and Conventional Retailers. Marketing Science, 17(3), 181-195.

Boulding, K. E. (1965). The Economics of Human Conflict. In E. B. McNeil (ed.), The Nature of Human Conflict, 172-191. Englewood Cliffs, N.J.: Pren-tice-Hall, Inc.

Brown, J. R. (1977). Toward Improved Measures of Distribution Channel Conflict. Pa-per presented at the Contemporary Marketing Thought: 1977 Marketing Educators' Conference, Hartford, Connecticut, B. A. Greenberg & D. N. Bellenger (ed.), 385-389. American Marketing Association.

Brown, J. R. (1979). Methods of Channel Conflict Resolution: Some Empirical Results.Paper presented at the 1979 Educators' Conference, N. Beckwith, M. Houston, R. Mittelstaedt, K. B. Monroe & S. Ward (ed.), 495-499. American Marketing Association.

Brown, J. R., & Day, R. L. (1981). Measures of Manifest Conflict in Distribution Channels. Journal of Marketing Research, 18(August), 263-274.

Brown, J. R., & Fern, E. F. (1992). Conflict in marketing channels: the impact of dual distribution. International Review of Retail, Distribution & Consumer Re-search, 2(2 April), 121-132.

Brown, J. R., & Frazier, G. L. (1978). The Application of Channel Power: Its Effects and Connotations. Paper presented at the 1978 Educators' Conference, Chi-cago, Illinois, S. C. Jain (ed.), 266-270. American Marketing Association.

Brown, J. R., Lusch, R. F., & Muehling, D. D. (1983). Conflict and Power-Dependence Relations in Retailer-Supplier Channels. Journal of Retailing, 59(4), 53-80.

Brown, J. R., Lusch, R. F., & Smith, L. P. (1991). Conflict and Satisfaction in an Industrial Channel of Distribution. International Journal of Physical Distribu-tion & Logistics Management, 21(6), 15-25.

Brynjolfsson, E., & Smith, M. D. (2000). Frictionless Commerce? A Comparison of Internet and Conventional Retailers. Management Science, 46(4), 563-585.

Bucklin, C. B., Thomas-Graham, P. A., & Webster, E. A. (1997). Channel con-flict: When is it dangerous? The McKinsey Quarterly(3), 36-43.

Byrne, B. M. (2001). Structural Equation Modeling with AMOS: Basic Concepts, Ap-plications, and Programming. Mahwah, New Jersey: Lawrence Erlbaum Asso-ciates, Inc., Publishers.

Cadotte, E. R., & Stern, L. W. (1979). A Process Model of Interorganizational Relations in Marketing Channels. In J. N. Sheth (ed.), Research in Market-ing, 127-158. Greenwich, CT: JAI Press.

Carver, T. B., & Vondra, A. A. (1994). Alternative Dispute Resolution: Why It Doesn't Work and Why It Does. Harvard Business Review, 72(3), 120-130.

Cespedes, F. V., & Corey, E. R. (1990). Managing Multiple Channels. BusinessHorizons, 33(4), 67-77.

_____________________________________________________________ REFERENCES ______

Chatterjee, S., Jamieson, L., & Wiseman, F. (1991). Identifying Most Influential Observations in Factor Analysis. Marketing Science, 10(2), 145-160.

Chin, W. W. (1998). Issues and Opinion on Structural Equation Modeling. MISQuarterly, 22(1), vii-xvi.

Chisnall, P. M. (1997). Marketing Research (5th ed.). Cambridge (UK): University Press.

Churchill, G. A. J. (1979). A Paradigm for Developing Better Measures of Market-ing Constructs. Journal of Marketing Research, 16(February), 64-73.

Coelho, F., Easingwood, C., & Coelho, A. (2003). Exploratory evidence of chan-nel performance in single vs multiple channel strategies. International Journal of Retail & Distribution Management, 31(11), 561-573.

Cohen, J. (1990). Things I Have Learned (So Far). American Psychologist, 45(12),1304-1312.

Collett, S. (1999, November 8). Channel Conflicts Push Levi to Halt Web Sales. Computerworld, 33, 8.

Cortese, A. E., & Stepanek, M. (1998, May 4). Good-bye to fixed pricing? Busi-ness Week, 70-76.

Coughlan, A. T., Anderson, E., Stern, L. W., & El-Ansary, A. I. (2001). MarketingChannels (6th ed.). Upper Saddle River, New Jersey: Prentice-Hall, Inc.

Cronin, J. J., & Baker, T. L. (1993). The Effects of a Distributor's Attribution of Manufacturer Influence on the Distributor's Perceptions of Conflict, Per-formance and Satisfaction. Journal of Marketing Channels, 3(2), 83-110.

Cunningham, M. T., & Pyatt, T. R. (1989). Marketing and Purchasing Strategies in the Distribution Channels of Mid-range Computers. European Journal of Marketing, 23(2), 130-143.

Dant, R. P., & Schul, P. L. (1992). Conflict Resolution Processes in Contractual Channels of Distribution. Journal of Marketing, 56(January), 38-54.

Day, G. S., & Nedungadi, P. (1994). Managerial Representations of Competitive Advantage. Journal of Marketing, 58(April), 31-44.

de Vaus, D. (2002). Analyzing Social Science Data: 50 Key Problems in Data Analysis(1st ed.). London: SAGE Publications Ltd.

Denzin, N. K., & Lincoln, Y. S. (2000). The Discipline and Practice of Qualitative Research. In N. K. Denzin & Y. S. Lincoln (eds.), The Handbook of Quali-tative Research, 1-28. Thousand Oaks, California: Sage Publications, Inc.

Dilts, J. C. (1985). A Cross-Channel Comparison of Channel Conflict. Paper presented at the 1985 Educators' Conference, R. F. Lusch, G. Ford, G. L. Frazier, C. Ingene, R. Howell, M. Reilly & R. Stampfl (ed.), 166-171. American Marketing Association.

Duarte, M., & Davies, G. (2003). Testing the conflict-performance assumption in business-to-business relationships. Industrial Marketing Management, 32, 91-99.

Ducati. (2005). The MH900e. Retrieved 12 March, 2005, from http://www.ducati.com/heritage/anni00/storia.jhtml

http://www.ducati.com/heritage/anni00/storia.jhtml

______ REFERENCES _____________________________________________________________

Dwyer, F. R. (1980). Channel-Member Satisfaction: Laboratory Insights. Journal of Retailing, 56(2), 45-65.

Eisenhardt, K. M. (1988). Agency- and Institutional- Theory Explanations: The Case of Retail Sales Compensation. Academy of Management Journal, 31(3), 488-511.

Eisenhardt, K. M. (1989). Building Theories from Case Study Research. Academyof Management Review, 14(4), 532-550.

Eliashberg, J., & Michie, D. A. (1984). Multiple Business Goals Sets as Determi-nants of Marketing Channel Conflict: An Empirical Study. Journal of Mar-keting Research, 21(February), 75-88.

Etgar, M. (1979). Sources and Types of Intrachannel Conflict. Journal of Retailing, 55(Spring), 61-78.

Forrester. (2004). Retail First Look: Research & Event Highlights From Forrester. Re-trieved 5 October, from http://www.forrester.com/FirstLook/Vertical/Issue/0,6454,181,00.html

Foster, J. R., & Shuptrine, F. K. (1974). Using Retailers' Perceptions of Channel Performance to Detect Potential Conflict. In T. V. Greer (ed.), CombinedProceedings of 1973 Spring and Fall Conferences of the American Marketing Asso-ciation, 118-123. Chicago: American Marketing Association.

Frazier, G. L. (1999). Organizing and Managing Channels of Distribution. Journalof the Academy of Marketing Science, 27(2), 226-240.

Frazier, G. L., Gill, J. D., & Kale, S. H. (1989). Dealer Dependence Levels and Reciprocal Actions in a Channel of Distribution in a Developing Country. Journal of Marketing, 53(January), 50-69.

Frazier, G. L., & Rody, R. C. (1991). The Use of Influence Strategies in Interfirm Relationships in Industrial Product Channels. Journal of Marketing, 55(January), 52-69.

Frazier, G. L., Sawhney, K., & Shervani, T. (1990). Intensity, Functions, and Inte-gration in Channels of Distribution. In V. A. Zeithaml (ed.), Review of Marketing. Chicago, IL: American Marketing Association.

Frazier, G. L., & Summers, J. O. (1984). Interfirm Influence Strategies and Their Application within Distribution Channels. Journal of Marketing, 48(Summer).

Friberg, R., Ganslandt, M., & Sandström, M. (2000). E-Commerce and Prices - The-ory and Evidence (No. S57). Stockholm: Handelns utredningsinstitut (HUI).

Friedman, L. G., & Furey, T. R. (1999). The Channel Advantage: Going to market with multiple sales channels to reach more customers, sell more products, make more profit. Oxford: Butterworth-Heinemann.

Ganesan, S. (1993). Negotiation Strategies and the Nature of Channel Relation-ships. Journal of Marketing Research, 15(May), 183-203.

Gaski, J. F. (1984). The Theory of Power and Conflict in Channels of Distribu-tion. Journal of Marketing, 48(Summer), 9-29.

http://www.forrester.com/FirstLook/Vertical/Issue/0,6454,181,00.html

_____________________________________________________________ REFERENCES ______

Gaski, J. F., & Nevin, J. R. (1985). The Differential Effects of Exercised and Un-exercised Power Sources in a Marketing Channel. Journal of Marketing Re-search, 22(May), 130-142.

Gattorna, J. (1978). Channels of Distribution Conceptualisations: A State-of-the-Art Review. European Journal of Marketing, 12(7), 471-512.

Gerbing, D. W., & Anderson, J. C. (1988). An Updated Paradigm for Scale De-velopment Incorporating Unidimensionality and Its Assessment. Journal of Marketing Research, 25(May), 186-192.

Geyskens, I., Steenkamp, J.-B. E. M., & Kumar, N. (1999). A Meta-Analysis of Satisfaction in Marketing Channel Relationships. Journal of Marketing Re-search, 36(May), 223-238.

Gilbert, A., & Bacheldor, B. (2000, March 27). The Big Squeeze. Informationweek,47-52.

Gilliland, D. I. (2004). Designing channel incentives to overcome reseller rejec-tion. Industrial Marketing Management, 33, 87-95.

Girishankar, S. (1998). IBM To Sell Product Line Over Web. Retrieved 17 Novem-ber, 2004, from http://www.techweb.com/wire/story/TWB19980219S0015

Goldkuhl, L. (2005). Multiple Marketing Channel Conflict with a Focus on the Internet.Unpublished Licentiate Thesis, Luleå University of Technology, Luleå.

Green, S. B., & Salkind, N. J. (2003). Using SPSS for Windows and Macintosh: Ana-lyzing and Understanding Data (3rd ed.). Upper Saddle River, NJ: Prentice Hall.

Gulati, R., & Garino, J. (2000). Get the Right Mix of Bricks & Clicks. HarvardBusiness Review, 78(3), 107-114.

Gundlach, G. T., Achrol, R. S., & Mentzer, J. T. (1995). The Structure of Com-mitment in Exchange. Journal of Marketing, 59(January), 78-92.

Gundlach, G. T., & Cadotte, E. R. (1994). Exchange Interdependence and Inter-firm Interaction: Research in a Simulated Channel Setting. Journal of Mar-keting Research, 31(November), 516-532.

Hair, J. F. J., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). MultivariateData Analysis (5th ed.). London: Prentice-Hall International (UK) Limited.

Hair, J. F. J., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate Data Analysis (6th ed.). Upper Saddle River: Pearson Prentice Hall.

Hair, J. F. J., Money, A. H., Samouel, P., & Page, M. (2007). Research Methods for Business. Chichester, England: John Wiley & Sons Ltd.

Heide, J. B., & John, G. (1992). Do Norms Matter in Marketing Relationships? Journal of Marketing, 56(2), 32-44.

Hilleke, K., & Butscher, S. A. (1997). How to use a two-product strategy against low-price competition. Pricing Strategy & Practice, 5(3), 108-115.

Hulin, C. (2001). Measurement. Journal of Consumer Psychology, 10(1&2), 55-69.

http://www.techweb.com/wire/story/TWB19980219S0015

______ REFERENCES _____________________________________________________________

Hunger, J. D., & Stern, L. W. (1976). An Assessment of the Functionality of the Superordinate Goal in Reducing Conflict. Academy of Management Journal, 19(4 (Dec.)), 591-605.

Hunt, S. D., & Nevin, J. R. (1974). Power in a Channel of Distribution: Sources and Consequences. Journal of Marketing Research, 11(May), 186-193.

Iacobucci, D. (2001). Measurement: Why use alpha if it is not a good measure of unidimensionality? Journal of Consumer Psychology, 10(1&2), 59-62.

Jelassi, T., & Leenen, S. (2003). An E-Commerce Sales Model for Manufacturing Companies: A Conceptual Framework and a European Example. EuropeanManagement Journal, 21(1), 38-47.

John, G., & Reve, T. (1982). The Reliability and Validity of Key Informant Data from Dyadic Relationships in Marketing Channels. Journal of Marketing Re-search, 19(November), 517-524.

Katsikeas, C. S. (1992). The Process of Conflict in Buyer-Seller Relationships at Domestic and International Levels: a Comparative Analysis. Journal of Mar-keting Management, 8, 365-381.

Katsikeas, C. S., & Piercy, N. F. (1991). The Relationship between Exporters from a Developing Country and Importers based in a Developed Country: Conflict Considerations. European Journal of Marketing, 25(1), 6-25.

Kelly, J. S., & Peters, J. I. (1977). Vertical Conflict: A Comparative Analysis of Franchi-sees and Distributors. Paper presented at the Contemporary Marketing Thought: 1977 Marketing Educators' Conference, Hartford, Connecticut, B. A. Greenberg & D. N. Bellenger (ed.), 380-384. American Marketing Association.

Ketilson, L. H. (1991). An examination of the use of legitimate power in market-ing channels. International Review of Retail, Distribution & Consumer Research, 1(5), 527-548.

Kiang, M. Y., Raghu, T. S., & Shang, K. H.-M. (2000). Marketing on the Inter-net - who can benefit from an online marketing approach? Decision Support Systems, 27, 383-393.

Kim, D., Cavusgil, S. T., & Calantone, R. J. (2006). Information System Innova-tions and Supply Chain Management: Channel Relationships and Firm Performance. Journal of the Academy of Marketing Science, 34(1), 40-54.

King, J. (1999). New covenants ease online channel war. Computerworld, 33, 24. Kochan, T. A., Huber, G. P., & Cummings, L. L. (1975). Determinants of In-

traorganizational Conflict in Collective Bargaining in the Public Sector. Administrative Science Quarterly, 20(1), 10-23.

Kotler, P. (1988). Marketing Management: Analysis, Planning, Implementation, and Control (6th ed.). Englewood Cliffs, New Jersey: Prentice Hall.

Kumar, N. (2005). The power of power in supplier-retailer relationships. IndustrialMarketing Management, 34, 863-866.

_____________________________________________________________ REFERENCES ______

Kumar, N., Scheer, L. K., & Steenkamp, J.-B. E. M. (1995). The Effects of Per-ceived Interdependence on Dealer Attitudes. Journal of Marketing Research, 32(August), 348-356.

Kung, M., Monroe, K. B., & Cox, J. L. (2002). Pricing on the Internet. Journal of Product & Brand Management, 11(5), 274-287.

Lee, D. Y. (2001). Power, conflict, and satisfaction in IJV supplier-Chinese dis-tributor channels. Journal of Business Research, 52, 149-160.

Lee, H. G. (1998). Do Electronic Marketplaces Lower the Price of Goods? Com-munications of the ACM, 41(1), 73-80.

Leenen, S., & Jelassi, T. (2005). Ducati Motorcycles (Italy): Riding traditional business channels or racing through the Internet. In T. Jelassi & A. Enders (eds.), Strategies for e-Business: Creating Value through Electronic and Mobile Commerce, 371-385. Essex, England: Pearson Education Limited.

Leonidou, L. C., Barnes, B. R., & Talias, M. A. (2006). Exporter-importer rela-tionship quality: The inhibiting role of uncertainty, distance, and conflict. Industrial Marketing Management, 35, 576-588.

Levy, M., & Weitz, B. A. (2004). Retailing Management (5th ed.). Boston, Mass.: McGraw-Hill Irwin.

Li, J., Mitra, J., & Matlay, H. (2004). E-commerce and management of channel conflict: evidence from small manufacturing firms in the UK. InternationalJournal of Technology Management, 28(7/8), 747-766.

Li, Z. G., & Dant, R. P. (1997). An Exploratory Study of Exclusive Dealing in Channel Relationships. Journal of the Academy of Marketing Science, 25(3),201-213.

Little, R. J. A. (1988). A Test of Missing Completely at Random for Multivariate Data with Missing Values. Journal of American Statistical Association, 83(404),1198-1202.

Lusch, R. F. (1976a). Channel Conflict: Its Impact on Retailer Operating Per-formance. Journal of Retailing, 52(2), 3-12, 89-90.

Lusch, R. F. (1976b). Sources of Power: Their Impact on Intrachannel Conflict. Journal of Marketing Research, 13(November), 382-390.

Lusch, R. F., & Brown, J. R. (1996). Interdependency, Contracting, and Rela-tional Behavior in Marketing Channels. Journal of Marketing, 60(October),19-38.

Mack, R. W., & Snyder, R. C. (1957). The analysis of social conflict - toward an overview and synthesis. Journal of Conflict Resolution, 1(June), 212-248.

Magrath, A. J., & Hardy, K. G. (1987). Avoiding the Pitfalls in Managing Distri-bution Channels. Business Horizons, 30(September-October), 29-33.

Marshall, C., & Rossman, G. B. (1999). Designing Qualitative Research (3rd ed.). Thousand Oaks: SAGE Publications, Inc.

McDonald, L. (1999). Managing Channel Conflict. Mortgage Banking, 60(1), 88-99.

______ REFERENCES _____________________________________________________________

McQuiston, D. H. (2001). A Conceptual Model for Building and Maintaining Relationships between Manufacturers' Representatives and Their Princi-pals. Industrial Marketing Management, 30, 165-181.

Menon, A., Bharadwaj, S. G., & Howell, R. (1996). The Quality and Effective-ness of Marketing Strategy: Effects of Functional and Dysfunctional Con-flict in Intraorganizational Relationships. Journal of the Academy of Marketing Science, 24(4), 299-313.

Miles, M. B., & Huberman, A. M. (1994). Qualitative Data Analysis: An Expanded Sourcebook (2nd ed.). Thousand Oaks: SAGE Publications, Inc.

Mohr, J., & Spekman, R. (1994). Characteristics of Partnership Success: Partner-ship Attributes, Communication Behavior, and Conflict Resolution Tech-niques. Strategic Management Journal, 15(2), 135-152.

Mohr, J. J., Fisher, R. J., & Nevin, J. R. (1996). Collaborative Communication in Interfirm Relationships: Moderating Effects of Integration and Control. Journal of Marketing, 60(July), 103-115.

Mohr, J. J., & Nevin, J. R. (1990). Communication Strategies in Marketing Channels: A Theoretical Perspective. Journal of Marketing, 54(October), 36-51.

Mols, N. P. (2000). The Internet and services marketing - the case of Danish retail banking. Internet Research: Electronic Applications and Policy, 10(1), 7-18.

Moore, R. A. (1990). The Conflict Gap in International Channel Relationships. Journal of Marketing Management, 6(3), 225-237.

Moriarty, R. T., & Moran, U. (1990). Managing Hybrid Marketing Systems. Har-vard Business Review, 68(6), 146-155.

Nicholls, J. A. F., Roslow, S., & Laskey, H. A. (1993). Perceptual and behavioural channel conflict: are they related? International Review of Retail, Distribution & Consumer Research, 3(1), 19-33.

O'Connor, P., & Frew, A. J. (2002). The Future of Hotel Electronic Distribution. Cornell Hotel and Restaurant Administration Quarterly, 43(3), 33-45.

Oliver, R. L. (1980). A Cognitive Model of the Antecedents and Consequences of Satisfaction Decisions. Journal of Marketing Research, 17(November), 460-469.

Ong, C., Elliott, G. R., & Armstrong, R. W. (1990). Power and intra-channel conflict in Australian petroleum retailing. International Review of Retail, Dis-tribution & Consumer Research, 1(1), 37-54.

Pearson, M., & Monoky, J. F. (1976). The Role of Conflict and Cooperation in Chan-nel Performance. Paper presented at the 1976 Marketing Educators' Confer-ence, Memphis, Tennessee, K. L. Bernhardt (ed.), 240-244. American Marketing Association.

Pearson, M. M. (1973). The Conflict-Performance Assumption. Journal of Purchas-ing, 9(February), 57-69.

_____________________________________________________________ REFERENCES ______

Perry, J. L., & Levine, C. H. (1976). An Interorganizational Analysis of Power, Conflict, and Settlements in Public Sector Collective Bargaining. TheAmerican Political Science Review, 70(4), 1185-1201.

Peterson, R. A., Balasubramanian, S., & Bronnenberg, B. J. (1997). Exploring the Implications of the Internet for Consumer Marketing. Journal of the Academy of Marketing Science, 25(4), 329-346.

Pitt, L. F., Berthon, P., & Berthon, J.-P. (1999). Changing Channels: The Impact of the Internet on Distribution Strategy. Business Horizons, 42(2), 19-28.

Pitt, L. F., Berthon, P., Watson, R. T., & Ewing, M. (2001). Pricing Strategy and the Net. Business Horizons, 44(2), 45-54.

Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Com-mon Method Biases in Behavioral Research: A Critical Review of the Lit-erature and Recommended Remedies. Journal of Applied Psychology, 88(5),879-903.

Podsakoff, P. M., & Organ, D. W. (1986). Self-Reports in Organizational Re-search: Problems and Prospects. Journal of Management, 12(4), 531-544.

Pondy, L. R. (1967). Organizational Conflict: Concepts and Models. AdministrativeScience Quarterly, 12(2), 296-320.

Potter, W. J. (1996). An Analysis of Thinking and Research About Qualitative Methods.Mahwah, New Jersey: Lawrence Erlbaum Associates, Inc.

Price, R. A. (1993). Effects of channel leadership behaviours under different levels of channel member interdependence. International Review of Retail, Distribu-tion & Consumer Research, 3(3), 261-278.

Pruden, H. O. (1969). Interorganizational Conflict, Linkage, and Exchange: A Study of Industrial Salesmen. Academy of Management Journal, 12(3), 339-350.

Rangaswamy, A., & Van Bruggen, G. H. (2005). Opportunities and Challenges in Multichannel Marketing: An Introduction to the Special Issue. Journal of In-teractive Marketing, 19(2), 5-11.

Rawwas, M. Y. A., Vitell, S. J., & Barnes, J. H. (1997). Management of Conflict Using Individual Power Sources: A Retailers' Perspective. Journal of Business Research, 40(1), 49-64.

Reve, T., & Stern, L. W. (1979). Interorganizational Relations in Marketing Channels. Academy of Management Review, 4(3), 405-416.

Robicheaux, R. A., & El-Ansary, A. I. (1976). A General Model for Understand-ing Channel Member Behavior. Journal of Retailing, 52(4), 13-30.

Rohm, A. J. (2001). Channel Strategy Transformation Within and Internet-Enabled Channels Environment: Findings and Implications Across Five Firms. University of Massachusetts Amherst.

Rosenberg, L. J. (1974). A New Approach to Distribution Conflict Management. Business Horizons, 17(October), 67-74.

Rosenberg, L. J., & Stern, L. W. (1971). Conflict Measurement in the Distribution Channel. Journal of Marketing Research, 8(November), 437-442.

______ REFERENCES _____________________________________________________________

Ross, R. H., & Lusch, R. F. (1982). Similarities between Conflict and Coopera-tion in the Marketing Channel. Journal of Business Research, 10(2), 237-250.

Rosson, P. J., & Ford, I. D. (1980). Stake, Conflict, and Performance in Export Marketing Channels. Management International Review, 20(4), 31-37.

Rosson, P. J., & Ford, I. D. (1982). Manufacturer-Overseas Distributor Relations and Export Performance. Journal of International Business Studies, 13(Autumn), 57-72.

Rueckert, R. W., & Orville, C. W. J. (1987). Marketing's Interaction with Other Functional Units: A Conceptual Framework and Empirical Evidence. Jour-nal of Marketing, 51(1), 1-19.

SAS. (2004). The SAS Group's Annual Report 2003 & Sustainability Report. Re-trieved 15 February, 2005, from www.sasgroup.net

SAS. (2005). The SAS Group's Annual Report & Sustainability Report 2004. Re-trieved 30 March, 2005, from http://www.sasgroup.net/SASGROUP_IR/CMSFOREIGNCONTENT/2004eng.pdf

SAS. (2007). SAS Group Annual Report & Sustainability Report 2006. Retrieved 19 June, 2007, from http://www.sasgroup.net/SASGROUP_IR/CMSForeignContent/2006eng.pdf

Schmidt, S. M., & Kochan, T. A. (1972). Conflict: Toward Conceptual Clarity. Administrative Science Quarterly, 17(3), 359-370.

Schul, P. L., Lamb, C. W. J., & Little, T. E. J. (1981). A Path Analysis of the In-trachannel Conflict Process. Paper presented at the The Changing Marketing Environment: New Theories and Applications: 1981 Educators' Confer-ence, Washington D.C., K. Bernhardt, I. Dolich, M. Etzel, W. Kehoe, T. Kinnear, W. J. Perreault & K. Roering (ed.), 39-42. American Marketing Association.

Schul, P. L., Pride, W. M., & Little, T. L. (1983). The Impact of Channel Leader-ship Behavior on Intrachannel Conflict. Journal of Marketing, 47(Summer),21-34.

Seidler, J. (1974). On Using Informants: A Technique for Collecting Quantitative Data and Controlling Measurement Error in Organization Analysis. Ameri-can Sociological Review, 39(December), 816-831.

Shuptrine, F. K., & Foster, J. R. (1976). Monitoring Channel Conflict with Evaluations from the Retail Level. Journal of Retailing, 52(1), 55-74.

Skarmeas, D., Katsikeas, C. S., & Schlegelmilch, B. B. (2002). Drivers of Com-mitment and its Impact on Performance in Cross-Cultural Buyer-Seller Relationshops: The Importer's Perspective. Journal of International Business Studies, 33(4), 757-783.

Smith, M. D., Bailey, J., & Brynjolfsson, E. (1999). Understanding Digital Markets: Review and Assessment. Retrieved October 17, 2004, from

http://www.sasgroup.net

http://www.sasgroup.net/SASGROUP_IR/CMSFOREIGNCONTENT

http://www.sasgroup.net/SASGROUP_IR/CMSForeignContent/2006en

_____________________________________________________________ REFERENCES ______

http://ebusiness.mit.edu/research/papers/140%20erikb,%20digital%20markets.pdf

Steinfield, C., Bouwman, H., & Adelaar, T. (2002). The Dynamics of Click-and-Mortar Electronic Commerce: Opportunities and Management Strategies. International Journal of Electronic Commerce, 7(1), 93-119.

Stern, L. W., & El-Ansary, A. I. (1977). Marketing Channels. Englewood Cliffs, New Jersey: Prentice Hall Inc.

Stern, L. W., & El-Ansary, A. I. (1988). Marketing Channels (3rd ed.). Englewood Cliffs, New Jersey: Prentice Hall.

Stern, L. W., & El-Ansary, A. I. (1992). Marketing Channels (4th ed.). Englewood Cliffs, New Jersey: Prentice-Hall.

Stern, L. W., & Gorman, R. H. (1969). Conflict in Distribution Channels: An Exploration. In L. W. Stern (ed.), Distribution Channels: Behavioral Dimen-sions, 156-175. Boston: Houghton Mifflin Company.

Stern, L. W., & Heskett, J. L. (1969). Conflict Management in Interorganization Relations: A Conceptual Framework. In L. W. Stern (ed.), Distribution Channels: Behavioral Dimensions, 288-305. Boston: Houghton Mifflin Com-pany.

Stern, L. W., Sternthal, B., & Craig, C. S. (1973a). Managing Conflict in Distribu-tion Channels: A Laboratory Study. Journal of Marketing Research, 10(2,May), 169-179.

Stern, L. W., Sternthal, B., & Craig, C. S. (1973b). A Parasimulation of Interor-ganizational Conflict. International Journal of Group Tensions, 3(1-2), 68-90.

Stern, L. W., Sternthal, B., & Craig, C. S. (1975). Strategies for Managing Interor-ganizational Conflict: A Laboratory Paradigm. Journal of Applied Psychology, 60(4), 472-482.

Strauss, A., & Corbin, J. (1998). Basics of Qualitative Research: Techniques and Proce-dures for Developing Grounded Theory (2nd ed.). Thousand Oaks, California: Sage Publications, Inc.

Sullivan, T. J. (2001). Methods of Social Research. Orlando: Harcourt College Pub-lishers.

Tsay, A. A., & Agrawal, N. (2004). Channel Conflict and Coordination in the E-Commerce Age. Production and Operations Management, 13(1), 93-110.

Tuite, M. (1972). Toward a Theory of Joint Decision Making. In M. Tuite, R. Chisholm & M. Radnor (eds.), Interorganizational Decision Making, 9-19. Chicago: Aldine Publishing Company.

Turban, E., King, D., Lee, J., Warkentin, M., & Chung, H. M. (2002). Electroniccommerce: a managerial perspective. New York: Pearson Education.

Urban, G. L. (2004). Digital marketing strategy. Upper Saddle River, New Jersey: Pearson Education, Inc.

Vaile, R. S., Grether, E. T., & Cox, R. (1952). Marketing In the American Economy.New York: The Ronald Press Company.

http://ebusiness.mit.edu/research/papers/140%20erikb,%20digital%20mark

______ REFERENCES _____________________________________________________________

Walker, O. C. J. (1972). The Effects of Learning on Bargaining Behavior. In F. C. Allvine (ed.), Relevance in Marketing and Marketing in Motion Combined Pro-ceedings of the 1971 Spring and Fall Conferences, 194-199. Chicago: American Marketing Association.

Webb, K. L. (1997). Understanding Hybrid Channel Conflict: A Conceptual Model and Four Case Studies. The University of North Carolina, Chapel Hill.

Webb, K. L. (2002). Managing channels of distribution in the age of electronic commerce. Industrial Marketing Management, 31(2), 95-102.

Webb, K. L., & Hogan, J. E. (2002). Hybrid channel conflict: causes and effects on channel performance. Journal of Business & Industrial Marketing, 17(5), 338-356.

Webb, K. L., & Lambe, C. J. (2007). Internal multi-channel conflict: An explora-tory investigation and conceptual framework. Industrial Marketing Manage-ment, 36, 29-43.

Weigand, R. E. (1977). Fit products and channels to your markets. Harvard Busi-ness Review, 55(January-February), 95-105.

Wilkinson, I. (1981). Power, Conflict, and Satisfaction in Distribution Channels - An Empirical Study. International Journal of Distribution and Materials Man-agement, 11(7), 20-30.

Woodside, A. G., & Wilson, E. J. (2003). Case study research methods for theory building. Journal of Business & Industrial Marketing, 18(6/7), 493-508.

Yin, R. K. (2003). Case Study Research: Design and Methods (3rd ed.). Thousand Oaks, California: Sage Publications, Inc.

Zahra, S. A., & Covin, J. G. (1993). Business Strategy, Technology Policy and Firm Performance. Strategic Management Journal, 14(6), 451-478.

Zettelmeyer, F. (2000). Expanding to the Internet: Pricing and Communications Strategies When Firms Compete on Multiple Channels. Journal of Marketing Research, 37(3), 292-308.

______________________________________________________________ APPENDIX 1 ______

1

Appendix 1 Examples of Cause-Related Issues from the Research Instrument used in the Three Channel Dyads (Rosenberg & Stern, 1971, p. 439)

Conflict Cause

Manufacturer-Distributors

Distributors- Dealers

Manufacturer- Dealers

Goals Distributors’ prime goal is to always increase his sales.

Distributor should be willing to sacrifice some of his short-run profits (one-year-basis) by fol-lowing manufacturer’s recommendation for making more profit in the long run.

Distributor does not want to sacrifice his in-dependence and discre-tion over his own busi-ness to the manufac-turer.

Dealer should prefer to tie up as little of his working capital as possible in equipment and parts in-ventory.

Dealer is basically a rug-ged individualist who wants the challenge of running his own business as he sees fit.

Dealer has little desire to establish himself among his customers as an “au-thority” in this product line.

Dealer often sells more Brand X in order to ob-tain incentive trips, al-though he may not make any profit on the extra sales.

Dealer’s ability to install and service Brand X for the customer is more im-portant than manufac-turer’s ability to produce the “best” quality brand in the industry.

It is to dealer’s advantage to add major sidelines to his business, rather than to concentrate fully on the primary product of Brand X.

Domains It is solely manufac-turer’s responsibility to project the product mix for the coming year and to base production schedules upon it.

Because of the high demand for Brand X during the summer, dis-tributor should pre-stock.

When offering funds for a joint promotion cam-paign, manufacturer is justified in specifying the exact ways in which funds may be spent by distributor and his deal-ers.

When dealer returns a de-fective product, he should not incur any cost (e.g., handling charges).

Distributor has the right to set sales quotas on dealer and to hold him ac-countable for them.

When dealer has a serious complaint, he has the right to go over distribu-tor’s head and discuss the matter with manufacturer.

Manufacturer should set and enforce standards for dealer, to be sure the ul-timate consumer is getting a fair value for his pur-chase of Brand X.

Dealers have a right to register opinions to manu-facturer regarding product design features and to get these suggestions incorpo-rated into the product.

Where breakdowns can be traced to manufacturer, dealer should receive labor and materials allowances from manufacturer.

______ APPENDIX 1 ______________________________________________________________

2

Percep-tions

Manufacturer is more willing to try new ideas in selling and sales pro-motion strategy than is distributor.

Manufacturer is inflexi-ble when distributor de-sires action on a special problem.

Manufacturer is rapidly approaching the day when product quality and shortage problems will be eliminated.

The more distributor helps dealer run his busi-ness, the greater dealer’s sales volume.

Distributor is too strict in enforcement of credit terms and discounts in-volving dealer.

Dealer is not promo-tional-minded enough.

Nearly every major manu-facturer’s technological capability to produce a good product is about equal in his industry.

Dealer’s success is more assured if he is tied to Brand X than if tied to any other brand.

Manufacturer is fair and adequate from dealer’s point of view.

Miscell-aneous

Distributor often en-counters difficulties in communicating with manufacturer and other distributors in order to obtain inventory items.

Manufacturer hires away some of distribu-tors’ best personnel.

Manufacturer withholds marketing research data from distributor.

Distributor’s salesmen are typically pleasant when taking orders but less co-operative when com-plaints are made.

Dealer does not imple-ment all the training re-ceived from distributor.

Distributor promises de-livery to meet dealer’s re-quest just to get past cur-rent problems, even at the risk of not being fully sure that distributor can make good his promises.

Dealers are now less loyal toward manufacturer’s brands because the large corporation has displaced the family-owned firm.

When manufacturer changes the specifications of a product, manufacturer is slow to inform dealer about the new product details.

Dealer’s showroom area is very well maintained for demonstrations.

______________________________________________________________ APPENDIX 2 ______

1

Appendix 2 Empirical Studies of Channel Conflict in Marketing Channels

Study Empirical Setting

Sample Size

Research Strategy

Research Ap-proach

Int-er-

Quali-tative

Quan-titative

net

(Assael, 1969) Manufacturer-Dealer(automobiles)US

9 manu-facturer rep-resentatives+ 81 dealers

Case study (open-ended interviews)

X

(Pruden, 1969) Producer-Distributors (wood building materials)US

44 sales calls made by 6 of producer’s salesmen

Survey (per-sonal inter-views)

X

(Rosenberg & Stern, 1971)

Manufacturer-Distributor-Dealers (rela-tively expen-sive household durables) US

4 + 8 manu-facturer rep-resentatives+ 11 distri-butors + 87 dealers

Survey(mail ques-tionnaire & personal in-terviews)

X

(Walker, 1972) Simulated Manufacturer-RetailerChannel(students) US

72 students (36 manu-facturer-retailergroups)

Experiment X

(Pearson,1973)

Vendors-Retailers(Groceries)US

55 respon-dents (from both vendor and retailer side of channel)

Survey(question-naire)

X

(Stern, Stern-thal & Craig, 1973a; Stern et al., 1973b; Stern, Stern-thal & Craig, 1975)

SimulatedManufacturer-WholesalerChannel(students) US

62student groups

Experiment X

(Foster & Shuptrine,1974)

Manufacturer-Wholesaler-RetailerChannel(specialty con-sumer product) US

167 retailers + 4 whole-salers

Survey(interviews)

X

______ APPENDIX 2 ______________________________________________________________

2


Sample Size

Research Strategy

Research Ap-proach

Int-er-

Quali-tative

Quan-titative

net

(Rosenberg, 1974)

Manufacturer-Reseller (air conditioning,cosmetics, fab-rics, groceries gasoline, paper, sewing ma-chines, soaps, etc)USMultiple Marketing Channels

8 manu-facturerexecutives

Personal in-terviews (in-terviewguide, open-ended ques-tions)

X

(Hunger & Stern, 1976)

SimulatedManufacturer-Distributor Channel(students) US

108 male under-graduatestudents (36 manu-facturer-distributor groups)

Experiment X

(Lusch, 1976b, 1976a)

Manufacturer-Dealers(automobiles)US

567 dealers Survey (mail ques-tionnaire)

X

(Pearson & Monoky,1976)

Wholesalers-Retailers(groceries)US

54 dyads (re-tail grocery byers + their grocery ven-dors

Survey(question-naire)

X

(Shuptrine & Foster, 1976)

Manufacturer-Retailer chan-nel (specialty consumerproduct sold through single-line retail stores)

157 retailers Survey (interviews)

X

(Brown, 1977) Simulated Manufacturer-Retailer chan-nel(students) US

47 under-graduatestudents

Experiment X

______________________________________________________________ APPENDIX 2 ______

3


Sample Size

Research Strategy

Research Ap-proach

Int-er-

Quali-tative

Quan-titative

net

(Kelly & Pe-ters, 1977)

Franchisee-Distributor (fast food, automotiveservice, hotel, industrial goods, con-sumer goods) US

27 franchi-sees + 12 distributors

Survey(personal in-terviews)

X

(Brown & Fra-zier, 1978)

Manufacturer-Dealers (auto-mobiles) US

26 dealers Survey (personal in-terviews)

X

(Brown, 1979) Manufacturer-Dealers (auto-mobiles) US


X

(Etgar, 1979) Manufacturer-Dealers (furni-ture, automo-biles, liquor, shoes, insur-ance, stereo) US


X

(Dwyer, 1980) Simulated Bilateral Du-opoly(students) US

80 students (20 quads; 2 manu-facturers +2 resellers/ quad)

Experiment X

(Rosson & Ford, 1980)

Manufacturer-Distributor (electronics,industrial leathers, etc.) Canada

19 manu-facturers in Canada ex-porting to the UK


X

(Brown & Day, 1981)

Manufacturer-Dealer(automobiles)US

210 dealers Survey (question-naire)

X

(Schul et al., 1981)

Franchisor-Franchisee(real estate brokers) US

349 franchi-sees

Survey(mail ques-tionnaire)

X

______ APPENDIX 2 ______________________________________________________________

4


Sample Size

Research Strategy

Research Ap-proach

Int-er-

Quali-tative

Quan-titative

net

(Wilkinson,1981)

Manufacturer-Retailer (75 brewer-hotel retailer dyads) Sydney

60 hotels (i.e. some hotels dealt with two breweries)


X

(Ross & Lusch, 1982)

Brokers-Wholesalers(groceries)US

54 matched pairs of bro-kers-wholesalers


X

(Rosson & Ford, 1982)

Manufacturers-Distributors (Canadianmanufacturers-UK distri-butors)

21 manu-facturer-distributor dyads


X

(Brown et al., 1983)

Suppliers-Retailers (beer, fast food, shoes, tires, automobiles, lawn movers) US

139 retailers Survey (personal in-terviews)

X

(Schul, Pride & Little, 1983)

Franchisor-Franchisee(real estate brokers) US

349 franchi-sees


X

(Anderson & Narus, 1984)

Manufacturers-Distributors (electroniccomponents)

153 distribu-tors


X

(Eliashberg & Michie, 1984)

Franchisor-Franchisee(industrial in-stallationsmanufacturer)

80 franchi-sor-franchiseedyads


X

(Dilts, 1985) Manufacturers-Retailers(audio and video home electronics) US Multiple Marketing Channels

78 retailers Survey (mail ques-tionnaire)

X

______________________________________________________________ APPENDIX 2 ______

5


Sample Size

Research Strategy

Research Ap-proach

Int-er-

Quali-tative

Quan-titative

net

(Gaski & Nevin, 1985)

Supplier/DealerChannel(heavy indus-trial machin-ery)

43 suppliers + 238 deal-ers(same manu-facturer)

Survey(mailquestion-naire)

X

(Arndt & Ogaard, 1986)

Manufac-turer/Importer-Dealer Chan-nel (automo-biles) US & Norway

55 US dyads + 104 Nor-wegian dy-ads

Survey(interviews)

X

(Achrol & Stern, 1988)

Supplier-Retailers (air conditioning,automobiles, books, agricul-tural equip-ment, com-puters, office supplies, TV/electronics etc)


X

(Cunningham& Pyatt, 1989)

Manufacturers-Distributors- End-users(mid-rangecomputer sys-tems)UK

3 manu-facturers+ 6 distribu-tors + 10 end-user or-gani-sations

Survey(personal in-terviews + mail ques-tionnaires)

X X

(Frazier et al., 1989)

Manufacturers-Dealers (tung-sten carbide tool industry) India


X

(Anderson & Narus, 1990)

Manufacturer- Wholesalers/Distributors (more than 100 industries)

213 manu-facturerfirms + 249 distributor firms


X

(Moore, 1990) Exporters-Distributors (UK manufac-turers, West German im-porters)

81 manu-facturers + 57 agents and distribu-tors


X

______ APPENDIX 2 ______________________________________________________________

6


Sample Size

Research Strategy

Research Ap-proach

Int-er-

Quali-tative

Quan-titative

net

(Ong et al., 1990)

Franchisor-Franchisee(petroleum) Perth,Australia

61 + 36 franchisees


X

(Brown et al., 1991)

Manufacturer-Distributor (light aircraft engine parts) North America

32 distribu-tors

LongitudinalSurvey(2 x identical mail ques-tionnaires)

X

(Frazier & Rody, 1991)

Suppliers-Distributors (industrial dis-tributors with sales revenues between US$ 1 to US$ 10 million)US

300 distribu-tors


X

(Katsikeas & Piercy, 1991)

Exporters-Importers (Greek export-ers, UK im-porters)

28 Exporter-Importer Dyads


X

(Ketilson,1991)

Wholesalers-Retailers (re-tailer-owned cooperative wholesaling-retailing sys-tem)Canada


X

(Brown & Fern, 1992)

SimulatedManufacturers-Wholesalers-Retailers Mul-tiple Market-ing Channels

22 student groups(4 manu-facturers+ 6 whole-salers + 12 retailers)

Experiment X

(Dant & Schul, 1992)

Franchisor-Franchisee(fast food res-taurants) US

176 franchi-sees


X

______________________________________________________________ APPENDIX 2 ______

7


Sample Size

Research Strategy

Research Ap-proach

Int-er-

Quali-tative

Quan-titative

net

(Katsikeas, 1992)

Buyers-Sellers(Greek manu-facturers trad-ing with UK importers) Greece

53 exporters (+ domestic buyer-sellerrelationshipsof 45 of these ex-porters)


X

(Cronin & Baker, 1993)

Manufacturer-Dealers(fluid power products) US


X

(Ganesan,1993)

Vendor-Retailers(5 department store chains)


X

(Nicholls,Roslow & Laskey, 1993)

Manufacturers-Retailers(boat industry) US

253 retailers Survey (personal in-terviews)

X

(Price, 1993) Manufacturer-RetailersManufacturer-Franchisees(furniture)Multiple Marketing Channels

1026 retail-ers+ 74 fran-chisees(same manu-facturer)


X

(Gundlach & Cadotte, 1994)

SimulatedManufacturer-Distributor (simulated mi-cro-computerindustry)

88 student groups(44 manu-facturers+ 44 dis-tributors)

Experiment(question-naire)

X

(Kumar et al., 1995)

Supplier-Dealer(automobiles)US


X

(Rawwas et al., 1997)

Wholesaler-Retailer(independent pharmacists)


X

______ APPENDIX 2 ______________________________________________________________

8


Sample Size

Research Strategy

Research Ap-proach

Int-er-

Quali-tative

Quan-titative

net

(Webb, 1997) Manufacturer/ Service Pro-vider-Retailer (computers, electricalequipments, hosieries, pag-ing services) USMultiple Marketing Channels

3 manu-facturers+ 1 service provider

Case Studies (58 personal open-ended interviews + 62 mail question-naires)

X X X

(Lee, 2001) Supplier-Distributors (brewery) China

95 distri-butors


X

(Rohm, 2001) Manufacturer-Retailer (athletic foot-wear, automo-tive, financial services, sta-tionery, print-ing equipment) USMultiple Marketing Channels

5 Manu-facturers

Case Studies (personal in-terviews)

X X

(Webb & Ho-gan, 2002)

Manufacturer/ Service Pro-vider-Retailer (computers, electricalequipments, hosieries, pag-ing services) USMultiple Marketing Channels

3 manu-facturers+ 1 service provider

Case Studies (62 mail question-naires)

X X

______________________________________________________________ APPENDIX 2 ______

9


Sample Size

Research Strategy

Research Ap-proach

Int-er-

Quali-tative

Quan-titative

net

(Duarte & Da-vies, 2003)

Service pro-viders-Resellers(financial ser-vices)UK

496 dyads (area manag-ers + agents)

Survey(mail ques-tionnaires)

X

(Jelassi & Leenen, 2003)

Manufacturer-Reseller (motorcycle manufacturer) ItalyMultiple Marketing Channels

1 manu-facturer

Case Study (personal open-ended interviews + telephoneinterviews)

X X

(Li, Mitra & Matlay, 2004)

Producer-Reseller (elec-tronic equip-ment)UKMultiple Marketing Channels

1 producer (or possibly 2)

Case Studies (Surveyof 10 Small- and Me-dium-sizedEnterprises + 10 per-sonal open-ended inter-views)

X (X)* X

(Goldkuhl,2005)**

Producer-Reseller(motorcycles,air travel) Italy, Sweden Multiple Marketing Channels

1 service provider + 1 manu-facturer

Case Studies (personal open-ended interviews)

X X

(Leonidou et al., 2006)

Exporter-Importer (consumer or industrial type products) US

201 export-ers


X

* The study includes some extent of quantitative data analysis, i.e. a questionnaire containing nine questions about current distribution channel structure, reasons for use of electronic channel, percep-tion of channel conflict and practices of channel management was distributed.

** The major part of this licentiate study is also included as Chapter 2 of this doctoral thesis.

______ APPENDIX 2 ______________________________________________________________

10


Sample Size

Research Strategy

Research Ap-proach

Int-er-

Quali-tative

Quan-titative

net

(Webb & Lambe, 2007)

Producer-Reseller (wire-less messaging, computer serv-ers, electrical equipment, hosieries)Multiple Marketing Channels

1 service provider + 3 manu-facturers

Case Studies (personal open-ended interviews)

X X

______________________________________________________________ APPENDIX 3 ______

1

Appendix 3 Interview Guide

INTERVIEW GUIDE

Name: _________________________________ Date: _________________

Company: ______________________________ Position: ______________

1) Describe your position at the company. How long have you been in this po-sition? What are your job responsibilities and objectives? What prior posi-tion(s) did you have?

2) How would you describe the business of your company and the markets which it serves?

3) What distribution channels are used to serve these markets. How and when did they evolve? (Internet)

4) Is your overall channel strategy viewed as a source of competitive advantage, or potentially so? If so, how? Does your distribution channel strategy differ in any important way compared to your competitors’ channel strategy?

5) When you decided to start selling on the Internet, what were the main con-cerns which you had?

6) What measures did you take to overcome or minimise these concerns?

7) What risks did you think of when you considered bypassing resellers?

8) When looking back, what would you have done differently (when adding the Internet)?

9) What do you think are the causes of conflict between your distribution channels?

10) Which cause is the most important?

______ APPENDIX 3 ______________________________________________________________

2

11) Has this changed in any way since the Internet was added as a distribution channel?

12) Has the Internet created more conflict among your channels? If so, in what way?

13) Has the Internet reduced channel conflict? If so, in what way?

14) How do you establish the goals and objectives for each channel?

15) Are there synergies between the different channels (products, pricing, pro-motion)?

16) To what extent do your channels perform the same task or activity?

17) To what extent do your various channels interact with the same clients or customers?

18) Do your channels’ managers argue about customers to serve? If so, how of-ten?

19) To what extent do your different channels market identical products and/or services?

20) Are your channels sometimes arguing with each other because of different views on some issues? If so, why?

21) How do you communicate your distribution strategy to your resellers?

22) Do you view communication as important in order to avoid channel con-flict? If so, in what way?

23) To what extent can you decide the marketing strategy of your resellers? (store, e-tailer etc)?

24) What issues do your channels argue about? (Customers served/Products or Services offered/Prices/Allocation of Resources/Compensation etc.)

25) How often do they argue about these issue(s)?

26) Do you think that these issues sometimes are dangerous to the company? If so, why? If not, why not?

______________________________________________________________ APPENDIX 3 ______

3

27) How important are these issues to the company’s overall performance, and ultimately, its profitability?

28) Among those issues you mentioned, which conflict issue is/was the most im-portant?

29) When do you think a channel conflict can become dangerous for the organi-sation?

30) How can you find out about a channel conflict within your distribution sys-tem? (Methods to measure/Customer complaints/Personnel com-plaints/Reseller complaints)

31) Is there any structured manner that helps identify and measure channel con-flicts?

32) Do you today have the same approach to Internet sales, compared to when you first introduced the Internet channel? (products, prices, brand name)

33) Was the channel conflict that occurred when adding the Internet channel positive to your company? If so, how?

34) Was the channel conflict that occurred when adding the Internet channel dangerous to your company? If so, how?

35) How did channel members respond to the channel conflict?

36) How did customers respond to the channel conflict?

37) How do you resolve disputes between your channels?(Guidelines, exchange-of-personnel, negotiations, etc./Price levels/Product versions/Brand name/Compensation/Relational norms)

38) What price levels do you utilize in your different channels? (Higher or lower online prices, identical prices in all marketing channels, auction pricing, etc.).

39) Do your online prices include shipping costs?

40) Do you view pricing as a source of channel conflict? If so, compared to other conflict sources, is it a bigger or smaller source of conflict?

41) Do the chosen price levels aim at avoiding channel conflict?

______ APPENDIX 3 ______________________________________________________________

4

42) Has the pricing method changed from before adding the Internet until to-day?

43) Can the products which are sold online be purchased through other distribu-tion channels? If not, how different are the products offered through various channels? (minor/major changes)

44) Do you think it is possible to sell the same product under different names and through different channels in order to charge different prices? What would customers’ reactions be?

45) Same scenario, but different brand names?

46) Does your company compensate any of the other distribution channels for online sales? If so, how?

47) Overall, what impact, if any, have the Internet and e-commerce had on your overall channel strategy and channel structure?

48) What impact has the characteristics of the Internet had on channel conflict, channel strategy and channel structure?

49) How satisfied are you with your current channel strategy? What would you like to change in this strategy? How do you see it in the future?

50) Is there anything else that you would like to add? Are there other people at your company that I should talk to?

______________________________________________________________ APPENDIX 4 ______

1

Appendix 4 Questionnaire

QUESTIONNAIRE

This survey focuses on your company’s relationship with SAS, and specifically with SAS’s online sales; it should take no more than 10 minutes to complete. Please indicate the extent to which you agree or disagree with each statement. So for example, if you strongly agree with a statement you might circle 7. If you strongly disagree you might circle 1. If you neither strongly agree nor strongly dis-agree, circle one of the numbers between 2 and 6 to properly reflect the actual strength of your feeling. Remember, there are no right or wrong answers - we are interested in what you feel.

1.a) How long has your firm been doing business with SAS? ________ years

b) How long have you been working in the travel agency industry? ________ years

c) How many people work at your travel agency (office)? ______________________________

SASSALESINFO.COM

Strongly Strongly Disagree Agree

2.a) Because of sassalesinfo.com, we utilise more information from SAS than we did before the 1 2 3 4 5 6 7

Extranet was implemented b) We get access to such information from SAS that we would not have 1 2 3 4 5 6 7

had access to before sassalesinfo.com was implemented

c) Because of sassalesinfo.com, the 1 2 3 4 5 6 7 information that we get from SAS is more accurate than before it was implemented

d) Sassalesinfo.com has contributed to decrease the level of disagreements 1 2 3 4 5 6 7

with SAS

e) The information that is available through sassalesinfo.com is structured 1 2 3 4 5 6 7 in such a way that it is easy to find

f) Too much information is available through sassalesinfo.com, which makes 1 2 3 4 5 6 7

it hard to find the needed information

g) Important information is lacking on sassalesinfo.com 1 2 3 4 5 6 7

h) If important information is lacking on sassalesinfo.com, WHAT is lacking?

___________________________________________________________________________

______ APPENDIX 4 ______________________________________________________________

2

COMMUNICATION

Strongly Strongly Disagree Agree 3.a) SAS has a good understanding of the type of information that is 1 2 3 4 5 6 7 needed by our company

b) SAS has a good understanding of the frequency of information that 1 2 3 4 5 6 7 is needed by our company

c) SAS keeps us well informed about events or changes that may affect us 1 2 3 4 5 6 7

d) SAS keeps us well informed about their products 1 2 3 4 5 6 7

e) SAS keeps us well informed about what is going on in their company 1 2 3 4 5 6 7

f) We keep SAS well informed about events or changes that may affect them 1 2 3 4 5 6 7

g) We hesitate to give SAS too much information 1 2 3 4 5 6 7

h) We regularly provide SAS with feedback about their products 1 2 3 4 5 6 7

i) We regularly provide SAS with feedback about our local market conditions 1 2 3 4 5 6 7

j) Over a typical four-week period, please estimate the frequency with which SAS provides

information to you via:Very Very

Infrequently Frequently

* Face-to-face interaction 1 2 3 4 5 6 7

* Telephone 1 2 3 4 5 6 7

* e-mail 1 2 3 4 5 6 7

* sassalesinfo.com 1 2 3 4 5 6 7

* Other, please specify

_____________________________ 1 2 3 4 5 6 7

______________________________________________________________ APPENDIX 4 ______

3

k) Over a typical four-week period, please estimate the frequency with which you provide

information to SAS via:Very Very

Infrequently Frequently

* Face-to-face interaction 1 2 3 4 5 6 7

* Telephone 1 2 3 4 5 6 7

* e-mail 1 2 3 4 5 6 7

* Amadeus, etc. 1 2 3 4 5 6 7


_____________________________ 1 2 3 4 5 6 7

DOMAINS


4.a) SAS has given us an exclusive territory for their products 1 2 3 4 5 6 7

b) The duties of SAS and the duties of our company are well-defined; only rarely is it not known who 1 2 3 4 5 6 7 should perform a specific function

c) Sometimes SAS tries to take over activities that we feel are our 1 2 3 4 5 6 7 responsibility

d) Via its direct-to-the-customer Internet sales, SAS serves the same 1 2 3 4 5 6 7 customers as we do

e) Via its direct-to-the-customer Internet sales, SAS markets 1 2 3 4 5 6 7 identical products as we do

f) We compete with SAS’s 1 2 3 4 5 6 7 direct-to-the-customer Internet sales

GOALS


5.a) We perceive that the goals of SAS are compatible with the goals of our company 1 2 3 4 5 6 7

b) We perceive that SAS and we have the same view on how to compete within our 1 2 3 4 5 6 7 company’s local market area

______ APPENDIX 4 ______________________________________________________________

4

Strongly Strongly Disagree Agreec) We perceive that it is feasible for both SAS and for us to meet or exceed our 1 2 3 4 5 6 7 goals

d) We disagree with SAS about how we can best achieve our respective goals 1 2 3 4 5 6 7

e) Success on the behalf of SAS’s direct-to-the-customer Internet sales 1 2 3 4 5 6 7 channel comes at the expense of our sales

RELATIONSHIP


6.a) It would be difficult for our firm to replace SAS with another supplier 1 2 3 4 5 6 7

b) There are other suppliers who could provide us with comparable product 1 2 3 4 5 6 7

lines

c) Maintaining SAS is critical to our company’s profitability 1 2 3 4 5 6 7

d) It would be difficult for our firm to replace the sales and profits 1 2 3 4 5 6 7 generated from SAS’s products

e) In our sales area, there are other travel agencies that could provide SAS 1 2 3 4 5 6 7

with comparable distribution

f) In our sales area, our travel agency generates a considerable share of SAS’ 1 2 3 4 5 6 7

sales and profits

g) In our sales area, it would be difficult for SAS to replace the tasks that 1 2 3 4 5 6 7 our firm performs

h) We expect our relationship with 1 2 3 4 5 6 7 SAS to continue for a long time

i) If another company offered us a better product line, we would most 1 2 3 4 5 6 7 certainly take them on, even if it

meant dropping SASj) If SAS requested it, we would be willing to make further investment 1 2 3 4 5 6 7

in supporting SAS’s product line

______________________________________________________________ APPENDIX 4 ______

5


k) Our relationship with SAS is a long-term cooperation 1 2 3 4 5 6 7

l) We have invested a lot in our relationship with SAS 1 2 3 4 5 6 7

m) When SAS added its Internet sales channel, we came close to terminating 1 2 3 4 5 6 7

our relationship with SAS

n) The relationship with SAS is 1 2 3 4 5 6 7 marked by a high degree of harmony

o) We argue frequently with SAS about 1 2 3 4 5 6 7 business issues

p) Our arguments with SAS are very 1 2 3 4 5 6 7 intense

q) How often do you perceive that disagreements occur with SAS on:

Very Very Infrequently Frequently

* Information exchanges 1 2 3 4 5 6 7

* Product aspects 1 2 3 4 5 6 7

* Pricing issues 1 2 3 4 5 6 7

* Payment terms 1 2 3 4 5 6 7

* Commission on sales 1 2 3 4 5 6 7

* SAS’s sales organisation 1 2 3 4 5 6 7

* Advertising and other promotional 1 2 3 4 5 6 7 activities

* SAS’s direct-to-the-customer 1 2 3 4 5 6 7 Internet sales today

* SAS’s direct-to-the-customer Internet sales when SAS introduced 1 2 3 4 5 6 7 that sales channel

* Our company’s sales effort 1 2 3 4 5 6 7


____________________________ 1 2 3 4 5 6 7

______ APPENDIX 4 ______________________________________________________________

6

r) How intense do you perceive that the disagreements with SAS have been on:

Very Very Weak Strong

* Information exchanges 1 2 3 4 5 6 7

* Product aspects 1 2 3 4 5 6 7

* Pricing issues 1 2 3 4 5 6 7

* Payment terms 1 2 3 4 5 6 7

* Commission on sales 1 2 3 4 5 6 7

* SAS’s sales organisation 1 2 3 4 5 6 7

* Advertising and other promotional 1 2 3 4 5 6 7 activities

* SAS’s direct-to-the-customer 1 2 3 4 5 6 7 Internet sales today

* SAS’s direct-to-the-customer 1 2 3 4 5 6 7 Internet sales when SAS introduced that sales channel

* Our company’s sales effort 1 2 3 4 5 6 7


____________________________ 1 2 3 4 5 6 7

FUNCTIONALITYIn every relationship disagreements sometimes occur. Regardless if you have had many or few disagreements with SAS, this part of the survey aims to capture your perception on which results your disagreements have led to.


7.a) Disagreements between SAS and our firm have considerably increased the productivity of our 1 2 3 4 5 6 7 working relationship

b) Disagreements with SAS stimulate us to find productive 1 2 3 4 5 6 7 solutions to our problems

c) Disagreements between SAS and our firm have damaged our working 1 2 3 4 5 6 7 relationship

______________________________________________________________ APPENDIX 4 ______

7


d) Disagreements between SAS and our firm have negatively affected 1 2 3 4 5 6 7 the SAS brand

e) We are very satisfied with how the disagreements that we have had 1 2 3 4 5 6 7 with SAS have been resolved f) With regards to SAS’s direct-to-the-customer Internet sales, 1 2 3 4 5 6 7 we are very satisfied with how the disagreements that we have had with SAS have been resolved

OUTCOMES


8.a) The relationship between our firm and SAS has been very productive 1 2 3 4 5 6 7

b) We have found the time and effort spent on the relationship with SAS 1 2 3 4 5 6 7 very worthwhile

c) The relationship between our firm and SAS has been very effective 1 2 3 4 5 6 7

d) We have a very rewarding relationship 1 2 3 4 5 6 7 with SAS

e) Overall, the return on investment from selling SAS has exceeded our firm’s 1 2 3 4 5 6 7 expectations

f) The growth in sales over time, from selling SAS, has exceeded our firm’s 1 2 3 4 5 6 7 expectations

g) The growth in profit over time, from selling SAS, has exceeded our firm’s 1 2 3 4 5 6 7 expectations

h) Overall, the financial results from selling SAS have exceeded our firms 1 2 3 4 5 6 7 expectations

i) When SAS added the direct-to-the- customer Internet sales of motorcycles, 1 2 3 4 5 6 7 it received a lot of media attention j) The media coverage was overall very positive when SAS added the direct-to-the-customer Internet sales 1 2 3 4 5 6 7 of motorcycles k) SAS’s direct-to-the-customer Internet sales of motorcycles has considerably 1 2 3 4 5 6 7 contributed to build the SAS brand

______ APPENDIX 4 ______________________________________________________________

8


l) Overall, SAS’s direct-to-the-customer Internet sales of motorcycles has had a 1 2 3 4 5 6 7 positive effect on our profits

m) Overall, SAS’s direct-to-the-customer Internet sales of motorcycles has had a 1 2 3 4 5 6 7 positive effect on our sales

n) Overall, SAS’s direct-to-the-customer Internet sales of motorcycles has had a 1 2 3 4 5 6 7 negative effect on our profits

o) We are very satisfied with the arrangement of our overall channel 1 2 3 4 5 6 7 system*

p) We are very satisfied with SAS’s decision to sell products direct-to-the- 1 2 3 4 5 6 7 customer via the Internet

q) We do not think that the current channel system* needs to be changed 1 2 3 4 5 6 7

r) Given the prerequisites, we feel that the current channel system* 1 2 3 4 5 6 7 is the best possible

* Channel system refers to the WHOLE sales- and distribution system, i.e. the system between your company and SAS, between your company and the end customer, and between SAS and the end customer

Note! Because of different paper size, the original questionnaire was exactly 3 dou-ble-pages long.

______________________________________________________________ APPENDIX 5 ______

Industrial Marketing & e-Commerce Luleå University of Technology

POSTAL ADDRESS: SE-971 87 Luleå VISITING ADDRESS: Universitetsområdet, Porsön, Luleå

TELEPHONE: +46 (0)920 49 10 00 HOMEPAGE: www.ltu.se

Appendix 5 Sample Cover Letter

April 2007

To: President/CEO Regarding: Doctoral Research

Dear Sir/Madam,

My name is Lena Goldkuhl and I am a PhD Candidate at Luleå University of Technology in Sweden. My research focuses on business relations between producers and their resel-lers.

In a previous study, managers at Ducati Motor Holding SpA (Ducati) gave their views on how their direct-to-the-customer Internet sales channel has affected the relationship with resellers. Based on that study, this questionnaire aims at capturing your (the reseller’s) per-spective on this important issue. The questionnaire is distributed to resellers in Canada, Great Britain, Ireland, Italy, Norway, Sweden and the USA. This study, together with the previous study, will serve as the foundation of my PhD thesis.

Please respond to the questionnaire based on your past interactions with Ducati, and send it back in the postage-paid envelope as soon as possible. The questionnaire takes around 10 minutes to complete.

This questionnaire is COMPLETELY ANONYMOUS. When I have received your survey I will note that you have responded, and thereafter all identification information will be destroyed. Only aggregated data will be published in the report, and therefore it will be impossible to identify specific individuals or organisations in the report. In other words, Ducati will not know how you responded to this questionnaire, so take your chance to influence them!

Your participation in this research is of course voluntarily, but the research stands and falls with the number of respondents. So besides the fact that your answers are very important to Ducati, they are also of great importance to my research.

Thank you in advance for your participation in this research!

Lena Goldkuhl

Industrial Marketing & e-Commerce Luleå University of Technology e-mail: [email protected]

http://www.ltu.se

mailto:[email protected]