Journal of Retailing and Consumer Services 24 (2015) 22–32

Contents lists available at ScienceDirect

Journal of Retailing and Consumer Services

http://d0969-69

n CorrFrance.

E-m

journal homepage: www.elsevier.com/locate/jretconser

Effects of loyalty program rewards on store loyalty

Lars Meyer-Waarden n

University Toulouse 1 Capitole, CRM CNRS, IAE School of Management & EM Strasbourg Business School-HuManiS (EA 7308), France

a r t i c l e i n f o

Article history:Received 19 December 2013Received in revised form31 December 2014Accepted 15 January 2015Available online 4 February 2015

Keywords:Loyalty programReward timingReward tangibilityReward compatibilityLoyaltyConjoint analysis

x.doi.org/10.1016/j.jretconser.2015.01.00189/& 2015 Elsevier Ltd. All rights reserved.

espondence address: 2, rue du Doyen-GabriFax: þ33 5 61 63 56.ail address: meyerwaarden@gmail.com

a b s t r a c t

This investigation examines consumers' preferred loyalty program (LP) designs across two retail con-texts, grocery retailing and perfumery, with varying degrees of personal involvement. The researchemploys in-store full profile conjoint analysis by using the following attributes: timing of the reward,reward compatibility with the store's image, and tangibility.

Our research reveals that the underlying effects of reward types on preferences and intended storeloyalty differ depending on the level of consumers' personal involvement. In sectors with high personalinvolvement, compatibility with the store's image and intangible rewards increase LP preference andloyalty intentions. The time required to obtain the reward (immediate/delayed) has no impact. In sectorswith low personal involvement, immediate and tangible rewards increase LP preference and loyaltyintentions. Compatibility with the store image has no impact.

& 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Many firms use customer relationship management instru-ments, such as loyalty programs (LPs), as key marketing activitiesfor customer information collection. LPs are vastly popular—forexample, 90% of Europeans and 90% of U.S. shoppers own at leastone loyalty card (Ferguson and Hlavinka, 2009). In 2010, thenumber of LP memberships in the United States exceeded 2.1 bil-lion memberships, growing by 16% from the previous year despitethe worldwide recession (Hlavinka and Sullivan, 2011). For ex-ample, research estimates that the U.K. pharmacy chain Boots in-vested 30 million British pounds in the launch of its AdvantageCard LP (Temporal and Trott, 2001), and the U.K. retailer Tesco hasspent an estimated 60 million pounds to operate its Clubcard LP(Bijmolt et al., 2010).

Despite their popularity, existing research challenges the effi-cacy of LPs because, in many cases, they offer rewards that fail toincrease loyalty (Leenheer et al., 2007; Liu, 2007; Meyer-Waarden,2007; Meyer-Waarden and Benavent, 2009). Fewer than half of LPmembers report that the programs add value, and the impact ofLPs on customer patronage lags behind most companies' ex-pectations (Ferguson and Hlavinka, 2009). Yet rewards should of-fer value (Bridson et al., 2008; García Gómez et al., 2012; Roehmet al., 2002). Overall, increasing the benefits and decreasing thecosts of using LPs increase loyalty (Demoulin and Zidda, 2008).

el-Marty , F-31000 Toulouse,

Nevertheless, heterogeneity in responsiveness exists across cus-tomer segments and industry sectors, as effectiveness depends onmarket characteristics. Therefore, this study contributes to re-search by assessing the following questions that remain in-sufficiently investigated (Bijmolt et al., 2010, pp. 207, 239):(1) Which type of reward creates customer value and enhances LPmembers' patronage intentions? and (2) What is the moderatingrole of personal involvement on LP effectiveness?

To contribute to a better theoretical and empirical under-standing of the effects of rewards, we propose a conceptual fra-mework that examines how rewards affect LP preferences ac-cording to three key variables that have insufficiently or not beenempirically investigated (Blattberg et al., 2008). Each of the vari-ables have been studied before, but no one has had all of themwithin the same study: (1) reward (in)tangibility, (2) compatibilitywith the image of the firm that offers the LP, and (3) time neces-sary to obtain rewards timing (immediate vs. delayed). We pro-pose that their relative impact on LP preference and loyalty in-tentions varies depending on consumers' personal involvement inthe product category (Yi and Jeon, 2003). This study thus examineshow different aspects of rewards affect preferences of a LP: type ofrewards (tangibility and compatibility) and timing of rewards. Weconduct conjoint analysis in a French grocery retailer and a per-fumery store. After presenting the results, we conclude with adiscussion, managerial implications, and avenues for furtherresearch.

Fig. 1. Conceptual framework.

L. Meyer-Waarden / Journal of Retailing and Consumer Services 24 (2015) 22–32 23

2. Theoretical background

2.1. Loyalty programs and loyalty

The American Marketing Association refers to loyalty programsas ‘continuity incentive programs offered by a firm to rewardcustomers and encourage repeat business’. Prior literature usesdifferent terms, including reward programs, frequency rewardprograms, frequent-shoppers programs, and frequent-flierprograms.

We adopt the term loyalty program (LP) to encompass all theseterms and various program designs that contain the followingcharacteristics (Blattberg et al., 2008; Leenheer et al., 2007): LPsconsist of integrated, structured and ruled (based on collection andredemption rules) systems of marketing actions that aim to en-courage enduring repeat purchases and increase the cost ofswitching by providing short- and long-term incentives (Meyer-Waarden, 2008). These rewards refer to any abstract (e.g., con-venience, hedonic, novelty, social recognition, self-esteem) orconcrete (e.g., economic savings, miles, points, discounts) stimulithat trigger consumers' internal cognitive responses (Vesel andZabkar, 2009; Drèze and Nunes, 2011; Kwong et al., 2011; Tietje,2002). Usually, LP members are rewarded with discounts, goods,services, personalized offers and tailored marketing efforts, orpreferential treatment (Meyer-Waarden, 2013). To induce sus-tainable effects on members' loyalty, LP participation and rewardsshould increase cost of switching (Bijmolt et al., 2010) and en-hance “true loyalty”, that is increase behavioral (e.g., cross pur-chases, repeat purchases, mean basket size; Ehrenberg, 1988) andattitudinal (relationship building through positives attitudes, trust,attachment; Morgan and Hunt, 1994) loyalty. A LP rewardsmembers for their loyalty on the basis of their past, current orfuture value to the firm, which is usually done through the accu-mulation of some form of LP currency based on purchase behavior.Rewards thus help create perceived value and satisfaction; im-prove economic decision-making and motivation.

2.2. Loyalty program perceived value and preference

Perceived value created is the relationship between the con-sumer's perceived benefits in relation to the perceived costs ofreceiving a good or a service, and represents a positive emotionalresponse (e.g. such as subjective feelings of pleasure or hedonicenjoyment) as well as a source of satisfaction and motivation,because the rewards fulfill a desire or a goal (Holbrook, 1996;Bagchi and Li, 2011).

A buyer who commits to an LP considers the costs and effortsrequired (e.g., membership fees, provision of personal informationto the firm, switching costs, transportation, changes to purchasebehavior) and compare these costs with the reward value. Theinfluence of LPs on member loyalty is contingent on the LP design(Keh and Lee, 2006). Specifically, LP design affects enrollment andloyalty. If the reward values are higher than the costs, the con-sumer decides to join the LP and change or increase behavioraland attitudinal loyalty. The nature and preference of rewards arethus decisive with regard to consumers' motivation to adopt anduse the LP as well as to change behavior and attitudes.

In this respect, two key analyses variables are recommended(Blattberg et al., 2008; O’Brien and Jones, 1995): (1) reward types,including tangibility and compatibility with the firm's image, and(2) reward timing. Yi and Jeon (2003) also suggest considering themoderating role of personal involvement. Each of the variables hasbeen studied before, but no one has had all of them within thesame study. We thus propose a conceptual model with threestages (see Fig. 1). The first and second stages examine how re-ward schemes (e.g. timing tangibility, compatibility of rewards)

affect first customers' rewards preferences of the LP and secondloyalty intentions. The third stage investigates the moderating roleof personal involvement.

3. Conceptual framework and hypotheses

3.1. Conceptual framework

Customers' value perception is a necessary condition for de-veloping brand loyalty through LPs (O’Brien and Jones, 1995). Re-tailers attempt to meet a wide range of requirements by offeringan array of rewards that cover different dimensions to increase thepreference of an LP (Meyer-Waarden, 2013). Both Roehm et al.(2002) and Dowling and Uncles (1997) suggest that three elementsof the LP determine its preference and thus can strengthen orweaken positive associations with the brand, resulting in in-creased or reduced loyalty: (1) tangibility, (2) compatibility withthe firm's image, and (3) reward timing. However, scant empiricalresearch has compared these elements.

Transaction theory (Thaler, 1983) explains the value derived bya customer from a LP exchange, which consists of two drivers:Acquisition utility represents the economic gain or loss from thetransaction realized within the LP. Transaction utility is associatedwith purchase within the LP and represents the pleasure (or dis-pleasure) of the financial deal per se.

3.1.1. Tangibility of the rewardReward tangibility depends on the relative level of abstraction.

Tangible or hard benefits include monetary incentives (e.g., dis-counts, vouchers), whereas intangible or soft rewards providepsychological, relational, emotional, and functional benefits (e.g.,preferential treatment, elevated sense of status, services, specialevents, entertainment, priority check-in; Arbore and Estes, 2013;Drèze and Nunes, 2009). Tangible rewards (e.g., free hotel stays,tickets) may be provided as rewards, but some research suggeststhat these forms of acquisition utility have an impact on short-term behavior but limited effects on relationship quality (Yi andJeon, 2003). Many programs now increase the customers' trans-action utility, which includes such intangible rewards as privilegedaccess to websites and members-only newsletters. Empirical evi-dence suggests that tangible rewards are preferred to intangibleones and create higher loyalty intentions (likelihood that a con-sumer will purchase a particular product or service in the future(Keh and Lee, 2006; Yi and Jeon, 2003). Thus

H1. The (a) preference of an LP (b) and store loyalty intentions are

L. Meyer-Waarden / Journal of Retailing and Consumer Services 24 (2015) 22–3224

higher for tangible than intangible rewards.

3.1.2. Compatibility of the reward with the firm's imageThe compatibility of the reward with the firm's image depends

on the overlap with positive associations of the sponsoring com-pany, which reinforce the value of the offer. Direct and compatiblerewards focus on the focal firm (e.g., “Buy 1 perfume, get a beautyservice”), whereas indirect incompatible rewards are unrelated(e.g., “Buy 1 perfume, get a DVD”). The more significant theoverlap, the more compatible the reward is with the focal firm.Acquisition utility is high for compatible benefits of the rewardprogram and includes economic rewards, which some researchsuggests may be more effective than incompatible awards (Kivetz,2005; Yi and Jeon, 2003). For example, consumers who purchase aperfume are likely to appreciate earning a beauty service. Not onlyis the relationship between the loyal behavior and program out-come consistent, but also it is clear that these customers are in-terested in the principal product cosmetics and beauty services. Incontrast, evidence suggests that providing incompatible rewards,typically, for extraneous goods or services, is a suboptimal rewardpractice and may even be harmful to promoting loyalty (cf. Kimet al., 2001). Thus

H2. The (a) preference of an LP (b) and store loyalty intentions arehigher for compatible than incompatible rewards.

3.1.3. Reward timingReward timing affects illusionary progress to obtain the LP

gratification, which in turn enhances post-enrollment inferences(e.g., pursuit, loyalty, recommendation likelihood; Bagchi and Li,2011; Kivetz et al., 2006). Rewarding in LPs may vary in the timingbetween the purchase behavior and the actual reward redemption.The time required to obtain rewards is thus a key variable, influ-encing both motivation and behavior (Bootzin et al., 1991; Hittet al., 1992). This shaping process usually occurs through twomechanisms resulting in successive reinforcements: “points pres-sure” and “rewarded behavior.” The points' pressure mechanism isthe short-term impact, in which customers increase their purchaserate to earn rewards. Conversely, the rewarded behavior me-chanism is the long-term impact, in which clients increase theirpurchase rate after they have received the reward (Taylor andNeslin, 2005). In designing LPs, managers therefore must decidewhether to offer immediate or delayed rewards, such that con-sumers must engage in the related effort for less or more time toobtain the reward (Drèze and Nunes, 2006; Nunes and Drèze,2006; Sorman, 1998). Empirical evidence suggests that, pre-ferences for reward timing vary under a number of situations.Consumers' preferences regarding rewards' timing shift as con-sumer effort increases (Kivetz, 2003). When the required con-sumer effort and timing are low, consumers prefer immediate low-magnitude, guaranteed rewards. But as required effort and timingincrease, they prefer larger rewards, even if they are less certain,providing evidence of a “lottery” effect. Nevertheless, all else beingequal, customers prefer immediate rewards whereas firms preferdelayed gratifications to build exit barriers (Bagchi and Li, 2011; Yiand Jeon, 2003). Thus

H3. The (a) preference of an LP (b) and store loyalty intentions arehigher for immediate than delayed rewards.

3.1.4. Personal consumer involvement as moderator variablePersonal involvement with a product has gained a central place

in the consumer research literature for the past three decades(Lesschaeve and Bruwer, 2010; Quester and Lim, 2003) as it isthought to have considerable influence over the consumer beha-viors and the decision making process. The moderating influence

of involvement on relationships between the marketing variables(including LPs) and consumers' attitudes, brand preference, per-ceptions, satisfaction, loyalty has been established (Laurent andKapferer, 1985; Traylor and Joseph, 1984). Although there arevarious views of involvement, it is generally accepted that a con-sumer's personal involvement in a product category reflects a stateof motivation, awareness, importance, attraction, interest, goal-directed emotional state that determines the personal relevance toproducts or services in a particular context of a purchase decision(Laurent and Kapferer, 1985). Involvement thus stems from theconsumer's perception that the product class meets importantvalues, goals, or interests while a consumer takes a choice deci-sion. Most literature classifies involvement as either high or low(Aurifeille et al., 2002; Celsi and Olson, 1988). The consequences ofperceived pertinence of a product category (high or low) include asearch for and processing of information and decision-making(Zaichkowski, 1985). High- and low-involvement consumers arebelieved to behave differently (Bei and Widdows, 1999). Con-sumers with high involvement having a high degree of interest forthe product or service, tend to be information-seekers resulting ina concomitant degree of knowledge and seek to maximize ex-pected satisfaction through an extensive choice process (Laurentand Kapferer, 1985; Hollebeek et al., 2007; Barber et al., 2008).Therefore, these consumers segments go through extensive stagesof awareness, comprehension attitudes and behaviors (Laurentand Kapferer, 1985).

As some product categories or sectors involve their consumersmore than others we suggest that the influence of consumers'personal involvement acts as a moderator of the process in whichthe type of LP rewards (compatibility, tangibility, timing) operateon preferences and loyalty intentions.

In situations with high personal consumer involvement, fortop-range or high-preference brands with high competitor differ-entiation, the brand purchased offers one of the primary motiva-tions, not the LP reward (Roehm et al., 2002). Customers partici-pate more actively in the search for information, and informationabout the product or brand becomes more important than rewardinformation. Therefore, value derives from the intrinsic char-acteristics of the product or the brand itself, not from the LP re-wards (Rothschild and Gaidis, 1981). Thus, customers perceivemore value from intangible rewards that are compatible with theimage or value proposition of the firm or the brand offering the LP,leading to high goal congruity and transaction utility (Daryantoet al., 2010). Even if customers prefer immediate rewards, all elsebeing equal, delayed rewards are probably more viable in highpersonal involvement conditions (Daryanto et al., 2010). Con-sumers who are highly involved with a LP are more ready to waitfor delayed rewards with high value rather than experiencingmore immediate, lower-value rewards (Keh and Lee, 2006). Theimmediacy of the reward becomes secondary because clients focusmore on the attributes of the brand than on the reward (Roehmet al., 2002). Thus, customers are more likely to accept intangible,delayed, uncertain future rewards with higher value (Keh and Lee,2006). The consequences of these rewards creating transactionutilities should include more sustainable loyalty and relationshipmotivation around the brand (Keh and Lee, 2006; Morgan andHunt, 1994; Yi and Jeon, 2003), rather than the program, by en-hancing customers' brand associations and attitudinal commit-ment (e.g., a sense of belonging; Dowling and Uncles, 1997), be-cause the intangible and direct nature of the rewards does notinterfere with the brand or draw attention away from the brand'sintrinsic characteristics (Roehm et al., 2002; Yi and Jeon, 2003;Drèze and Nunes, 2009; Phillips Melancon et al., 2010). Behavioralreinforcement (rewarded behavior) due to the satisfaction withthe rewards, combined with a learning effect due to future re-wards, emerges (Frisou and Yildiz, 2011; Rothschild and Gaidis,

L. Meyer-Waarden / Journal of Retailing and Consumer Services 24 (2015) 22–32 25

1981; Taylor and Neslin, 2005). Consumers who are satisfied witha LP are more ready to wait for delayed rewards with high valuerather than experiencing more immediate, lower-value rewards.These results are consistent with the notion that high-involvementconsumers are ready to wait for delayed rewards (Yi and Jeon,2003). Considering these explanations we posit therefore the fol-lowing hypotheses. In high personal involvement conditions:

H4. The (a) preference of an LP (b) and store loyalty intentions arehigher for intangible than tangible rewards.

H5. The (a) preference of an LP (b) and store loyalty intentions arehigher for compatible than incompatible rewards.

H6. The (a) preference of an LP (b) and store loyalty intentions arenot different for immediate and delayed rewards.

On the other hand, in situations with low personal involve-ment, for low-range or low-preference brands with low compe-titor differentiation, the LP reward is one of the primary motiva-tions, not the brand purchased. Purchases are characterized byhigh inertia and geographic proximity. Preference is derived fromthe LP rewards, and less from the intrinsic characteristics of theproduct, brand or store itself. This causes customers to regard thefirm's reward compatibility as secondary and the tangibility andreward timing as primary creating acquisition utilities and nottransaction utilities (Rothschild and Gaidis, 1981). Thus, customersperceive more value from immediate, tangible rewards regardlessof whether these rewards have high goal congruity or are (in)compatible with the image of the firm offering the LP (Daryantoet al., 2010). The concrete features of an immediate, tangible re-ward, such as face value or presentation (e.g., percentages, dollardiscounts), are more relevant for determining their favorability(Roehm et al., 2002). Though attractive to customers, such rewardsmight cause inefficiency due to high unit costs (Palmatier et al.,2009) and may create interference in attention to the brand's in-trinsic characteristics, which might induce spurious and short-term loyalty because customers do not necessarily search for astronger relationship with the company (Keh and Lee, 2006;Phillips Melancon et al., 2010; Yi and Jeon, 2003). Delayed rewardsare therefore less attractive than immediate ones (Keh and Lee,2006). Consumers who are not highly involved with a LP are lessready to wait for delayed rewards with high value and want toexperience more immediate, lower-value rewards (Keh and Lee,2006). Alternatively, dissatisfied consumers prefer more im-mediate and lower-magnitude rewards (Keh and Lee, 2006). Theseresults are consistent with the notion that low-involvement con-sumers prefer more immediate rewards (Yi and Jeon 2003). Con-sumers do not change their purchasing behavior substantively butrather make the most of the immediate rewards before resuminghabitual behaviors, without any reinforcement in the long run

Table 1Hypothesized effects of rewards on a) LP preference b) store loyalty inte

H1: Tangibility-(a) LP preference, (b) store loyalty intentionH2: Compatibility-(a) LP preference, b) store loyalty intentionH3: Timing of-(a) LP preference, (b) store loyalty intention

High personal category involvementH4: Tangibility-(a) LP preference, (b) store loyalty intentionH5: Compatibility-(a) LP preference, (b) store loyalty intentionH6: Timing of-(a) LP preference, (b) store loyalty intention

Low personal category involvementH7: Tangibility-(a) LP preference, (b) store loyalty intentionH8: Compatibility-(a) LP preference, (b) store loyalty intentionH9: Timing of-(a) LP preference, (b) store loyalty intention

Notes: The “4” sign indicates stronger preference (utility) and intendedloyalty, and the “o” sign reveals weaker preference (utility) and intend

(Rothschild and Gaidis, 1981). “Points pressure” behavior, notcombined with a learning effect due to future rewards, thusemerges (Frisou and Yildiz, 2011; Taylor and Neslin, 2005). Withthe reward in hand, consumers have no reason to buy more andreturn to old habits (Meyer-Waarden, 2013). Considering theseexplanations we posit therefore the following hypotheses: In lowpersonal involvement conditions:

H7. The (a) preference of an LP (b) and store loyalty intentions arehigher for tangible than intangible rewards.

H8. The (a) preference of an LP (b) and store loyalty intentions arenot different for compatible than incompatible rewards.

H9. The (a) preference of an LP (b) and store loyalty intentions arehigher for immediate than delayed rewards.

Our model thus examines how the rewards (tangibility, com-patibility, and reward time) affect the preference of the LP andinduce store loyalty intentions. Personal involvement serves as amoderator variable (for a summary of our hypotheses, see Table 1).

4. Methodology and data

We first develop our scenarios, questionnaire measure instru-ments and pretest them in order to purify them by a measurementmodel analysis. We then apply the questionnaire scales to ran-domly selected LP members in two sales outlets situated in Tou-louse (France). One of the points of sales belongs to a marketleading grocery retailer in France (20% market share just behindCarrefour with 20.5%; Kantar May 2014), with 561 self-servicestores; the other outlet belongs to a French leader in the dis-tribution of high value cosmetics, with 1220 points of sales. Thesetwo settings differ in the degree of personal involvement (personalinvolvement in the purchase of grocery goods should be lowerthan in perfume purchases; Zaichkowski, 1985) in order to see ifour results hold in these different consumption domains. Per-fumeries as a product category have all the attributes that Laurentand Kapferer (1985) argue are the source of involvement. It has theability to provide the consumer with the pleasure value. Further-more, this category has a significant sign value and is perceived bysome as an important product. There is also a perceived risk whenpurchasing a perfume or cosmetics.

4.1. Conjoint profiles

We employ a full profile conjoint analysis. Conjoint analysis issuitable to test fictive profiles. The method offers a realistic choicesimulation and calculates global utility or preference scores—thatis, preferences for a product based on global judgments of the full

ntions.

Tangible 4 intangibleCompatible4 incompatibleImmediate4delayed

Intangible4tangibleCompatible4 incompatibleImmediate¼delayed

IntangibleotangibleCompatible¼ incompatibleImmediate4delayed

loyalty, the “¼” sign means equal preference (utility) and intendeded loyalty.

Table 2Orthogonal array and holdout sample.

Attribute Compatibility Tangibility Timing of rewardLevel Strong (1) Tangible (1) Immediate (1)

Weak (2) Intangible (2) Delayed (2)

Validation sampleProfile 1 1 1 1Profile 2 1 1 2Profile 3 1 2 1Profile 4 1 2 2Profile 5 2 1 1Profile 6 2 1 2Profile 7 2 2 1Profile 8 2 2 2

Holdout sampleProfile 1 1 2 1Profile 2 2 1 2

Table 4Discriminant and convergent validity: Personal involvement.

Importance Interest Attraction Sign Value

Groc Perf Groc Perf Groc Perf Groc Perf

Importance .83a .90a

.68 .81Interest .02n .06n .94a .86a

.004 .003 .89 .74Attraction .53n .22n .16n .16n .79a .8a

.29 .048 .02 .025 .63 .64Sign value .17n .09n .07n .06n .31n .28n .89a .79a

.03 .008 .00 .003 .09 .078 .80 .62

a Diagonal elements are the square roots of the AVE of the concerned con-structs or factors.

n Significant at po .01.

L. Meyer-Waarden / Journal of Retailing and Consumer Services 24 (2015) 22–3226

profiles (Green and Srinivasan, 1990). Extensive literature hasshown its robustness and adequacy since the early work of Greenand Srinivasan (in the 1970s).

Two sets of eight fictitious LP profiles, generated by an ortho-gonal design (SPSS Orthoplan), feature three attributes with twolevels each (eight profiles; see Tables 2–4):

1.

TabMe

S

Im

In

A

S

It

Tangibility of the reward: tangible or intangible.

2. Compatibility of the reward with the image of the store: strong

or weak.

3. Time before obtaining the reward: immediate or delayed.

To test for predictive validity, we add a holdout sample of twosupplementary LP profiles generated by SPSS to the eight profilesfrom the validation sample.

Next, 20 experts representing the two stores (e.g., program,outlet, or marketing managers) identify rewards that correspondto each hypothetical profile in the conjoint design. To ensurecomparable programs, the monetary value of the reward con-sistently averages around 6% (Yi and Jeon, 2003). For example, animmediate €4 reward corresponds to 6.2% of the value of the

le 3asurement model (CFA): personal involvement.

ize Shopping in the store …

portance/pertinence …is important (5)….not impor…is essential (5) ….not essentVariance extractedCronbach's alpha

terest … is interesting (5)… not inte… is of major concern (5) … nVariance extractedCronbach's alpha

ttraction … is valuable (5) …not valuab…is vital (5) … not vital (1) to… worries me (5) ….does notVariance extractedCronbach's alpha

ign value … means a lot (5)… not mean… is of great value (5) … is ofVariance extractedCronbach's alpha

em removed “is very exciting … not very exχ2/sigRMSEAo .05GFI/AGFI/CFIZ .90

average shopping basket of €60 in both stores. A €30 delayed re-ward, obtained after about eight purchases of €60 on average (or€480 in total), similarly corresponds to a 6.2% value (for the hy-pothetical programs of both stores, see Appendix). Pretests ofthese profiles (N¼420 students form the universities Toulouse andStrasbourg, France) show that the comprehension of the attributesand levels is well understood for the respondents and confirm thecredibility, comparability, and clarity of the conjoint design.

4.2. Measurement scales and purification

Store loyalty intentions and involvement are measured withquestionnaire items, while the LP attributes (compatibility, timeand tangibility) and their respective levels are not measured di-rectly via questionnaire as they appear in the conjoint scenarios.The conjoint measurement methodology calculates the partialutilities or preferences of the attribute levels (compatible/in-compatible, tangible/intangible, delayed/immediate) from the or-dinal ranking of the eight full profiles by using an OLS regression(see point 4.4).

The measure of personal involvement relies on Zaichkowski's(1985) semantic differential scale. All items used five-point Likert

Grocery Perfumery

tant (1) to me. .988 .982ial (1) to me. .973 .964

25% 23%.96 .86

resting (1) to me. .781 .744o concern to me (1). .768 .756

19% 18%.83 .89

le (1) to me. .849 .939me. .718 .919worry me (1). .609 .866

18% 18%.83 .89

s a lot (1) to me. .881 .985no value to me (1). .789 .975

13% 11%.81 .87

citing to me”2.58/.2 2.88/.5.04 .04.98/.97 .96/.95

L. Meyer-Waarden / Journal of Retailing and Consumer Services 24 (2015) 22–32 27

scales (1¼“strongly disagree” to 5¼“strongly agree”), such thatrespondents indicated their degree of agreement with a series ofstatements about the stimulus object. This scale focuses on thepersonal relevance, interest, and attraction of the category. First,an exploratory factor analysis (oblimin rotation) done with theabove mentioned student sample (N¼420) highlights four factors(Zaichkowski, 1985) that account for 75% and 70% of the variancein the grocery retailer and perfumery stores, respectively: im-portance or pertinence, interest, attraction, and sign value (Gerb-ing and Anderson, 1988). Cronbach's alphas range from .85 and .95for the dimensions, which thus appear reliable. We perform twoconfirmatory factor analyses (CFA) on the basis of respectivelyrandomly selected 401 grocery retailer and 376 perfumery LPmembers (these respondents are not included in the final study).All indicators provide evidence of good model fit (GFI4 .9;AGFI4 .8, RMSEAo .05; χ2 (CMIN)40.1; Fornell and Larcker,1981). The Cronbach's alpha, are again all higher than.80 for thefour dimensions, which confirms reliability. We demonstrateconvergent validity (all item loadingsZ .7 on the latent variable).Discriminant validity is confirmed (latent variables share morevariance with their respective items than with other latent vari-ables and the variance extracted for all constructs is greater thanthe generally accepted value of .50; Fornell and Larcker, 1981). Weperform a manipulation check on personal involvement with the401 grocery retailer and 376 perfumery LP members. The averageaggregate score of involvement is higher for the perfumery thanfor the grocery-retailing store (4.1 vs. 1.7; po .01).

Finally, the behavioral store loyalty intention is measured by aJuster scale (0¼“no intention to repurchase in the store,”100¼“absolutely sure to repurchase in the store”), which assessesthe likelihood that a consumer will purchase in the store inquestion systematically for all relevant purchases in the futurebecause of its LP (Uncles and Lee, 2006; Wright et al., 2002).Previous research has found that the scale in its many applicationsis superior to other intentions scales as a predictive measure offuture purchase behavior (Brennan and Esslemont, 1994).

4.3. Final surveys

Between January and September 2012, 999 LP members of thegrocery retailer and 1100 LP members of the perfumery are ran-domly selected in the outlets of Toulouse to participate in personalinterviews that last 15 min. Survey times and dates range from10:00 A.M. to 8:00 P.M., Monday to Saturday. All gender and agecategories are well represented. In total, 82–90% hold loyalty cardswith the respective stores for more than a year (Table 5).

The interviews begin with questions about the level of personalinvolvement when purchasing from the outlet in question. For thesubsequent tests we use the average aggregate score of personalinvolvement that is significantly higher for purchases in the

Table 5Sample description.

Grocery (%) Perfumery (%)

GenderFemale 55 53Male 45 47

Age16–30 30 2931–45 37 3546þ 33 36

LP membershipLess than one year 10 18More than one year 90 82

perfumery than in the grocery store (3.7 vs. 1.6; po .01; Table 6).Next, respondents read the eight fictitious LP profiles and re-

wards corresponding to each store (see Appendix) and classifythese descriptions in order of preference, from the most desired(1) to the least (8). A random rotation before each survey helpsavoid systematic bias. For each LP, respondents indicate behavioralstore loyalty intentions, due to the LP, on the 0–100% loyalty in-tention scale. Finally, as a test of predictive validity, a holdoutsample of two supplementary LP holdout profiles (see Appendix)undertakes the same procedure of preference classification frommost desired (1) to the least desired (2). These evaluations do notenter into the calculation of partial utilities in the conjoint ana-lysis. For cross-validation, we randomly divide each sample in half,to obtain four sub-samples (N1Grocery¼500, N2Grocery¼499;N1Perfumery¼550, N2Perfumery¼550).

4.4. Assessment of partial utilities and LP preferences

We employ the SPSS Conjoint using an OLS regression (usingthe ranking of the 8 profiles) to calculate for each individual(N¼999 LP members of the grocery retailer, N¼1100 LP membersof the perfumery) the partial utilities of the attribute levels(compatible/ incompatible, tangible/intangible, delayed/im-mediate). An additive linear model without interactions betweenattributes fits best according to a test of predictive validity (Greenand Srinivasan, 1990), in which the theoretical and real utilities ofboth holdout samples are compared. The predictive validity of theadditive linear model without interactions is better (po .01) thanthat with the interactions. The correlations of the theoretical(calculated on the basis of partial utilities) and actual preferencerankings are high for all samples and subsamples (ρGroc¼ .91,ρPerf¼ .85; ρN1Groc¼ .89, ρN2Groc¼ .93; ρN1Perf¼ .83, ρN2Perf¼ .87;see Table 7).

5. Results

5.1. Overall results for the whole sample

The reward timing is the top choice criterion (36%), followed byreward tangibility (35%) and compatibility (29%), for determiningLP preference (Table 8). Preferences are higher for immediate thendelayed rewards (.36 vs. � .36), for tangible than for in intangiblerewards (.23 vs. � .23), and compatible than incompatible rewards(.19 vs. � .19). H1a, H2a and H3a are confirmed.

Two regression models are estimated measuring the impact of(a) the relative importance of rewards' attributes (time, tangibility,compatibility) as well as personal involvement and (b) the con-joint partial utilities (preferences) of attribute levels (immediate/delayed; strong compatibility/weak compatibility; tangible/in-tangible) on intended store loyalty (measured on the 0–100%Juster Scale). The Kolmogorov–Smirnov tests are not significant(p4 .05), so the normality postulate is respected. The values of theDurbin–Watson tests and the variance inflation factor are less than2, indicating that multicollinearity is not a problem (seeTables 9 and 10).

The regression model measuring the impact of the relativeimportance of rewards' attributes on store loyalty shows thattiming has a negative impact (b¼� .42, po .01), tangibility(b¼ .36) has a positive influence on loyalty intentions (po .01).Reward compatibility with the store's image as well as personalinvolvement are not significant (p4 .1) (Table 9).

The regression model measuring the impact of the conjointpartial utilities of attribute levels (preferences) on store loyaltyintentions, shows that immediate (b¼ .24), compatible (b¼ .31)and tangible (b¼ .29) rewards have positive (and higher)

Table 6Personal involvement.

Size Shopping in the store … Grocery Perfumery Total

Importance/pertinence …is important (5)….not important (1) to me. 1.5 3.9 2.7…is essential (5) ….not essential (1) to me. 1.5 3.5 2.5

Interest … is interesting (5)… not interesting (5) to me. 2.1 4.5 3.3… is of major concern (5) … no major concern to me. 1.5 3.5 2.5

Attraction … is valuable (5) …not valuable (1) to me. 1.6 3.7 2.7…is vital (5) … not vital (1) to me. 1.1 2.9 2.0… worries me (5) ….does not worry me (1). 1.5 3.7 2.6

Sign value … means a lot (5)…does not mean anything (1) to me. 1.8 3.8 2.8… is of great value (5) … is of no value to me (1). 1.9 4.2 3.1

Involvement average (over the 10 items) by sector 1.6nn 3.7nn 2.6

Scales range from 1 to 5.nn Significantly different at po .01.

Table 7Predictive validity: correlation of prediction first choices in theholdout sample.

Model without interaction Model with interactions

Grocery: .91n Grocery: .10Perfumery: .85n Perfumery: .11N1Groc¼ .89n, N2Groc¼ .93n N1Groc¼ .09, N2Groc¼ .11N1Perf¼ .83n, N2Perf¼ .87n N1Perf¼ .11, N2Perf¼ .11

n Differences between methods are significant at po .01.

Table 8Relative importance of attributes and conjoint partial utility levels (total sampleN¼2099 grocery retailer and perfumery).

Attribute Relative importanceattribute

Level Conjoint partialutility

Time 36% (1) Immediate .36Delayed � .36

Compatibility 29% (3) Strong .19Weak � .19

Tangibility 35% (2) Tangible .23Intangible � .23

Note: The conjoint partial utility is set equal to preference.

Table 9Effects of the relative importance of attributes on intended store loyalty.

Constant Totalsample

Groceryretailer

Perfumery Scheffé test dif-ferences (groceryvs. perfumery)2.5 2.8

Standardized regression coefficients

Time � .42 nn,t¼5.95

� .83 nn,t¼6.24

� .009 ns,t¼ .24

nn, F¼96.3

Compatibility .25 ns,t¼ .61

.01 ns,t¼ .81

.40 nn, t¼8.75 nn, F¼99.1

Tangibility .36 nn,t¼5.56

.75 nn,t¼4.23

� .029 n,t¼3.35

nn, F¼97.03

Personalinvolvement

.20 ns,t¼ .41

.02 ns,t¼ .73

.39nn, t¼5.99 nn, F¼89.12

R2 .64 .63 .65Durbin–Watson 1.63 1.61 1.62Kolmogorov–Smirnov

.34 .38 .29

nn po .01, n po .05, ns: non-significant.

Table 10Effects of conjoint partial utility levels on intended store loyalty.

Constant Totalsample

Groceryretailer

Perfumery Scheffé test dif-ferences (groceryvs. perfumery)3.43 3.49

Standardized regression coefficients

Tangible .29 n,t¼4.34

.38 n,t¼3.61

.20 n, t¼4.38 n, F¼33.2

Intangible � .24 n,t¼ 3.32

� .30 n,t¼3.28

� .18 n, t¼3.39 n, F¼9.9

Strongcompatibility

.31 n,t¼2.56

.08 ns,t¼ .61

.54 nn, t¼6.64 nn, F¼92.3

Weakcompatibility

� .16 n,t¼2.12

� .03 ns,t¼ .73

� .30 n,t¼4.61

nn, F¼94.32

Immediate .24 n,t¼3.23

.48 n,t¼3.75

.01 ns, t¼ .51 nn, F¼83.61

Delayed � .15 n,t¼2.23

� .23 n,t¼2.92

� .08 ns,t¼ .92

nn, F¼95.9

R2 .65 .64 .67Durbin–Watson 1.75 1.89 1.67Kolmogorov–Smirnov

.70 .66 .75

nn po .01, n po .05, ns: non-significant.

Table 11Test of the moderator role of personal involvement in the store category (grocery/perfumery).

χ2 Value df

LP attributes-preferenceIndependent 3815.51 2178Constrained 3920.38 2256Δ 104.86 78p .02

LP attributes-loyalty intentionsIndependent 4631.12 2518Constrained 4799.76 2639Δ 168.64 121p .01

L. Meyer-Waarden / Journal of Retailing and Consumer Services 24 (2015) 22–3228

influences, whereas delayed (b¼� .15), compatible (b¼� .16) andintangible (b¼� .24) rewards have negative (thus lower) influ-ences (po .05) (Table 10). The compatibility levels (strong/weak)are not significant (p4 .1). H1b, H2b and H3b are confirmed.

Table 12Relative importance of attributes and conjoint partial utility levels (subsamplesgrocery and perfumery).

Grocery retailer

Attribute Relative importance attribute Level Conjoint partialutility

Sample Total N¼999 (subsample N1¼500/N2¼499)

Time 50% (51%/49%) Immediate .50 (.51/.49)s¼ .14

(1) Delayed � .50 (� .51/� .49)s¼ .14

a)nn b)nn, F¼89.12

Compatibility 10% (9%/11%) Strong .10 (.09/.11)s¼ .06

(3) Weak � .10 (� .09/� .11)s¼ .06

a)nn b)ns F¼1.23

Tangibility 40% (42%/38%) Tangible .39 (.42/.36)s¼ .12

(2) Intangible � .39 (� .42/� .36)s¼ .12

a)n b)n, F¼41.61

Perfumery

Attribute Relative importanceattribute

Level Conjoint partialutility

Sample Total N¼1100 (subsample N1¼550/N2¼550)

Time 22% (23%/21%) Immediate .22 (.23/.21)s¼ .09

(3) Delayed � .22 (� .23/� .21)s¼ .09

a)nn b)ns, F¼3.1

Compatibility 48% (49%/49%) Strong .47 (.47/.47)s¼ .17

(1) Weak � .47 (� .47/� .47)s¼ .17

a)n b)nn, F¼127.2

Tangibility 30% (31%/29%) Tangible .30 (.30/.30)s¼ .13

(2) Intangible � .30 (� .30/� .30)s¼ .13

a)n b)n, F¼32.3

a) χ2 Test difference importance attributes (grocery vs. perfumery): nn po .01,n po .05, ns: non-significant.b) Anova: nn po .01, n po .05, ns: non-significant.

L. Meyer-Waarden / Journal of Retailing and Consumer Services 24 (2015) 22–32 29

5.2. Moderating effects of personal involvement according to theretailing contexts

A 2(compatible vs. in compatible)�2(immediate vs.delayed)� (tangible� intangible) between-subjects ANOVA with apost-hoc Scheffé test shows that the two grocery contexts (codedfor this target with 1¼grocery retailing with low involvement, and2¼perfumery with high involvement) differ in terms of the im-pact of (a) the attributes and, as well as (b) the conjoint partialutilities of attribute levels on LP preferences and intended storeloyalty (see Tables 9, 10, and 12). A test of the moderator personalinvolvement with a multi-group analysis is realized with AMOS(Hayes and Matthes, 2009) and shows significant differences be-tween the two grocery contexts (see Table 11). A χ2 comparison ofthe (un)constrained models supports the moderating effects ofpersonal involvement on LP reward scheme preferences(χ2¼3.815, χ2¼3.920, po .05) and finally store loyalty intentions(χ2¼4.631, χ2¼4.799, po .01).

In the grocery retailer store, with low personal involvement,

the reward timing is the top choice criterion (50%), followed byreward tangibility (40%) and compatibility (10%), for determiningLP preference (see Table 12). The same results appear across bothsub-samples (N1Grocery, N2Grocery). In the perfumery, with highpersonal involvement, the compatibility of the reward with theretailer's image is the most important variable (48%), followed bytangibility (30%) and timing (22%), for both the total sample andthe sub-samples. A χ2 test is significant (po .01 and po .05) andshows that the importance of the attributes differ in both sectors.

In the perfumery, with high personal involvement, tangibilityhas a significant main effect on LP preferences (F(32.3), po .01).Preferences are higher for tangible than for in compatible rewards(.30 vs. .30, po .01). We reject H4a (as preferences are not higherfor intangible than tangible rewards). Compatibility has a sig-nificant main effect on LP preferences (F(127.2), po .01). LP pre-ferences of the LP are higher for compatible than for in compatiblerewards (.47 vs. � .47, po .01). We confirm H5a. The main effect oftiming of reward is not statistically significant (F(3.1), p4 .1), evenif the preferences are higher for immediate and delayed rewards(.22 vs. � .22, but p4 .1). We thus confirm H6a (the preference isnot different for immediate and delayed rewards).

In the grocery retailer store, with low personal involvement,tangibility has a significant main effect on LP preferences (F(41.61),po .05). Preferences are higher for tangible than for in intangiblerewards (.39 vs. � .39, po .05). We confirm H7a. Compatibility hasnot a statistically significant effect on LP preferences (F(1.23), p4 .1),even if the preferences of the LP are higher for compatible than forincompatible rewards (.10 vs. � .10, p4 .1). We confirm H8a (thepreference is not different for compatible and incompatible re-wards). The main effect of timing of reward is significant (F(89.12),po .01). The preferences are higher for immediate then delayedrewards (.50 vs. � .50, po .01). We thus confirm H9a.

The regression model measuring the impact of the relativeimportance of rewards' attributes on store loyalty shows that forthe perfumery, compatibility has the greatest influence on beha-vioral store loyalty intentions (b¼ .40, po .01). Tangibility exerts anegative impact on this variable (b¼� .029, po .05), and timebefore obtaining the reward is not significant (p4 .1). Personalinvolvement has a positive (b¼ .39) significant impact on loyalty(po .01) (see Table 9). Conversely, for the grocery retailer, tangi-bility (b¼ .75) has the greatest influence on loyalty intentions(po .01). Reward timing has a negative impact (b¼� .83, po .01),and compatibility with the store's image as well as personal in-volvement are not significant (p4 .1).

The regression model measuring the impact of the conjointpartial utilities of attribute levels (preferences) on store loyaltyintentions, shows that in the perfumery store, strongly compatible(b¼ .54) and tangible (b¼ .20) rewards increase behavioral storeloyalty intentions, whereas weak compatibility (b¼� .30) and in-tangible rewards (b¼� .18) have negative influences (po .05) (seeTable 10). H4b (intangible rewards increase loyalty intentionsmore than tangible ones) is thus rejected; H5b (compatible re-wards increase loyalty intentions more than incompatible ones) isconfirmed. The time elapsed before obtaining the reward (im-mediate /delayed) is not significant (p4 .1). H6b (no difference ofthe impact of immediate or delayed rewards on loyalty intentions)is confirmed.

For the grocery retailer, immediate (b¼ .48) and tangible(b¼ .38) rewards have positive influences, whereas delayed(b¼� .23) and intangible (b¼� .30) rewards have negative influ-ences (po .05). The compatibility levels (strong/weak) are notsignificant (p4 .1). H7b (tangible rewards increase loyalty inten-tions more than intangible ones), H8b (no difference of the impactof compatible or incompatible rewards on loyalty intentions) and9b (immediate rewards increase loyalty intentions more than de-layed ones) are confirmed.

Table 13Summary of results and hypotheses.

H1: Tangibility-(a) LP preference, (b) loyalty intention Tangible4 intangible Valid./valid.H2: Compatibility-(a) LP preference, (b) loyalty intention Compatible4 incompatible Valid./validH3: Timing of-(a) LP preference, (b) loyalty intention Immediate4delayed Valid./valid.

High personal category involvementH4: Tangibility-(a) LP preference, (b) loyalty intention Intangible4tangible Reject./rejectH5: Compatibility-(a) LP preference, (b) loyalty intention Compatible4 incompatible Valid./validH6: Timing of-(a) LP preference, (b) loyalty intention Immediate¼delayed Valid./valid

Low personal category involvementH7: Tangibility-(a) LP preference, (b) loyalty intention Intangibleotangible Valid./valid.H8: Compatibility-(a) LP preference, (b) loyalty intention Compatible¼ incompatible Valid./validH9: Timing of-(a) LP preference, (b) loyalty intention Immediate4delayed Valid./valid.

Notes: The “4” sign indicates stronger preference (utility) and intended loyalty, the “¼” sign means equal preference (utility) and intended loyalty, and the “o” sign revealsweaker preference (utility) and intended loyalty.

L. Meyer-Waarden / Journal of Retailing and Consumer Services 24 (2015) 22–3230

6. Discussion

LPs have gained popularity since American Airlines launched itsfrequent-flier program in 1981. Despite their importance, little isknown about how LPs influence post-enrollment value percep-tions. The key results presented extend the previous studies ofDowling and Uncles (1997) as well as Yi and Jeon (2003). Our re-search reveals that the underlying effects of reward types onpreferences and intended store loyalty differ depending on thelevel of consumers' personal involvement in the store. Table 13summarizes the hypotheses and results.

Overall, we confirm H1a and b, which stipulate that the pre-ference of an LP and, consequently, store loyalty intentions arehigher for tangible than intangible rewards. This contrasts theresults of Roehm et al. (2002) who show that intangible rewardsare preferred in all sectors. In support of H2a and b as well as Yiand Jeon (2003), LP preference and store loyalty intentions in-crease for rewards that are compatible with the image of the storeoffering the LP. They lead to high goal congruity and transactionutility (Daryanto et al., 2010). Finally, we confirm H3a and b, whichstipulate that immediate rewards compared to delayed rewardslead to higher preference and store loyalty intentions for (Yi andJeon, 2003).

Involvement clearly moderates the relationship between the LPreward scheme and LP preferences (Laurent and Kapferer, 1985).

In high involvement conditions, in contrast with H4a and b(which are rejected), which stipulate that the preference of an LPand, consequently, store loyalty intentions are higher for in-tangible than tangible rewards, the results show superiority fortangible rewards (e.g., cosmetics are preferred to beauty serviceswith the same monetary value). This contrasts the results ofRoehm et al. (2002) who show that intangible rewards are pre-ferred in all sectors. In support of H5a and b as well as Yi and Jeon(2003), LP preference and store loyalty intentions increase for re-wards that are compatible with the image of the store offering theprogram (e.g., beauty treatments in the perfume store). They leadto high goal congruity and transaction utility (Daryanto et al.,2010). Not only is the relationship between the loyal behavior andprogram outcome consistent, but also it is clear that these custo-mers are interested in the related products. In contrast, evidencesuggests that providing indirect rewards is a suboptimal rewardpractice and may even be harmful to promoting loyalty (cf. Kimet al., 2001). Finally, we confirm H6a and b, which stipulate that inhigh personal involvement situation delayed and immediate re-wards lead to similar preference and store loyalty intentions. Butimmediate rewards lead to higher preference and store loyaltyintentions than delayed rewards, which have negative value, eventhough the differences are not significant. Our results are thus inline with Yi and Jeon (2003).

In low personal involvement conditions, the results are in linewith Roehm et al. (2002) and validate H7a and b, because thepreference of the LP and store loyalty intentions are higher fortangible than intangible rewards. In line with H8a and b, rewardsthat are compatible with the store's image have a similar pre-ference and trigger not more store loyalty intentions than in-compatible rewards (so our results are in line with Yi and Jeon(2003)). Finally, in line with Yi and Jeon (2003) and our H9a and b,preference and store loyalty intentions are higher for immediatethan delayed rewards, which create acquisition utilities; the pre-ference is derived from the LP rewards, and less from the intrinsiccharacteristics of the product, brand or store itself.

To summarize, we find that in both contexts, immediate re-wards, compatible rewards, and tangible rewards have the stron-gest value (Yi and Jeon, 2003). Weaker compatibility, intangible,and delayed rewards offer the lowest value (regardless of thesample or sub-samples analyzed).

The effects of reward compatibility, tangibility and rewardtiming on preferences are moderated by consumers' personal in-volvement. Under high personal involvement, compatibility hasthe highest significant effect on preferences of the LP; compatiblerewards are perceived to be more valuable than incompatible re-wards and offer highest transaction utilities, which are preferredby consumers on the long term (Verhoef, 2003). These transactionutility benefits have been shown to have enduring effects on long-term brand loyalty (Roehm et al., 2002). Under low personal in-volvement, timing of the reward has the highest significant effecton LP preferences; immediate rewards are perceived to be morevaluable than delayed rewards and create acquisition utilities.Nevertheless, the delayed and intangible rewards have less nega-tive effects in high than low involvement contexts. Therefore,proposing delayed and intangible rewards in high personal in-volvement situations “destroys” less value. Rewards with weakercompatibility with the image of the store offering the programresult in more negative effects in high than low involvementcontexts. Similarly, proposing weaker-compatibility rewards is less“value destructive” in low than high involvement contexts. Sectorsthat offer weaker-compatibility and immediate rewards enablecustomers to acquire them rapidly. Though attractive to custo-mers, they may create interference in attention to the brand's in-trinsic characteristics, (that can weaken the intrinsic character-istics of the brand's capital) which might induce spurious andshort-term loyalty because customers do not necessarily search fora stronger relationship with the company (Keh and Lee, 2006;Phillips Melancon et al., 2010). Instead clients focus on acquiringthe reward rather than the purchase of the product or brand,which becomes secondary. The probability of defection increaseswhen consumers attain this reward (Roehm et al., 2002).

L. Meyer-Waarden / Journal of Retailing and Consumer Services 24 (2015) 22–32 31

7. Managerial implications

This study thus deepens our understanding of how value per-ception of the LP rewards affects store loyalty. Consumers' perso-nal involvement influences the relative importance of these routesand has to be considered as an important factor in designing a LP.This is beneficial for brand managers to understand aspects howloyalty schemes influence customer value creation. For example,the effectiveness of LP may be undermined when an incompatiblereward is granted in the high-involvement situation or when adelayed reward is adopted in the low-involvement situation. Ourfindings can be used to design LP more effectively as they highlightthe need for differentiated management of rewards according tothe sectors and their associated involvement. Tangibility, com-patibility, and reward timing are key variables that enable man-agers to enhance LP or store loyalty. Managers must determinewhich rewards have the strongest influence on the value of LP. Theconcepts vary, depending on the business sector. A category in-fluences consumers' personal involvement, which in turn re-inforces or weakens the links among LP rewards, preferences, andloyalty (Kivetz et al., 2006).

8. Limitations and future research directions

From a methodological perspective, conjoint analysis suffersfrom methodological problems that might weaken the results. Forinstance, behavioral research shows that one of the main as-sumptions of conjoint analysis is IIA (independence of irrelevantalternatives) does not stand up in some consumer behavior si-tuations (Hausman and McFadden, 1984).

From a theoretical perspective, this study examines only alimited amount of the influence of LP on consumers' store loyaltyintentions. Further studies should consider the effects of rewardson real purchase behavior. Other research could integrate attitu-dinal loyalty indicators. Studies that include customer character-istics (e.g., variety seeking, inertia, purchasing orientations) mighttest how these variables moderate the link between LP utility andbrand loyalty. Finally, further research should include financialindicators, such as customer lifetime value (Hardie et al., 2005),because LP success ultimately depends on financial contributions.

Concerning the LPs' time dimension, future research couldbuild on these efforts by probing deeper into consumer evalua-tions of reward magnitude and timing. Specifically, do consumerssimply assess magnitude based on perceived monetary value, ordo they adopt a more comprehensive evaluation that accounts forother aspects of value (e.g., perceived sacrifice, preferential treat-ment, relational benefits)?

It might be fruitful to include some others customer char-acteristics. To complete the explanation about the timing of re-ward, beyond the level of involvement, it could be also relevant toinvestigate the temporal orientation (Drèze and Nunes, 2006).Another avenue of research is to analyze the link between thecustomer satisfaction and the LP (Demoulin and Zidda, 2008). Fi-nally, as markets become saturated with LP (18 LP membershipsper U.S. household; Hlavinka and Sullivan, 2011) consumers mightreact negatively and oppose different forms of resistance to them(El Euch Maalej and Roux, 2012). It would be definitively worth-while to investigate this topic.

Appendix. Design (hypothetical LP)

Classify the following 8 fictive LPs according to your preference(with most desired being “1” and the least desired “8”). Then,express your intention to repurchase in the store because of each

of the 8 LP (from 0% to 100%¼“no intention to absolutely sure”).8 scenarios for perfumery store (to calculate partial utilities)LP 1: Immediate discount worth €4 for purchases of over €60.LP 2: Discount worth €30 after 8 purchases of €60 (or €480 in

total).LP 3: Immediate personalized beauty treatment worth €4 for

purchases of over €60.LP 4: Beauty treatment worth €30€ after 8 purchases of €60 (or

€480 in total).LP 5: Immediate reward of 1 movie DVD worth €4 for pur-

chases of over €60.LP 6: Free admission to Disneyland worth €30 after 8 purchases

of €60 (or €480 in total).LP 7: Immediate 1-month Car Assistance insurance worth €4

for purchases of over €60.LP 8: 1-year Car Assistance insurance worth €30 after 8 pur-

chases of €60 (or €480 in total).8 Scenarios for grocery retailing store (to calculate partial

utilities).LP 1: Immediate discount worth €4 for purchases of over €60.LP 2: Discount worth €30 after 8 purchases of €60 (or €480 in

total).LP 3: Immediate priority check-out worth €4 rewarded for

purchases of over €60.LP 4: Home delivery service worth €30 rewarded after 8 pur-

chases of €60 (or €480 in total).LP 5: Immediate movie DVD worth €4 for purchases of over

€60.LP 6: Free admission to Disneyland worth €30 after 8 purchases

of €60 (or €480 in total).LP 7: Immediate 1-month Car Assistance insurance worth €4

for purchases of over €60.LP 8: 1-year Car Assistance insurance worth €30 after 8 pur-

chases of €60 (or €480 in total).Now, classify the following 2 fictive LPs according to your

preference (with the most desired being “1” and the least desired“2”). Then, express your intention to repurchase in the store be-cause of each of the 2 LPs (from 0% to 100%¼“no intention toabsolutely sure”).

Holdout sample perfumery (test predictive validity; not used tocompute partial utilities)

LP 1: Immediate beauty service worth €4 for purchases of over€60.

LP 2: 2 books worth €30 after 8 purchases of €60 (or €480 intotal).

Holdout sample grocery retailer (test predictive validity: notused to compute partial utilities)

LP 1: Immediate home delivery service worth €4 for purchasesof over €60.

LP 2: 2 books worth €30 after approximately 8 purchases of €60(or €480 in total).

References

Arbore, A., Estes, Z., 2013. Loyalty program structure and consumers’ perceptions ofstatus: Feeling special in a grocery store? J. Retail. Consum. Serv. 20 (5),439–444.

Aurifeille, J.-M., Quester, P.G., Lockshin, L., Spawton, T., 2002. Global vs internationalinvolvement-based segmentation: a cross-national exploratory study. Int.Mark. Rev. 19 (4), 369–386.

Bagchi, R., Li, X., 2011. Illusionary progress in loyalty programs: magnitudes, rewarddistances, and step-size ambiguity. J. Consum. Res. 37 (5), 888–901.

Barber, N., Ismail, J., Dodd, T., 2008. Purchase attributes of wine consumers withlow involvement. J. Food Prod. Mark. 14 (1), 69–86.

Bei, L.-T., Widdows, R., 1999. Product knowledge and product involvement asmoderator of the effectsof information on purchase decisions: a case studyusing the perfect information frontier approach. J. Consum. Aff. 33 (1), 165–186.

Bijmolt, T., Dorotic, M., Verhoef, P., 2010. Loyalty programs: generalizations on their

L. Meyer-Waarden / Journal of Retailing and Consumer Services 24 (2015) 22–3232

adoption, effectiveness and design. Found. Trends Mark. 5 (4), 197–258.Blattberg, R.C., Kim, B., Neslin, S.A., 2008. Database Marketing. Analyzing and

Managing Customers. Springer, New York.Bootzin, R.R., Bower, G.H., Crocker, J., Hall, E., 1991. Psychology Today. McGraw-Hill,

New York.Brennan, M., Esslemont, D., 1994. The accuracy of the juster scale for predicting

purchase rates of branded, fast-moving consumer goods. Mark. Bull. 5, 47–52.Bridson, K., Evans, J., Hickman, M., 2008. Assessing the relationship between loyalty

program attributes, store satisfaction and store loyalty. J. Retail. Consum. Serv.15 (5), 364–374.

Celsi, R., Olson, J., 1988. The role of involvement in attention and comprehensionprocesses. J. Consum. Res. 15 (2), 210–224.

Daryanto, A., De Ruyter, K., Wetzels, M., Patterson, P., 2010. Service firms andcustomer loyalty programs: a regulatory fit perspective of reward preferencesin a health club setting. J. Acad. Mark. Sci. 38, 604–616.

Demoulin, N., Zidda, P., 2008. On the impact of loyalty cards on store loyalty: doesthe customers’ satisfaction with the reward scheme matter? J. Retail. Consum.Serv. 15 (5), 386–398.

Dowling, G.R., Uncles, M., 1997. Do customer loyalty programs really work? SloanManag. Rev. 38 (4), 71–82.

Drèze, X., Nunes, J.C., 2006. The endowed progress effect: how artificial advance-ment increases effort. J. Consum. Res. 32 (4), 504–512.

Drèze, X., Nunes, J.C., 2009. Feeling superior: the impact of loyalty program struc-ture on consumers’ perceptions of status. J. Consum. Res. 35 (6), 890–905.

Drèze, X., Nunes, J.C., 2011. Recurring goals and learning: the impact of successfulreward attainment on purchase behavior. J. Mark. Res. 48 (2), 103–115.

Ehrenberg, A.S.C., 1988. Repeat-buying: facts, theory and applications, 2nd ed.Edward Arnold, London; Oxford University Press, New York.

El Euch Maalej, M., Roux, D., 2012. Répertoires de critiques et conflits des mondes:une approche conventionnaliste des programmes de fidélisation. Rech. Appl.Mark. 27 (4), 60–94.

Ferguson R., Hlavinka K., 2009. The 2009 Colloquy loyalty marketing census. Col-loquy, http://www.colloquy.com/files/2009-COLLOQUY-CensusTalk-White-Pa-per.pdf.

Frisou, J., Yildiz, H., 2011. Consumer learning as a determinant of a multi-partnerloyalty program’s effectiveness: a behaviorist and long-term perspective. J.Retail. Consum. Serv. 18 (1), 81–91.

Fornell, C., Larcker, D.F., 1981. Evaluating structural equation models with un-observable variables and measurement error. J. Mark. Res. 18 (1), 39–50.

García Gómez, B., Gutiérrez Arranz, A., Gutiérrez Cillán, J., 2012. Drivers of customerlikelihood to join grocery retail loyalty programs. An analysis of reward pro-grams and loyalty cards. J. Retail. Consum. Serv. 19 (5), 492–500.

Gerbing, D., Anderson, J., 1988. An updated paradigm for scale development in-corporating unidimensionality and its assessment. J. Mark. Res. 25 (2), 186–192.

Green, P., Srinivasan, V., 1990. Conjoint analysis in marketing: new developmentswith implications for research and practice. J. Mark. 54 (4), 3–19.

Hardie, B., Lok Lee, K., Fader, P., 2005. Counting your customers the easy way: analternative to the Pareto/NBD model. Mark. Sci. 24 (2), 275–284.

Hausman, J.A., McFadden, D., 1984. Specification tests for the multinomial logitmodel. Econometrica 52, 1219–1240.

Hayes, A.F., Matthes, J., 2009. Computational procedures for probing interactions inOLS and logistic regression: SPSS and SAS implementations. Behav. Res.Methods 41 (3), 924–936.

Hitt, D.D., Marriott, R.G., Esser, J.K., 1992. Effects of delayed rewards and task in-terest on intrinsic motivation. Basic Appl. Soc. Psychol. 13 (4), 405–414.

Hlavinka, K., Sullivan, J., 2011. The Art and Science of Building Customer value.COLLOQUY loyalty census., Census Talk (April). http://www.colloquy.com.

Holbrook, M., 1996. Customer value: a framework for analysis and research. Adv.Consum. Res. 23, 138–142.

Hollebeek, L.D., Jaeger, S.R., Brodie, R.J., Balemi, A., 2007. The influence of in-volvement on purchase intention for new world wine. Food Qual. Prefer. 18 (8),1033–1049.

Keh, H.T., Lee, Y., 2006. Do reward programs build loyalty for services? The mod-erating effect of satisfaction on type and timing of rewards. J. Retail. 28 (2),127–136.

Kim, B.D., Shi, M., Srinivasan, K., 2001. Reward programs and tacit collusion. Mark.Sci. 20, 99–120.

Kivetz, R., 2003. The effects of effort and intrinsic motivation on risky choice. Mark.Sci. 22 (4), 477–502.

Kivetz, R., 2005. Promotion reactance: the role of effort-reward congruity.

J. Consum. Res. 31 (4), 725–736.Kivetz, R., Urminsky, O., Zheng, Y., 2006. The goal-gradient hypothesis resurrected:

purchase acceleration, illusionary goal progress, and customer retention. J.Mark. Res. 43 (1), 39–58.

Kwong, J.Y., Sorman, D., Ho, C.Y., 2011. The role of computational ease on the de-cision to spend loyalty program points. J. Consum. Psychol. 21 (2), 146–156.

Laurent, G., Kapferer, J.-N., 1985. Measuring consumers’ involvement profiles. J.Mark. Res. 22 (1), 41–53.

Leenheer, J., Bijmolt, T., van Heerde, H.J., Smidts, A., 2007. Do loyalty programsenhance behavioral loyalty? A market-wide analysis accounting for en-dogeneity. Int. J. Res. Mark. 24 (1), 31–47.

Lesschaeve, I., Bruwer, J., 2010. The importance of consumer involvement andimplications for new product development. In: Jaeger, S.R., MacFie, H. (Eds.),Consumer-Driven Innovation in Food and Personal Care Products, WoodheadFood Series no. 195. Woodhead Publishing Ltd., Cambridge, pp. 386–423.

Liu, Y., 2007. The long-term impact of loyalty programs on consumer purchasebehavior and loyalty. J. Mark. 70 (4), 19–35.

Meyer-Waarden, L., 2007. The effects of loyalty programs on customer lifetimeduration and share-of-wallet. J. Retail. 83 (2), 223–236.

Meyer-Waarden, L., 2008. The influence of loyalty programme membership oncustomer purchase behaviour. Eur. J. Mark. 42 (1–2), 87–114.

Meyer-Waarden, L., 2013. The impact of reward personalisation on frequent flyerprogrammes’ perceived value and loyalty. J. Serv. Mark. 27 (3), 183–194.

Meyer-Waarden, L., Benavent, C., 2009. Grocery retail loyalty program effects: self-selection or purchase behavior change? J. Acad. Mark. Sci. 37 (3), 345–358.

Morgan, R., Hunt, S., 1994. The commitment-trust theory of relationship marketing.J. Mark. 58 (3), 20–38.

Nunes, J.C., Drèze, X., 2006. The endowed progress effect: how artificial advance-ment increases effort. J. Consum. Res. 32, 504–512.

O’Brien, L., Jones, C., 1995. Do rewards really create loyalty? Harv. Bus. Rev. 73,75–82.

Palmatier, R.W., Jarvis, C.B., Bechkoff, J.R., Kardes., F.R., 2009. The role of consumergratitude in relationship marketing. J. Mark. 73 (5), 1–45.

Phillips Melancon, J., Noble, S., Noble, C., 2010. Managing rewards to enhance re-lational worth. J. Acad. Mark. Sci. 39 (3), 341–362.

Quester, P.G., Lim, A.L., 2003. Product involvement/brand loyalty: Is there a link? J.Prod. Brand Manag. 12 (1), 22–38.

Roehm, M.L., Bolman Pullins, E., Roehm Jr., H.A., 2002. Designing loyalty-buildingprograms for packaged goods brands. J. Mark. Res. 39 (2), 202–214.

Rothschild, M., Gaidis, W., 1981. Behavioral learning theory: its relevance to mar-keting and promotions. J. Mark. 45 (2), 70–78.

Sorman, D., 1998. The illusion of delayed incentives: evaluating future effort-moneytransactions. J. Mark. Res. 35 (4), 427–438.

Thaler, R., 1983. Transaction utility theory. In: Bagozzi, Richard P., Tybout, Alice M.(Eds.), NA-Advances in Consumer Research, vol. 10. Association for ConsumerResearch, Ann Abor, MI, pp. 229–232.

Taylor, G., Neslin, S., 2005. The current and future sales impact of a retail frequencyreward program. J. Retail. 81 (4), 293–305.

Temporal, P., Trott, M., 2001. Romancing the Customer: Maximising Brand ValueThrough Powerful Relationship Management. John Wiley & Sons, New York.

Tietje, B.C., 2002. When do rewards have enhancement effects? An availabilityvalence approach. J. Consum. Psychol. 12 (4), 363–373.

Traylor, B., Joseph, B., 1984. Measuring consumer involvement in products. Psychol.Mark. 1 (2), 65–77.

Uncles, M., Lee, D., 2006. Brand purchasing by older consumers: an investigationusing the Juster scale and the Dirichlet model. Mark. Lett. 17 (1), 17–29.

Verhoef, P.C., 2003. Understanding the effect of customer relationship managementefforts on customer retention and customer share development. J. Mark. 67 (4),30–45.

Vesel, P., Zabkar, V., 2009. Managing customer loyalty through the mediating role ofsatisfaction in the DIY retail loyalty program. J. Retail. Consum. Serv. 16 (5),396–406.

Wright, M., Sharp, A., Sharp, B., 2002. Market statistics for the Dirichlet model:using the juster scale to replace panel data. Int. J. Res. Mark. 19 (1), 81–90.

Yi, Y., Jeon, H., 2003. The effects of loyalty programs on value perception, programloyalty and brand loyalty. J. Acad. Mark. Sci. 31 (3), 229–240.

Zaichkowski, J., 1985. Measuring the involvement construct. J. Consum. Res. 12 (4),350–371.