Customer Retailer Loyalty

Journal of Retailing 80 (2004) 249–263

Customer retailer loyalty in the context of multiple channel strategies David W. Wallacea,∗ , Joan L. Gieseb,1 , Jean L. Johnsonb,2 b

a Department of Marketing, College of Business, Illinois State University, Normal, IL 61790-5590, USA Department of Marketing, College of Business and Economics, Washington State University, Pullman, WA 99164-4730, USA

Abstract With an increasingly competitive retail environment and decreasing customer switching costs, customer retailer loyalty is a critical goal for merchants of all types. We investigate customer retailer loyalty in the context of multiple channel retailing strategies. Results show that multiple channel retail strategies enhance the portfolio of service outputs provided to the customer, thus enhancing customer satisfaction and ultimately customer retailer loyalty. These results suggest that multiple channel retailing can be a useful strategy for building customer retailer loyalty. © 2004 New York University. Published by Elsevier. All rights reserved. Keywords: Customer loyalty; Multiple channel retailing; Internet marketing

Introduction Customer loyalty generates numerous benefits and hence is a critical aim of many marketing strategies (Jacoby & Chestnut 1978). Most importantly, customer loyalty creates a stable pool of customers for a firm’s product or service (Oliver 1997). A small shift in customer retention rates can make a large difference for earnings, and this influence accelerates over time. Loyal customers buy more, are willing to pay higher prices, and generate positive word of mouth, thus suggesting a strong link between loyalty and profitability (Reichheld 1993; Wright & Sparks 1999; Zeithaml, Berry, & Parasuraman 1996). Customer loyalty encapsulates both loyalty to the retailer and loyalty to the brand. Brand loyalty, in particular, has been extensively studied (Day 1969; Jacoby & Chestnut 1978; Oliver, 1997); however, little research has been conducted on the critical role of retailer loyalty. Customer retailer loyalty is of extreme interest to merchants, because high customer acquisition costs are difficult to recoup without repeat purchasing. This is ironic, particularly with the advent of In∗

Corresponding author. Tel.: +1 309 438 5066; fax: +1 309 438 5510. E-mail addresses: [email protected] (D.W. Wallace), [email protected] (J.L. Giese), [email protected] (J.L. Johnson). 1 Tel.: +1 509 335 6354; fax: +1 509 335 3865. 2 Tel.: +1 509 335 1877; fax: +1 509 335 3865.

ternet retailing, because increased competition and minimal customer switching costs make customers increasingly difficult to retain (Srinivasan, Anderson, & Ponnalovu 2002). Thus, efforts to enhance customer loyalty may be a critical defensive strategy for retailers: the existing customer base is both retained for the retailer and denied to its competitors (Fornell 1992; Jacoby & Chestnut 1978). Our key research question concerns the implications of a multiple channel interface for building customer retailer loyalty. Do merchants who invest in multiple channels receive a payoff in terms of customer loyalty? Do these multiple channels influence the drivers of customer satisfaction and ultimately loyalty? This question is surprisingly unanswered to date; empirical research has not adequately considered market-level responses to multiple channel retailing strategies (Homburg, Hoyer, & Fassnacht 2002; Reinartz, Krafft, & Hoyer, 2004). This is a compelling issue and of utmost importance to researchers (Peterson & Balasubramanian 2002), as well as practitioners (Reda 2002b). Multiple channel retailers simultaneously employ an array of channels consisting of retail stores, mail order catalogs, and web sites often targeting the same customer. Merchants may undertake multiple channel strategies for a variety of reasons: to gain legitimacy with key stakeholders (DiMaggio & Powell 1983); to respond to competitor actions within the industry (e.g., Grewal, Comer, & Mehta 2001); to save on transaction costs (Dutta, Heide, Bergen, & John 1995); and

0022-4359/$ – see front matter © 2004 New York University. Published by Elsevier. All rights reserved. doi:10.1016/j.jretai.2004.10.002

250

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

to increase market coverage (Friedman & Furey 2003). While these are no doubt important strategic reasons, it is also critical for both managers and researchers to gain insight into market-level responses to multiple channel retailing strategies and their implications for customer retailer loyalty and retailer profits. The major contribution of this research is to conceptualize and empirically demonstrate the critical role of multiple channel retailing in attaining strategic imperatives from their multiple channel customer markets—customer satisfaction that leads to customer retailer loyalty. The remainder of this paper proceeds as follows. First, we develop a conceptual framework and hypotheses. Next, we describe the research design incorporating a multi-mode survey undertaken in cooperation with an industry partner (a major regional retailer of specialty sporting goods). Empirical results from the main study data and a follow-up study are presented. The paper closes with a discussion of implications and future research suggestions.

Conceptual development With the advent of multiple channel retailing, the interface between merchants and their customers has become much more complex. Merchants typically augment their core product offerings with service outputs (e.g., product selection, attribute information, and extended hours of operation) provided before, during, and after purchase (Bucklin 1966; Stern & El-Ansary 1992). A portfolio of complementary channels makes available a greater and deeper mix of service outputs to the final customer (Frazier & Shervani 1992). With more

service outputs available across several channels, final customers have an opportunity to engage with a retailer over multiple contact points; this can occur during a single purchase or over multiple purchases. Because customers can have more of their service output needs met through multiple channels (and, even if not accessed, customers perceive that their needs can be met easily with multiple channels), we argue that multiple channel strategies develop customer retailer loyalty through enhanced satisfaction. With an increasingly competitive retail climate, retailers are extremely motivated to increase satisfaction and subsequently build customer loyalty. Multiple channel strategies may help retailers achieve this important goal. Fig. 1 displays the framework that drives our conceptual development. In general, we argue that increased multiple channel shopping by the customer increases the available package of service outputs and the potential number of contact points between the retailer and the customer. As the figure indicates, we theorize that this has major strategic consequences in customer satisfaction and customer retailer loyalty. Customer retailer loyalty Day (1969) argued that for true loyalty to be in effect the customer must both have a favorable attitude towards a product and purchase it repeatedly. The attitude component distinguishes between true loyalty and “spurious” loyalty due to high switching costs or a dearth of other choices. Because the evaluative nature of attitude towards something, in this case the retailer, may develop over time and have lasting duration

Fig. 1. Multiple channel loyalty framework.

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

(Eagly & Chaiken 1998), this aspect of loyalty includes a more enduring tendency that generalizes across multiple purchase instances. Dick and Basu (1994) characterized loyalty as a “relative attitude,” an appraisal of a behavioral choice relative to its alternatives. Merging these two perspectives, we conceptualize customer retailer loyalty as the customer’s attitudinal and behavioral preference for the retailer when compared with available competitive alternatives. In the case of multiple retailing channels, a merchant manages the store, catalog, and web site; many brands are offered, but only one retailer is identified across the channels. Thus, the real benefit of a multiple channel strategy is that this strategy encourages customers to be loyal to a specific retailer regardless of which channel(s) customers access and regardless of which brand they purchase. With more channel opportunities for the retailer to provide positive customer experiences via greater service outputs and contact with customers, we would expect the outcome to be increased customer retailer loyalty; that is, a preference for a particular retailer relative to competitors. Multiple channel strategies Multiple channel strategies may take a variety of forms. Friedman and Furey (2003) differentiate between channel mix and channel integration. In the former, each channel functions independently of the others as a stand-alone unit, providing a package of services that appeals to a particular group of customers; customers may visit different channels on different shopping occasions, depending on their current needs (Hansell 2002). Channel integration, on the other hand, involves a synergistic combination of channel functions (G¨orsch 2000). A clothing item can be tried on at a retail store and later purchased from the affiliated catalog, or items can be ordered online for express store pickup. It is plausible that multiple channel strategies in a competitive environment may have the effect of reducing customer loyalty. As channels multiply, the retailer’s market coverage increases. This increased customer access leads to decreases in the customers’ information search costs and increased price transparency. The resulting increased competition may lead to lower prices, higher price elasticities, frequent price changes, and narrow price dispersion—classic symptoms of market competition (Brynjolfsson & Smith 2000; Tang & Xing 2001). All of these factors decrease customer switching costs while at the same time increase customer motivation to switch, leading to an erosion in customer loyalty. We argue, however, that the cross-channel synergies made possible by a multiple channel retailing strategy will serve to enhance customer retailer loyalty. Channel types differ in their abilities to perform various service outputs (Bucklin, Ramaswamy, & Majumdar 1996). For example, retail stores provide excellent opportunities for prepurchase trial, instant gratification, and personalized attention, while Internet sites provide expanded accessibility, product information, and novelty (Grewal, Iyer, & Levy 2004). Catalogs offer superior

251

color and image reproduction while putting product information at a shopper’s fingertips without necessitating computer and Internet access. Because of these differential strengths, customers may use different channels for different purposes or at different times; for example, a gift book could be purchased online where it can be wrapped and drop shipped directly to its intended recipient, while a customer may purchase clothing at a retail store where it can be tried on. In addition to these single-channel uses for different purposes, technology creates opportunities for cross-channel synergies, for example, in-store Web kiosks and the combination of online ordering with express pick up (Peterson & Balasubramanian 2002). Because multiple complementary channels provide more, and more diverse, service outputs than single-channel strategies, when a merchant adds complexity to its customer interface by increasing customer contact points, it thereby expands both the quantity and possible combinations of service outputs available to its customers. For example, a customer may want to try on multiple styles of shoes, but may not be able to arrive at a decision until later in the evening; rather than heading back to the mall, the customer can order the exact product desired online. Conversely, a customer may feel confident ordering a product online sight-unseen, knowing that the product is returnable to a local retail store. Thus, we would expect that these multiple customer contact points would have important positive implications for customer shopping behavior, and these will lead to increased customer loyalty. This expectation represents the foundation for our research question and study design. Is the loyalty formation process influenced by multiple channel shopping behavior? Multiple channel shopping Market-level response is the strategic imperative for any marketing strategy. This research is concerned in particular with market response to multiple channel retailing. Recent market research suggests that customer shopping patterns have evolved to take advantage of the new multiple channel environment. Results indicate that 35 percent of the consumers surveyed shopped using some combination of catalogs, bricks-and-mortar stores, and the Internet; 66 percent said that they had visited one channel before purchasing from another (Saunders 2002). Jupiter Research has labeled this class of customers the “multi-channel shoppers” (Reda 2002a). These customers “are combining various channels and approaches, searching online to buy offline, searching offline to buy online—and everything in between” (Wind & Mahajan 2002, p. 65). We define customer multi-channel employment as the number of different channels a customer visits in making a purchase. It is important to note that like multiple channel retailing strategies, customer multi-channel employment manifests itself in a variety of ways. The most important distinction in this context is between shopping that crosses a merchant’s different channels (e.g., when a customer researches prod-

252

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

ucts at a Best Buy retail store and purchases at BestBuy.com) and shopping that crosses not only channels but merchants as well (e.g., researching with the Crutchfield catalog and purchasing from Comp USA’s retail store). Retailers whose customers engage in the former are most directly positioned to take advantage of the possibilities for differentiation provided by a multiple channel strategy and hence reap the benefits of enhanced customer loyalty. These customer shopping habits provide the most managerially actionable strategic options, in that cross-channel synergies as described above can be designed into a merchant’s multiple channel system. Multiple channel service outputs We define the multi-channel portfolio of service outputs as the package of service outputs that customers perceive to be available from the merchant’s combined mail order, retail, and Internet channels. Channels differ in the package of service outputs that they provide. As customers visit more channels, they thereby enhance the total package of service outputs on which they can draw. Thus, we predict that through multiple channel shopping, customers will encounter a higher level of available service outputs. The influence of multiple channel shopping on the perception of available service outputs forms the basic premise on which this customer retailer loyalty model is developed. Therefore: H1 . Customer multi-channel employment positively influences the perceived level of the multi-channel portfolio of service outputs. Satisfaction and customer retailer loyalty In the traditional expectancy disconfirmation satisfaction model (Oliver 1980), satisfaction results from the comparison of an initial standard and perceived variance from that standard. In this paradigm, customer expectations form the standard of comparison. Positive disconfirmation occurs when the customer evaluates the discrepancy between what actually occurred and these expectations thereby concluding that their expectations were met or exceeded. This positive disconfirmation leads to increased satisfaction (e.g., Oliver 1980; Spreng, MacKenzie, & Olshavsky 1996). Satisfaction is generally considered to be a customer’s summary affective response to the experience (Giese & Cote 2000; Oliver 1980). Because satisfaction is the “seed” out of which loyalty develops (Oliver 1999, p. 42), enhancing satisfaction is an important means for achieving loyalty. The objectives of loyalty strategies are to increase customer loyalty, increase the purchase amount of loyal customers, decrease customer loyalty to competitors, and decrease customer switching (Jacoby & Chestnut 1978). Therefore, offering a multi-channel portfolio of service outputs as a means to satisfy customer needs, leads to increased customer satisfaction and hence increased customer loyalty (Shankar, Smith, & Rangaswamy 2003).

In a multiple channel setting, customers enter into retail exchanges with expectations regarding the appropriate levels of service outputs they should encounter. Customers have learned to expect an increasingly high level of service outputs most likely because of their past experience across many channels. The convergence of catalog, retail store and the Web leads customers to expect more options and more service outputs (Wind & Mahajan 2002). These expectations form the standard of comparison for their current experience with the merchant. Customers with complex needs expect a complex portfolio of service outputs. For example, a customer who enters a store expecting considerable advice in choosing a product after examining product assortment and prices on the Internet site, and gets that advice and accuracy, will be more satisfied than one who needs such advice but is not able to get it. With the advent of multi-channel shopping, customers have expectations regarding the service outputs available to them from all of a retailer’s channels. Consistent with the expectancy disconfirmation paradigm, when customers find higher levels of service outputs available, their expectations are more likely to be positively disconfirmed, and this leads to increased satisfaction. Also consistent with the expectancy disconfirmation paradigm, when customers find lower levels or lower quality of service outputs available, their expectations will be negatively disconfirmed resulting in decreased satisfaction. There are two satisfaction issues here: raised customer expectations and the multiple channel merchant’s capability to meet those expectations in service output performance. Research has shown that multiple channel customers have higher expectations because of the nature of the multiple channel system (Shankar et al. 2003). As an advance to this logic, our contention is that multiple channel merchants are positioned to satisfy those expectations because they offer channel options designed to be in proportion to the service outputs expected. Not only do retailers employ multiple channel strategies to respond to competitors (Grewal et al., 2001); most importantly, they develop multiple channel strategies to respond to customer expectations. Thus, multiple channel customers are more likely to have their service output needs met, and hence: H2 . An increase in the multi-channel portfolio of service outputs leads to increased positive disconfirmation of the customer’s expectations. H3 . Disconfirmation positively influences satisfaction. Satisfying the end customer is strategically crucial to all channel members. It is seen as a means to competitive advantage (Day & Nedungadi 1994). Among its most strategic consequences, satisfaction leads to increased customer retention, decreased price elasticities, lower customer acquisition costs, and lower transaction costs (Anderson, Fornell, & Lehmann 1994; Lemon, White, & Winer 2002). Here we focus on a key variable associated with customer retention: customer retailer loyalty.

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

Theory suggests that satisfaction feeds back into the system to influence ensuing intentions and behaviors (Howard & Sheth 1969; Oliver 1980). A customer’s satisfaction with the shopping experience should reflect well on the merchant. When a retailer pursues a multiple channel strategy, the depth and diversity of service outputs it provides to customers is enhanced. This, as we have argued, leads to increased customer satisfaction; and an increase in satisfaction has been shown to result in increased customer loyalty (Fornell 1992; Oliver 1999). Although there are low switching costs to consummating purchases and intense pressure to lure competitors’ customers, we contend that when complex service output needs are satisfied with an individual merchant’s multiple channel portfolio of service outputs, then:

253

multiple channel customers than single-channel customers in that multiple channel customers perceive more service outputs to be available to them and have more potential contact points with the retailer; they are, therefore, more likely to have their expectations positively disconfirmed, and hence more likely to be satisfied. Finally, we expect that the influence of satisfaction on loyalty will be likewise higher for multiple channel customers. In sum, we argue that the conceptual model presented in Fig. 1 is more effective for multiple rather than single-channel customers. H5 . Our model (service outputs influence disconfirmation, which drives customer satisfaction, and thereby builds customer retailer loyalty) will better explain loyalty for multiple rather than single-channel customers.

H4 . Customer satisfaction will positively influence customer retailer loyalty. Method Some question has been expressed in the practitioner literature as to whether multiple channel customers are likely to be more or less loyal than single-channel customers (Reda 2002a). Some merchants feel that multiple channel customers are simply casting about for the best prices and hence will have no real loyalty to any particular merchant. Others argue that multiple channel retailing provides an opportunity to attract customers who may have been exposed to a product in one channel, but who are participating in another channel when they are ready to buy. A study by BizRate.com and J.C. Williams group found that dual channel customers spent an average of $600 more in a focal store than store-only shoppers (Pastore 2001). Similarly, another recent survey showed that multiple channel shoppers tend to spend more than singlechannel shoppers and that channels can effectively drive traffic to each other (Saunders 2002). Cross-channel synergies are clearly influencing customer purchase decisions. Recall, again, the example of a customer who tries on a pair of shoes in a store, but only decides later in the evening to purchase them via the Internet. It is likely that the focal retailer’s web site will be top-of-mind for the customer or at least included in the customer’s consideration set of alternative web sites (Reda 2002a). Channels differ in the levels and forms of service outputs that they provide (Bucklin, Ramaswamy, & Majumdar 1996; Grewal, Iyer, & Levy 2004). Because of this, multiple channel customers are able to assemble a more diverse and complete portfolio of service outputs than single-channel customers. As hypothesized in H1 , we expected that the customer’s use of multiple channels would influence the level of service outputs that the customer finds available. Furthermore, we expect that multiple channel shopping, relative to single-channel shopping, will have strong implications for customer/retailer relationships. In particular, we anticipate that the development of customer retailer loyalty will differ between single and multiple channel customers. Our model, which predicts that service outputs influence positive disconfirmation, satisfaction, and ultimately retailer loyalty, should be more descriptive of

The first phase of research involved ten semi-structured qualitative interviews with consumers. The interviews ensured that our research questions were relevant and aided in developing service output domains. Participants consistently reported engaging in multiple channel shopping behaviors and being influenced by the kinds of service outputs they encountered. We found consistent customer response to four key service outputs: convenience, selection, waiting time, and product information. Thus, they form the basis for our measures. The quantitative research targeted the customers of a single industry multiple channel partner, a large regional retailer of specialty outdoor sporting goods with bricks and mortar, mail order, and Internet businesses. By focusing on a single industry, the effects of variance not due to the measured constructs are reduced. The industry is structured along traditional lines, in which the product moves from manufacturers/suppliers to distributors or retailers and then on to end customers; thus, results from this research should generalize well to other specialty and shopping goods industries. By focusing on a single multiple channel retailer, this research design limits uncontrollable environmental variables. At the same time, it allows us to capture both single and multiple channel customers, since not all customers of a multiple channel retailer will take advantage of all channels. As well, single-channel retailers by definition can’t have their own multiple channel customers. Although this $5 billion industry has seen strong growth in recent years (Outdoor Industry Association 2002), its core distribution has traditionally been through small, independent bricks and mortar retail locations. The industry as a whole is currently struggling with a variety of multi-channel related issues, which threaten to inflict major structural changes. Manufacturers are experimenting with direct selling, and retailers are experimenting with opening additional channels. This industry is therefore an appropriate setting for research into multiple channel retailing strategies.

254

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

Sample and data collection The survey, with an incentive, was administered to the industry partner’s customers via store intercept, mail and email. The mixed mode approach ensures that the perspectives of a broad selection of the merchant’s customers are represented and mitigates coverage errors or other biases resulting from data collection mode. For example, a mail survey alone might generate a disproportionate number of responses from catalog-only shoppers. The survey explores customer experiences with the retailer as well as its competitors during a recent purchase. The mail survey was sent to 600 randomly selected recent (within the last three months) purchasers, with 151 completed surveys returned, for a 25 percent response rate. For the Internet customer survey, we culled the partner’s email list for 4,000 recent purchasers; they were sent an email from the president endorsing the research and a link to the survey site. Of these, 326 filled out the online questionnaire, for a response rate of 8.2 percent. Finally, 103 retail customers filled out the questionnaire on-site in the retail store. The retailer’s customer database indicated that our respondents had substantial average life-to-date purchases, both in dollar volume and in number of transactions, suggesting that these customers have repeated experience with the retailer and have developed related attitudes over time. There is no significant difference in purchase volume between respondents and nonrespondents (p = .33) and no differences between early and late respondents on any of the variables in this study (Armstrong & Overton 1977); we conclude that non-response bias should not be a problem.

Measure development Measure development followed standard research procedures (Churchill 1979). Reflective measures were adapted from existing scales, while the formative measures were new. The questionnaire document was peer reviewed and pretested on a student sample (n = 132), and items were reworded or dropped based on the pretest. After Confirmatory Factor Analysis (CFA), composites were formed by summing scale items. Please see Appendix A for all measures.

Customer multi-channel employment In order to assess the extent to which each respondent is a multi-channel shopper (that is, the more channels you visit, the more of a multi-channel shopper you are), we asked customers about the service outputs accessed from all mail order, retail, and Internet channels combined. If service outputs were accessed from three channels, the customer is a three-channel shopper, while if only one channel is accessed, the customer is a single-channel shopper, and so forth; thus, our measure of multi-channel employment is derived from the categories of service outputs reportedly accessed.

Multi-channel portfolio of service outputs We requested that each customer, in all three data collection modes, estimate the likelihood that each of the merchant’s mail order, Internet, and retail channels can deliver the four primary service outputs (information, product selection, delivery time, convenience). Similar to Fishbein’s (1963) expectancy value model, customers rated the ability of each of the industry partner’s channels to deliver each output; the outputs are then weighted in terms of their importance to the customer. The logic behind weighting service outputs by their importance is that available service outputs alone are not sufficient to sway customer responses. Service outputs that are available, but are of no interest to the shopper, are likely to be of little influence. The customer must both perceive the service output to be available and have some reason to think that the service output matters to him/her, in order to find it valuable. For example, a customer may perceive that the retail store can provide the quickest product delivery time; yet if the customer is purchasing a pair of overstock skis in July, delivery time is of very little importance. Four weighted values result (one each for convenience, waiting and delivery time, assortment, and information); these are then summed to create a single multi-attribute service output score. A high score indicated that all four-service outputs were important and were perceived to be readily available from all of the retailer’s channels. We used seven-point scales to measure availability (1 = Very Unlikely, 7 = Very Likely) and importance (1 = Not at all Important, 7 = Very Important). Disconfirmation and satisfaction Disconfirmation was operationalized using three items based on Oliver (1980), with end points reflecting the problems (1 = Much More Serious, 7 = Much Less Serious), benefits, and overall purchase experience (1 = Much Worse Than Expected, 7 = Much Better Than Expected). Customer satisfaction was measured using a two-item, seven-point scale (1 = Very Dissatisfied/Terrible and 7 = Very Satisfied/Delighted) (Oliver 1980; Westbrook 1980). The disconfirmation measure operationalizes the evaluation of the actual experience relative to a priori expectations, as opposed to the satisfaction measure which reflects the summary response resulting from the disconfirmation of expectations. Typically, a positive relationship should exist between disconfirmation and satisfaction; however, results to the contrary have been found (e.g., Spreng & Olshavsky 1993). Customer retailer loyalty This construct was measured to capture both behavioral and attitudinal elements (Dick & Basu 1994; Srinivasan et al. 2002). To tap the behavioral domain, we measured the number of the focal merchant’s competitors patronized by the customer using a checklist of such competitors provided by the industry partner. In general, greater switching behavior indicates lower loyalty to any one retailer. To measure the attitudinal element, customers were asked to compare the industry partner to its competition (1 = Much Worse Than Av-

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

erage, and 7 = Much Better Than Average). Since these are experienced customers, this measure captures a comparison with available alternatives that has developed over time.

Control variables We control for a number of variables that, while not the focus of this research, are likely to influence our variables: product type, depth of shopping experience, age, income, and purchase involvement. Customers are likely to engage in less information search and devote considerably less effort for convenience goods than for shopping and specialty goods (Darby & Karni 1973). Product type is controlled for by the research design, which only addresses shopping and specialty goods. Expectations based on shopping experience may significantly influence perceptions of and satisfaction with the current shopping experience (Churchill & Surprenant 1982; Spreng et al. 1996). It is also likely that customers with more shopping experience are more accustomed to the service outputs made available to them by the merchants that they encounter. For example, they are better at navigating web sites, reading size charts in catalogs, asking appropriate questions of retail sales people, and so forth. We employ a four-item measure that taps a customer’s experience at researching and purchasing similar products. We also control for age and income levels. These are related to socioeconomic status and are likely to influence purchase volume, as well as the type and quantity of service outputs accessed. Online shoppers, in particular, are thought to be from higher socioeconomic groups (Wolfinbarger & Gilly 2003). Age and income may also make for more demanding customers, thus lessening the impact of service output availability on satisfaction and loyalty. Purchase involvement, the personal relevance a purchase has for a customer, is also included as a control variable. Customers who are highly involved with the purchase will engage in greater effort processing information (Celsi & Olson 1988) and attend more closely to the services made available to them. We expect that highly involved customers will be more likely to notice and assess a high availability of service outputs, and thus treat it as a control variable that may confound

255

the influence of available service outputs on satisfaction. Our measure is based on a nine-item version of Zaichkowsky’s (1985) widely used scale, adapted as appropriate for purchase involvement (Lichtenstein, Bloch, & Black 1988).

Results The validity of the five reflective measures was assessed using confirmatory factor analysis. In the first CFA model, which included disconfirmation, multi-channel service outputs, and customer satisfaction, all loadings were significant and ranged from .57 to .95, well above Nunnally and Bernstein’s (1994) suggested cutoff of .4. The model was 2 significant (χ(24) = 53.59; p < .001) which is not surprising given this test’s known sensitivity to large sample sizes (Bollen 1989). However, alternative fit indices also suggest that the model fits the data reasonably well: the GFI, NFI and CFI statistics are at or above .98, and RMSEA = .05 (Bagozzi & Yi 1988). In addition, results show evidence of reasonable consistency (Fornell & Larcker 1981). Average variance extracted (AVE) ranged from .66 to .75, while reliabilities ranged from .79 to .92. Finally, discriminant validity can be inferred in that 95 percent confidence intervals around the construct Φs do not contain 1 (Anderson & Gerbing 1988). Thus, it can be concluded that, though related, disconfirmation and satisfaction are distinct and separate constructs. The second CFA model included the control variables (involve2 ment and experience) (χ(26) = 229; p = .000). The NFI, CFI, and GFI statistics all exceed .92; RMSEA is .11. AVE was .82 for involvement and .66 for experience; reliabilities were .96 and .89, respectively. Finally, 95 percent confidence intervals around the construct Φs once again do not contain 1 (Anderson & Gerbing 1988). Table 1 summarizes CFA results and Table 2 shows descriptive statistics. An examination of Fig. 1 suggests that the entire system should be estimated simultaneously; we thus test our hypotheses using LISREL. The initial structural model including all controls had a χ2 of 161.84; GFI was .94, CFI was .64, and NFI was .63, while RMSEA was .13. Having

Table 1 Measure validation Construct

Range of factor loadings

AVEa

Reliability (ρn )b

Satisfaction Disconfirmation Multi-channel portfolio of service outputs Goodness-of-fit statistics

.65–.95 .57–.88 .76–.92 NFI .98

.79 .81 .92

CFI .99

.66 .59 .75 GFI .98

Involvement Experience Goodness-of-fit statistics

.82–.96 .74–.89 NFI .94

CFI .94

.82 .66 GFI .92

a b

Average variance extracted calculated as per Fornell and Larcker (1981). Reliability calculated per Fornell and Larcker (1981).

χ2 (df, p) = 98 (38, .000)

RMSEA .046 .96 .89 RMSEA .11

χ2 (df, p) = 229 (26, .000)

256

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

Table 2 Correlation coefficients and descriptive statistics Construct

Number of channels

Service outputs

Service outputs Disconfirmation Customer satisfaction Loyalty Age Income Experience Involvement

.207** −.002 −.016 −.062 .027 −.013 .046 −.034

.246** .212** .026 −.068 −.038 .136** .067

Mean SD

1.75 .70

∗ ∗∗

330.76 112.32

Disconfirmation

.456** .253** .072 −.038 .129** .114*

Customer satisfaction

.204** .047 .028 .144** .079

16.55 3.07

12.49 1.81

Loyalty

Age

Income

Experience

.035 −.06 −.125** .095*

.456** .051 −.054

.116** −.092*

−.056

14.09 2.68

37.40 11.17

79694 43734

23.40 3.96

Involvement

44.54 11.16

Significant at the .05 level (two-tailed). Significant at the .01 level (two-tailed).

Table 3 LISREL results combined data Dependent variable

Independent variable

Service outputs

Controls Experience Involvement

.13 .07

3.15** 1.80*

Predictors Number of channels

.19

4.72**

Controls Age Experience Involvement

.10 .09 .10

2.38* 2.24* 2.40*

Predictors Service outputs

.22

5.56**

Controls Experience

.09

2.44*

Predictors Disconfirmation

.44

11.85**

Controls Experience

−.15

−3.81**

Predictors Satisfaction

.23

5.52**

Disconfirmation

Satisfaction

Loyalty

Standardized parameter estimate

t-Value

2 χ(17) = 41.96; p < .001; RMSEA = .05, GFI = .98; CFI = .92; NFI = .87. ∗ Significant at the .05 level. ∗∗ Significant at the .01 level.

ascertained that some control variables had no effect on our endogenous constructs, a more parsimonious model was estimated (see, e.g., Nijssen, Sing, Sirdeshmukh, & Holzmueller 2003). In the new model, χ2 was 41.96 and RMSEA was .05, while GFI, CFI and NFI were .98, .92, and .87, respectively. Table 3 presents results of this second model. H1 predicted that customer multi-channel employment positively influences the multi-channel portfolio of service outputs. H1 is supported (γ = .19; p < .01). H2 predicted that the multichannel portfolio of service outputs would positively influence disconfirmation. H2 is also supported (β = .22; p < .01).

H3 predicted that disconfirmation would influence customer satisfaction. Consistent with past findings, this prediction is supported (β = .44; p < .01). H4 predicted that customer satisfaction would influence customer retailer loyalty; this prediction is also supported (β = .23; p < .01). Turning to the control variables, age and income did not influence the multi-channel portfolio of service outputs, but experience (γ = .13; p < .01) and involvement (γ = .07; p < .05) did. Older customers (γ = .10; p < .05), more experienced customers (γ = .09; p < .05), and more involved customers (γ = .10; p < .05) were more likely to have their expectations positively disconfirmed, while income did not have a significant effect. Interestingly, the only control variable to influence overall satisfaction was experience (γ = .09; p < .05). Equally interesting is the negative influence of shopping experience on loyalty (γ = −.15; p < .01).1 Apparently, the more shopping customers engage in, the less loyal they are to any given merchant. These results suggest that multiple channel strategies, as depicted in Fig. 1, do contribute towards building customer retailer loyalty. To test H5 , we split the sample into two groups, single and multiple channel customers, based on their multiple channel usage scores; that is, we converted this measure into a dichotomous variable and used it to split the sample. There were 160 single-channel customers, and 264 multiple channel customers. Because a Chow test was necessary to test the poolability of these models across the two groups, we ran separate 3SLS models for each group; this analytical method simultaneously estimates a system of linear equations and permits the error terms of each equation to covary (Ramanathan 1998). Each 3SLS model included three equations. In the first equation, the dependent variable was dis1 Examination of our model suggests that it would be appropriate to test for whether satisfaction completely mediates the effect of disconfirmation on loyalty. We used three OLS regression models for this purpose (Baron & Kenny 1986). Results suggest that there is partial mediation. In addition to its indirect effect, disconfirmation has a direct effect as well (b = .21; p < .001). While the satisfaction paradigm does not specifically predict such an effect, apparently loyalty is directly influenced when retailers successfully meet customer expectations.

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

257

Table 4 3SLS regression results: single versus multiple channel customers Independent variable

Dependent variable

Single-channel customers

Multiple channel customers

Standardized parameter estimate Disconfirmation

Controls Age Income Experience Involvement Predictors Service outputs

Satisfaction

Controls Age Income Experience Involvement Predictors Disconfirmation

Loyalty

∗ ∗∗

t-Value

Standardized parameter estimate

t-Value

.10 −.17 .06 .12

1.09 −1.98* .71 1.56

.14 −.02 .22 .10

2.19* −.80 3.65** 1.71

.22

2.88*

.20

3.46**

−.11 .09 .05 .02

−1.21 .95 .66 .20

−.07 .15 .00 −.02

.51

2.18*

.84

Controls Age Income Experience Involvement

.34 −.05 −.26 −.03

.32 −.45 −2.50* −.35

−.07 −.03 −.23 .05

Predictors Satisfaction

.81

1.55

.83

−1.00 2.25* .02 −.25 4.62**

−.91 −.31 −2.58* −.70 3.22**

Significant at the .05 level. Significant at the .01 level.

confirmation, in the next it was satisfaction, and in the final equation, it was loyalty. The Chow test was significant for the disconfirmation equation (p = .002), and marginal for the loyalty equation (p = .15), suggesting that some differences do exist in how these models function for multiple channel and single-channel customers. H5 stated that H2 –H4 would be more strongly supported for multiple channel customers. There appears to be support for this assertion; R2 for the single-channel group is 8 percent, while R2 for the multiple channel group is 19 percent. Overall, the constructs do a better job of explaining the behavior of multiple rather than single-channel customers. This difference may be manifested in two ways: disconfirmation has a stronger influence on satisfaction for multiple channel customers (b = .84; p < .01) than for single-channel customers (b = .52; p < .05); in addition, satisfaction is a significant predictor of loyalty for multiple channel customers (b = .83; p < .01), while it is not for singlechannel customers (b = .81; p < .20). These results suggest that multiple channel retailing can be an effective means for building customer retailer loyalty. Table 4 presents these results. Our finding that satisfaction is a better predictor of loyalty for multiple channel customers than single-channel customers is consistent with recent empirical research. Shankar et al. (2003) found that the relationship between loyalty and satisfaction is higher online than offline. They argue that this may be because the online medium makes it easier for satisfied customers to choose the service provider again. We can

extend this argument to the single versus multiple channel context. A multiple channel customer perceives an increased ability to satisfy their complex needs via enhanced service outputs and more points of contact with a specific merchant. Multiple channel service outputs and multiple points of contact make it easier for a multiple channel customer’s satisfaction to manifest itself in the form of loyalty. This explanation assumes that an increased level of available service outputs is more likely to confirm customer expectations and, mediated by satisfaction, increase retailer loyalty. There is an intriguing relationship between customer expectations level and the resulting level of satisfaction. The most plausible explanation for our results is that raised multiple channel customer expectations actually lead to increased perceived performance (Yi 1990), which leads in turn to positive disconfirmation. On the other hand, high expectations would appear to be more difficult to meet, thereby raising the likelihood and magnitude of negative disconfirmation (Anderson 1973; Olshavsky & Miller 1972). Multiple channel shoppers might reasonably be expected to have higher expectations, thereby requiring that multiple channel performance needs to be correspondingly better in order to avoid negative disconfirmation. It might thus be argued that multiple channel shoppers could be expected to be less, not more, satisfied. Both arguments are evident in Spreng et al.’s (1996) findings of a negative direct effect of expectations on positive disconfirmation, as well as a positive indirect effect through perceived performance; interestingly, the positive indirect ef-

258

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

fect was slightly stronger than the negative direct effect. We conducted a follow-up study to examine these alternative perspectives. Follow-up study The follow-up study was designed to extend our results by examining negative market response and competitive effects. The basic contention in our main research study is that the use of multiple channels provides an opportunity to expose the customer to a larger and more diverse package of service outputs, and this has critical strategic implications in the form of enhanced customer satisfaction leading to customer loyalty with the retailer. A key challenge for a multiple channel retailer is the integration of its various channels in order to provide the customer a seamless shopping experience such that competitive switching pressures are lessened. Our major assumption throughout has been that the multiple channel retailing strategy has been implemented successfully. What happens when this is not the case? Cross-channel failures can occur in a number of ways. For example, a clothing item purchased online may not be returnable to a local store, or a product displayed in the store cannot be found in the catalog. In such cases, customer expectations are negatively, rather than positively disconfirmed, and the multiple channel strategy turns into a liability rather than a loyalty-enhancing asset. If cross-channel synergies can produce loyal customers, it may be that cross-channel failures can produce customers that are not loyal. Our key challenge in conducting this research was to identify customers of our industry partner who may have been the victims of cross-channel failures. We developed a mailing list of 1,000 “lost” customers; that is, randomly selected past customers of our industry partner who had not purchased from them in the last 36 months but who were still active purchasers with other firms in the same industry. Why had these active purchasers apparently dropped this retailer from their consideration sets? Have these apparently “lost” customers stopped patronizing the retailer due to cross-channel failures or something else? A survey was mailed to the home addresses of these 1,000 past customers. The packet included the survey, a stamped return envelope, and a cover letter with a response incentive. We received 140 responses; after accounting for 65 nondeliverable addresses, this represented a 15 percent response rate. Survey responses were matched to customer sales history in the industry partner’s database. There was not a significant difference in the number or dollar amount of life-to-date purchases between the mailing list and our 140 respondents, nor was there a significant difference between early and late respondents on either of these metrics, suggesting that the respondents were representative of the mailing list. The survey included the same measures for disconfirmation and overall satisfaction used in the main study data collection, along with measures of cross-channel service performance problems (e.g., I saw something in the catalog but

couldn’t find it on the web site) and inflated multiple channel expectations (e.g., I expected more product information). Since customers may cease to patronize a store for reasons not directly related to multiple channel retailing, questions related to general failings in regard to service, pricing, and merchandise were included (e.g., the hours were inconvenient). In addition, questions were asked concerning competition (e.g., other stores have a better selection). Results suggest that while there was no significant difference between “lost customers” and respondents from the main study data collection in level of disconfirmation (t(663) = .01); lost customers did express a lower level of overall satisfaction (t(694) = −3.8; p = .000). Given this finding, we estimated a regression model in which the lost customer’s overall satisfaction was a direct function of disconfirmation, cross-channel failures, multiple channel expectations, and competition. The regression model is significant (F(106) = 22.4; p < .001), with these independent variables explaining 47 percent of the variance in overall satisfaction. As expected, disconfirmation is the strongest predictor of satisfaction (b = .49; p < .001). Inflated expectations about multiple channel service outputs lead to reduced satisfaction (b = −.27; p < .01), as does strength of competition (b = −.18; p = .04). Problems caused by multiple channel issues are not a significant predictor of overall satisfaction (b = .10; p = .15). Interestingly, the most common reason cited by customers for their lack of recent patronage was that they had simply lost the habit of shopping there; others were price and lack of recent need.

Discussion The key research question focused on the market-level response to multiple channel retailing, particularly in the forms of customer satisfaction and retailer loyalty; that is, do retailers who invest in multiple channels receive a payoff in terms of customer loyalty? Our results emphatically suggest that multiple channel retailers do receive a customer loyalty payoff. Although market-level response is the ultimate critical measure of retailer effectiveness and profitability, the impact of multiple channel strategies on market-level responses has received much less attention by researchers (Homburg, Hoyer, & Fassnacht 2002) than other multiple channel issues, such as competitive response. Therefore, based on seminal service outputs and satisfaction/loyalty theories, the results of our research prompt researchers and managers to focus on these critical market-level responses to multiple channel strategies. According to our results, the retailer loyalty payoff occurs because: customers perceive an enhanced portfolio of service outputs provided by the multiple channels; customers’ complex needs are more likely to be satisfied with a synergistic combination of service outputs; and satisfaction based on multiple channel strategies results in customers having loyalty to the provider of the multiple channels, the retailer. Im-

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

portantly, our results provide a preliminary indication that the empirical model linking multiple channel strategies to customer retailer loyalty was more effective for multiple channel customers than for those customers employing only a single channel. The results of our follow-up investigation imply that, not surprisingly, these lost customers are less satisfied overall than those surveyed in the main data collection. Interestingly, dissatisfaction does not appear to result from cross-channel failures related to the complexity inherent in multiple channel strategies; but rather to a failure to deliver the overall level of service outputs customers expect of a multiple channel retailer (inflated expectations issue), failure of a single channel to deliver a specific service output, or competitive reasons to switch to another retailer. Price is a generally important competitive issue in satisfying customers. These follow-up results, paired with our main study results, support the conclusion that multiple channel retailing is an important means of enhancing customer satisfaction and retailer loyalty, provided that retailers make available and deliver general levels of service outputs commensurate with multiple channel customers’ high expectations. On the other hand, the strategy of multiple channel retailing does not appear to diminish customer satisfaction; that is, when multiple channel customers are dissatisfied and ultimately switch to another retailer, the causes focus on single-channel failures (as opposed to multiple channel interaction failures) and competitive advantages, particularly price advantages. These results suggest that multiple channel strategies result in more satisfied and loyal multiple channel customers; these results do not suggest that the unique complexities of multiple channel interactions result in more dissatisfied multiple channel customers.

Future research Future research is needed to address the unique role of specific service outputs in different channels and how these unique roles impact satisfaction and retailer loyalty. It is accepted that different channels provide different mixes of service outputs and that customers may access different channels depending on their current needs (Friedman & Furey, 2003; Hansell 2002); most importantly, future research needs to address the efficacy of these different mixes, both alone and in combination. For example, how does customer response to live online help compare with response to 1–800 support or in-person selling? What do customers get out of in-store sources of information such as point-of-purchase materials and employees that they might not get from a web site or catalog? When retailers transition from single to multiple channel strategies, they thereby gain the ability to provide these different mixes of services to their customers. We are not necessarily advocating that all retailers should pursue multiple channel strategies. It is an interesting and important question to examine under what product market conditions, and

259

for what types of retailers, this transition to multiple channel strategies is particularly efficacious. In addition, future research should examine the market response to specific service outputs provided by combined channels, such as online ordering with retail pickup or in-store web kiosks. Combining channels to provide service outputs is emerging as a merchandising trend. The question remains whether such a multiple channel strategy creates strategically important market-level responses. Friedman and Furey (2003) introduce an interesting distinction between channel mix and channel integration strategies. Empirical research needs to be done on this important distinction and on other multiple channel strategies that may be available to managers. In particular, what is the efficacy of each strategy, and what are their managerial challenges and implications? Multiple channel retailing introduces a host of competitive issues. For example, competition among a retailer’s different channels would appear to be detrimental to the synergistic combination of services required of successful multiple channel retailing. What managerial strategies are effective at curbing such competition? The purpose of this research has been to examine market response to an existing multiple channel retailing strategy. More research needs to be done on the reasons why such strategies are implemented in the first place. Competitive reasons surely loom large here, as indeed they do for any marketing initiative. What are the implications for competitive response when retailers announce or implement multiple channel strategies? Finally, competition among multiple channel retailers and the risk of free-riding on services may make it difficult to profitably support products with the kinds of services advocated here. What forms of compensation and control mechanisms are needed on the part of suppliers and wholesalers to align the interests and objectives of different retailers with their upstream suppliers? Although we used cross-sectional data from a customer group with considerable experience shopping from the retailer, we did not track these customers across several purchase episodes; instead loyalty is based on historical sales amounts and number of past purchase episodes rather than true longitudinal data. A longitudinal study, however, would provide valuable insight into other factors that enhance satisfaction and retailer loyalty, as well as insight into what factors diminish satisfaction resulting in “lost” customers and what factors cause customers to switch retailers even though satisfaction may be unchanged. Future research should also explore multiple channel strategic implications of customers who purchase from one channel on one occasion and another channel on another occasion and customers who use different channels for different purposes. Our research is designed to assess the impact of multiple channel shopping on customer satisfaction and, ultimately, customer retailer loyalty, for the customers of a single retailer. With this design, there is variance in the number of channels used by the customer but no variance in the number of channels offered by the retailer. Also, with this de-

260

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

sign, there is less variance in customer response because of the main study’s focus on only current customers. We contend that there is little reason to expect that the multiple channel customers of this retailer would be different in any systematic way from the multiple channel customers of other retailers, and that this general framework—customers who visit multiple channels are thereby exposed to more service outputs and this has beneficial effects on satisfaction and loyalty—should apply in other settings. The follow-up study was designed to investigate market-level responses of “lost” customers. Nonetheless, future research should compare the responses of customers to single versus multiple channel retailers. In addition, future research should further expand on differences between single versus multiple channel customers. Indeed, fertile future research should track the development of loyalty, utilizing current customers and noncustomers, over multiple purchase instances and multiple channels, as well as, across multiple retailers that employ different multiple channel strategies.

Conclusion It has been observed that with the advent of Internet retailing, the new “frictionless market” may lead to increased price competition (Brynjolfsson & Smith 2000). Retailers that wish to compete successfully must find some way of providing differential value. This research, based on actual industry data, suggests one important means to this end. By expanding the portfolio of service outputs available to the

customer, multiple channel strategies may be employed to enhance the retailer’s overall value proposition (Porter 2001), leading to enhanced customer satisfaction and managerially relevant outcomes. This research has investigated two strategically critical yet under-researched areas: multiple channel retailing strategies and customer retailer loyalty. Multiple channel distribution strategies have become increasingly important in industry (Reda 2002b) and in academic research (Frazier 1999), and customer retailer loyalty has become similarly consequential. The major contribution of this research is to apply classic channels and satisfaction theory to these domains, and to show that retailers who pursue multiple channel strategies enhance the service outputs the customers find available, thereby satisfying their customers; this satisfaction enhances the strategically important outcome of customer retailer loyalty. In addition, we specifically model this process for two different groups, single and multiple channel customers, and show that it can be particularly effective for the latter. Thus, we conclude that multiple channel strategies, in and of themselves, are useful in satisfying multiple channel customers’ high expectations and retaining customers.

Acknowledgements The authors thank our industry partner for their generous involvement in this project. The authors also thank the Department of Marketing at Washington State University for financial support.

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

Appendix A. Construct operationalizations

261

262

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263

References Anderson, Eugene W., Fornell, Claes, & Lehmann, Donald R. (1994). Customer satisfaction, market share, and profitability: Findings from Sweden. Journal of Marketing, 58(July), 53–66. Anderson, James C., & Gerbing, David W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), 411–423. Anderson, Rolph E. (1973). Consumer satisfaction: The effect of disconfirmed expectancy on perceived product performance. Journal of Market Research, 10(February), 38–44. Armstrong, J. Scott, & Overton, Terry S. (1977). Estimating nonresponse bias in mail surveys. Journal of Marketing Research, 14(August), 396–402. Bagozzi, Richard P., & Yi, Youjae. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(Spring), 74–94. Bollen, Kenneth A. (1989). Structural equations with latent variables. New York: John Wiley and Sons. Brynjolfsson, E., & Smith, M. (2000). Frictionless commerce? Comparison of Internet and conventional retailers. Management Science, 46(April), 563–585. Bucklin, Louis P. (1966). A theory of distribution channel structure. Berkeley: IBER Special Publications. Bucklin, Louis P., Ramaswamy, Venkatram, & Majumdar, Sumit K. (1996). Analyzing channel structures of business markets via the structure-output paradigm. International Journal of Research in Marketing, 13, 73–87. Celsi, Richard L., & Olson, Jerry C. (1988). The role of involvement in attention and comprehension processes. Journal of Consumer Research, 15(September), 210–224. Churchill, Gilbert A. (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 16(February), 64–73. Churchill, Gilbert A., & Surprenant, Carol. (1982). An investigation into the determinants of customer satisfaction. Journal of Marketing Research, 491–504. Darby, Michael R., & Karni, Edi. (1973). Free competition and the optimal amount of fraud. Journal of Law and Economics, 16(April), 66–86. Day, George S. (1969). A two-dimensional concept of brand loyalty. Journal of Advertising Research, 9(3), 29–36. Day, George S., & Nedungadi, Prakash. (1994). Managerial representations of competitive advantage. Journal of Marketing, 58(April), 31–44. Dick, Alan S., & Basu, Kunal. (1994). Customer loyalty: toward an integrated conceptual framework. Journal of the Academy of Marketing Science, 22(Spring), 99–113. DiMaggio, Paul J., & Powell, Walter J. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48, 147–160. Dutta, Shantanu, Bergen, Mark, Heide, Jan, & John, George. (1995). Understanding dual distribution: The case of reps and house accounts. Journal of Law, Economics, and Organization, 11(1), 189– 204. Eagly, Alice H., & Chaiken, Shelly. (1998). Attitude structure and function. In T. Daniel, Gilbert, T. Susan, Fiske, & Gardner Lindzey (Eds.), The handbook of social psychology. Boston, MA: McGraw-Hill. Fishbein, M. (1963). An investigation of the relationship between beliefs about an object and the attitude toward that object. Human Relations, 16, 233–240. Fornell, Claes. (1992). A national customer satisfaction barometer: The Swedish experience. Journal of Marketing, 56(January), 6–22. Fornell, Claes, & Larcker, David F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 28(August), 382–388.

Frazier, Gary L. (1999). Organizing and managing channels of distribution. Journal of the Academy of Marketing Science, 27(2), 226– 240. Frazier, Gary L., & Shervani, Tasadduq A. (1992). Multiple channels of distribution and their impact on retailing. In R. A. Peterson (Ed.), The future of U.S. retailing: An agenda for the 21st century (pp. 217–238). New York: Quorum Books. Friedman, Lawrence G., & Furey, Timothy R. (2003). The channel advantage. Burlington, MA: Butterworth-Heinemann. Giese, Joan L., & Cote, A. Joseph. (2000). Defining consumer satisfaction, Academy of Marketing Science Review, http://www.amsreview.org/amsrev/theory/giese01-00.html. G¨orsch, Daniel. (2000). The impact of hybrid channel structures on the customer purchase process: A research outline. In Proceedings of IRIS 123 Laboratorium for Interaction Technology. Grewal, Dhruv, Iyer, Gopalkrishnan R., & Levy, Michael. (2004). Internet retailing: Enablers, limiters and market consequences. Journal of Business Research, 57(7), 703–713. Grewal, Rajdeep, Comer, James M., & Mehta, Raj. (2001). An investigation into the antecedents of organizational participation in busninessto-business electronic markets. Journal of Marketing, 65(July), 17–33. Hansell, Saul. (2002, December 14). A retailing mix: on internet, in print and in store. The New York Times. Homburg, Christian, Hoyer, Wayne D., & Fassnacht, Martin. (2002). Service orientation of a retailer’s business strategy: Dimensions, antecedents, and performance outcomes. Journal of Marketing, 66(October), 86–101. Howard, J. A., & Sheth, Jagdish N. (1969). The theory of buyer behavior. New York: John Wiley & Sons Inc. Jacoby, Jacob, & Chestnut, Robert W. (1978). Brand loyalty measurement and management. New York: John Wiley & Sons. Lemon, Katherine N., White, Tiffany Barnett, & Winer, Russell. (2002). Dynamic customer relationship management: incorporating future considerations into the service retention decision. Journal of Marketing, 66(January), 1–14. Lichtenstein, Donald R., Bloch, Peter H., & Black, William C. (1988). Correlates of price acceptability. Journal of Consumer Research, 15(September), 243–252. Nijssen, Edwin, Singh, Jagdip, Sirdeshmukh, Deepak, & Holzmueller, Hartmut. (2003). Investigating industry context effects in consumerfirm relationships: Preliminary results from a dispositional approach. Journal of the Academy of Marketing Science, 31(Winter), 46–60. Nunnally, Jum C., & Bernstein, Ira H. (1994). Psychometric theory. New York, NY: McGraw-Hill. Oliver, Richard. (1980). A cognitive model of the antecedents and consequences of satisfaction decisions. Journal of Marketing Research, 22(November), 460–469. Oliver, Richard. (1997). Satisfaction: A behavioral perspective on the consumer. New York: McGraw-Hill. Oliver, Richard. (1999). Whence consumer loyalty? Journal of Marketing, 63(Special issue 1999), 33–44. Olshavsky, Richard W., & Miller, John A. (1972). Consumer expectations, product performance, and perceived product quality. Journal of Marketing Research, 9(February), 19–21. Outdoor Industry Association. (2002). State of the industry report, http://www.outdoorindustry.org/SOI2002.pdf. Pastore, Michael. (2001). Multi-channel customers key to retail success, http://www.ecrmguide.com/columns/print/0,10383 902831,00.html. Peterson, Robert A., & Balasubramanian, Sridhar. (2002). Retailing in the 21st century: Reflections and prologue to research. Journal of Retailing, 78, 9–16. Porter, Michael. (2001, March). Strategy and the Internet. Harvard Business Review, 63–78. Ramanathan, Ramu. (1998). Introductory econometrics with applications. Fort Worth, TX: Harcourt Brace. Reda, Susan. (2002a). Active multi-channel shoppers may be a liability, less loyal than other on-line shoppers. Stores, 84(September), 78–82.

D.W. Wallace et al. / Journal of Retailing 80 (2004) 249–263 Reda, Susan. (2002b). The multi-channel CEO: What does it take to succeed? Stores, 84(October), 26–32. Reichheld, Frederick. (1993). Loyalty based management. Harvard Business Review, 71(March–April), 64–73. Reinartz, Werner, Manfred, Krafft, & Wayne, D. Hoyer. (2004). The customer relationship management construct: its measurement and impact on performance. Journal of Marketing Research, 41(August), 293–305. Saunders, Christopher. (2002). Multiple channel buyers worth pursuing. Internet Advertising Report, January 28, http://www.internetnews.com. Shankar, Venkatesh, Smith, Amy, & Rangaswamy, Arvind. (2003). Customer satisfaction and loyalty in online and offline environments. International Journal of Research in Marketing, 20(June), 153– 175. Spreng, Richard A., MacKenzie, Scott B., & Olshavsky, Richard W. (1996). A Reexamination of the determinants of consumer satisfaction. Journal of Marketing, 60(July), 15–32. Spreng, Richard A., & Olshavsky, Richard W. (1993). A desires congruency model of consumer satisfaction. Journal of the Academy of Marketing Science, 21(Summer), 169–177. Srinivasan, Srini S., Anderson, Rolph, & Ponnavolu, Kishore. (2002). Customer loyalty in E-commerce: An exploration of its antecedents and consequences. Journal of Retailing, 78, 41–50.

263

Stern, Louis W., & El-Ansary, Adel I. (1992). Marketing channels. Englewood Cliffs: Prentice-Hall. Tang, Fang-Fang, & Xing, Xiaolin. (2001). Will the growth of multichannel retailing diminish the pricing efficiency of the web? Journal of Retailing, 77, 319–333. Westbrook, Robert A. (1980). A rating scale for measuring product/service satisfaction. Journal of Marketing, 44(Fall), 68–72. Wind, Y., & Mahajan, V. (2002). Convergence marketing. Journal of Interactive Marketing, 16, 64–74. Wolfinbarger, Mary, & Gilly, Mary C. (2003). eTailQ: Dimensionalizing, measuring and predicting retail quality. Journal of Retailing, 79, 183–198. Wright, Claire, & Sparks, Leigh. (1999). Loyalty saturation in retailing: Exploring the end of retail loyalty cards? International Journal of Retail and Distribution Management, 27(October), 429–440. Yi, Youjae. (1990). A critical review of consumer satisfaction. In Valarie A. Zeithaml (Ed.), Review of marketing (pp. 68–123). Chicago, IL: American Marketing Association. Zaichkowsky, Judith Lynne. (1985). Measuring the involvement construct. Journal of Consumer Research, 12(December), 341–352. Zeithaml, Valerie A., Berry, Leonard L., & Parasuraman, A. (1996). The behavioral consequences of service quality. Journal of Marketing, 60(April), 31–46.