Online consumer retention: contingent effects of online shopping habit and online shopping experience

Transcription

1 (2007) 16, & 2007 Operational Research Society Ltd. All rights reserved X/07 $ consumer retention: contingent effects of online shopping habit and online shopping experience Mohamed Khalifa 1 and Vanessa Liu 2 1 Department of Information Systems, City University of Hong Kong, Kowloon Tong, Hong Kong SAR, China; 2 School of Management and Department of Information Systems, New Jersey Institute of Technology, University Heights, Newark, NJ, U.S.A. Correspondence: Mohamed Khalifa, Department of Information Systems, City University of Hong Kong, Tat Chee Avenue, Kowloon Tong, Hong Kong SAR, China. Tel: (852) ; iskhal@cityu.edu.hk Abstract In this study, we further develop the information systems continuance model in the context of online shopping, using a contingency theory that accounts for the roles of online shopping habit and online shopping experience. Specifically, we argue and empirically demonstrate that although conceptually distinct, online shopping habit and online shopping experience have similar effects on repurchase intention. They both have positive mediated effects through satisfaction and moderate the relationship between satisfaction and online repurchase intention. The results of a survey study involving 122 online customers provide strong support for our research model. We also identify after-sale service, transaction efficiency, security, convenience, and cost savings as important online shopping usefulness drivers. Theoretical and practical implications include establishing a contingency theory to more fully explain online customer retention as well as guidelines for development of customer relationship management initiatives. (2007) 16, doi: /palgrave.ejis Keywords: online customer retention; online shopping habit; online shopping experience Received: 28 January 2005 Revised: 18 August nd Revision: 12 June rd Revision: 6 January th Revision: 14 June 2007 Accepted: 20 September 2007 Introduction Customer retention is considered by both scholars and practitioners to be one of the critical success factors for retail businesses with its implications for cost savings and profitability (Doyle, 2003). The cost of acquiring new customers is five to seven times that of retaining existing ones. Furthermore, retained customers enhance profitability with their lower sensitivity to price changes and their higher likelihood of referring new customers (Doyle, 2003). Customer retention is an even more challenging issue in the context of online shopping, where severe competitors exist and the switching costs for customers are minimal (Anderson & Srinivasan, 2003). It is therefore important to identify the major determinants of online customer retention. One such important factor identified in previous research is satisfaction. Its direct relationship with retention appears intuitive (Anderson & Srinivasan, 2003). Indeed, the marketing literature confirms that customer satisfaction is one of the main drivers of repurchase as verified in various different industrial and social contexts (e.g., see Rust & Zahorik, 1993; Rust et al., 1995; Hallowell, 1996). When a customer is satisfied with a particular internet store, he or she is more likely to shop there again. In a study

2 consumer retention Mohamed Khalifa and Vanessa Liu 781 (Jones & Suh, 2000), satisfaction is found to be a strong single predictor of repurchase behaviour, sometimes explaining over 50%. Most of the prior studies, however, were conducted in the traditional shopping context. It is not clear whether their findings can be applied to the online shopping setting. Some information systems (IS) research (e.g., Bhattacherjee, 2001a, b) adopted the marketing approach in explaining system usage continuance, suggesting perceived usefulness to be another factor affecting continuance in addition to satisfaction. A customer is more likely to buy again from an internet store when he or she finds shopping there useful. Well established in the traditional IS paradigm, perceived usefulness is a central notion that explains the usage or intended usage of information technologies (Davis et al., 1989). Drawing on the expectation disconfirmation theory, Bhattacherjee (2001a) modelled IS continuance intention as a direct outcome of both satisfaction and perceived usefulness. However, the adequacy of explaining repurchase intention using satisfaction alone is questionable (Capraro et al., 2003). It has been reported that only 15 35% of satisfied customers did return (Reichheld, 1996). The strength of the effect of satisfaction on retention also varies, subject to factors like nature of industry (Jones & Sasser, 1995) and other social factors (Oliver, 1999). It is therefore important to examine the role of potential moderators in attaining a better understanding of the relationship between satisfaction and repurchase intention in the online context (Oliver, 1999; Capraro et al., 2003; Nijssen et al., 2003). Of particular interest to this study is the role of online shopping habit, which refers to the habit of shopping on the internet in general. Habit is not a new phenomenon to the marketing literature and has been examined in the traditional retailing context. Several empirical studies (e.g., Murray & Haubl, 2005) show that retention can be achieved indirectly when a habit exists. Habit is considered as an important factor in explaining repeated purchases (Quinn & Wood, 2005). Nevertheless, most previous research on habit focus on the temporal dimension of the construct only (i.e., the frequency of the behaviour) with little consideration to the effect of the context in which a habit is practised (Khare & Inman, 2005). The online channel represents an innovative shopping context with a number of unique attributes, for example, interactivity, flexibility of navigation, etc. (Childers et al., 2001). The novelty of online shopping implies that some of the effects of the determinants of online repurchase may be contingent upon the development of the habit of using the online channel. Satisfaction, for example, may not necessarily lead to intention to return to an internet store in the absence of an online shopping habit. An individual may not repurchase from a cyber merchant when he or she has not formed a habit of online shopping despite his/her satisfaction with previous transactions. An important implication of this contingency is that the retention theories developed and validated in the physical shopping context need further enhancement to account for the effect of online shopping habit. Indeed, some researchers suggested that a weak habit may lower the sensitivity of consumer retention to satisfaction (Anderson & Srinivasan, 2003). The habit construct is also of appeal to practitioners. According to a recent report, online retailers consider the development of online shopping habit to have a major impact on internet sales (Cisco Systems, 2003). It is therefore important to examine the role of habit in the online shopping context in general and its effect on online repurchase intention in particular. In addition, the research on habit often involves the use of prior experience as its surrogate measure (e.g., East et al., 1994). Instead of examining the habit of doing general shopping online, a number of studies looked into broad experience with internet shopping. Prior experience is indeed one of the key ingredients for habit formation (Triandis, 1971). Some scholars, however, argue that experience is conceptually distinct from habit and hence its validity as a proxy for habit becomes doubtful (e.g., Ajzen, 1991). To throw light on this ambiguity, we study prior online shopping experience in this research and compare its effects with those of online shopping habit. In this study, we integrate the marketing and IS theories to develop a better model for explaining the repurchase intention of online customers. Specifically, we adopt a contingency approach including online shopping habit as the moderator of the relationship between satisfaction and repurchase intention. Furthermore, we clarify the distinction between habit and prior experience by performing a comparative analysis. Our research model therefore consists of two separate levels of research variables. Repurchase intention and its hypothesised predictors (satisfaction and perceived usefulness) are internet retailer-specific while the moderators (online shopping habit and online shopping experience) are related to the online shopping context in general. By refining the existing online customer retention models, our paper also responds to Pateli & Giaglis (2004) s call for the development of more structured theoretical tools in the electronic business literature. In the next section, we introduce and present the theoretical foundation of the research model. We then describe the research context and methodology, followed by a discussion of the empirical results. We conclude the paper by discussing the practical implications of the findings and suggesting directions for future research. The research model With the growing maturity of the customer retention literature, this study extends early definitional research and develops a research model with a focus on more detailed ontological analyses of the relationships among the constructs underlying the customer retention phenomenon (Pateli & Giaglis, 2004). The theoretical development of our research model is grounded in the IS

3 782 consumer retention Mohamed Khalifa and Vanessa Liu Perceived Usefulness Shopping Satisfaction Figure The research model. + Shopping Habit/ Experience Repurchase Intention continuance theory (Bhattacherjee, 2001a). In justifying the integration of the technology acceptance model (TAM) (Davis et al., 1989) with the disconfirmation theory (Oliver & DeSarbo, 1988), Bhattacherjee (2001a) argued that IS continuance decisions are very similar to consumers repurchase intention. He conducted his study in the context of online banking systems. In his model, continuance intention is directly driven by satisfaction and perceived usefulness. Following the marketing approach, Bhattacherjee (2001a) measured satisfaction as an evaluative outcome of a comparison process between expected and perceived performance known as confirmation (Oliver & DeSarbo, 1988). Bhattacherjee s (2001a) model is based on theories originally developed and validated in physical environments, with which the customers are quite familiar. We argue that the application of these theories to the online environment is contingent upon the formation of online shopping habit. Our research model (see Figure 1) extends Bhattacherjee s (2001a) study to capture the mediated and moderating effects of online shopping habit on the formation of repurchase intention. 1 Repurchase intention, perceived usefulness, and satisfaction Repurchase intention has been studied extensively by marketing scholars. It is typically defined as the intention to repeatedly purchase a particular product or service over time (see Copeland, 1923). Repurchase mirrors and constitutes one important dimension of loyalty (Jacoby & Chestnut, 1978; Soderlund et al., 2001). As the issues of conversion (converting customers to the online channel; Ho & Wu, 1999) and stickiness (continuation of using the online channel; Khalifa et al., 2002) are of particular interest to internet retailers, the definition of repurchase intention originated from marketing may not be in the best fit. In view of this, we define repurchase as the re-usage of the online channel to buy from a particular retailer. Unlike other metrics such as click through ratios, repurchase can better capture the process of retaining 1 We exclude the confirmation construct as its effects are fully mediated through satisfaction (Oliver & DeSarbo, 1988). Its inclusion would not add to the explanation of the focal variable in this study, i.e., online repurchase intention. + + prior customers with repeat business (Nemzow, 1999). As such, repurchase can be considered as continuance behaviour (i.e., continue to shop from the same internet store). The usage of a continuance model to explain this behaviour is therefore appropriate. Consistent with most prior IS and marketing studies, our focus is continuance intention rather than the behaviour itself. Intention is considered to be strongly correlated with volitional behaviour (i.e., a behaviour that a person can decide at will to perform or not perform, Ajzen, 1991). Volitional behaviours typically involve a number of alternatives among which a choice can be made at discretion. Intention can accurately predict behaviours that are generally under one s control (see Ajzen, 1988). The strong association between intention and volitional behaviour is validated empirically (Davis et al., 1989; Taylor & Todd, 1995). Repurchase is a form of volitional behaviour, as customers generally enjoy more than one option of which store to buy again. Additional theoretical justification for this association comes from the cognitive dissonance theory (Festinger, 1957), which stipulates that individuals attempt to relieve the psychological tension due to discrepancies between intention and behaviour by being consistent with their intention. There is also a practical consideration for focusing on intention. In the context of continuance, it is difficult to tell whether an individual who has not repurchased for a certain period of time has indeed discontinued. According to the TAM, perceived usefulness is one of the major determinants of intention formation. The notion was originally developed in the organisational setting, denoting the extent to which an individual s job performance is enhanced with the use of a specific technology (Davis et al., 1989). In the context of online shopping, it refers to the salient beliefs of customers regarding the instrumentality of repurchase (Davis et al., 1989). For example, online shopping may be perceived as more useful when it confers convenience and cost savings. The original operationalisation of perceived usefulness (Davis et al., 1989) involved scales that are generalisable across both applications and contexts. The initial usefulness measures were developed with respect to the use of several office applications, namely, an system, a graphics system, and a file editor application. The measures include quality of work, control over work, work more quickly, critical to my job, increase productivity, job performance, accomplish more work, effectiveness, makes job easier, and useful (Davis et al., 1989, p. 329, Table 6). Ajzen, 1991 subsequently argues that researchers should consider salient beliefs that are specific to the context, as it is only at the level of specific beliefs that we can learn about the unique factors that induce one person to engage in the behaviour of interest. He further contends that these salient beliefs must be elicited from the respondents themselves, or in pilot work from a sample of respondents that is representative of the research population. Some examples of usefulness factors identified in the

4 consumer retention Mohamed Khalifa and Vanessa Liu 783 e-commerce literature include faster search, improved search and buying, increased shopping productivity, enhanced shopping performance (e.g., see Gefen & Straub, 2000; Koufaris, 2002). Accordingly, we have conducted a belief elicitation process (described later in the Methodology section) involving a sample of our respondents (i.e., online shoppers) to identify salient usefulness beliefs specific to the online shopping context and the internet technology, resulting in five usefulness factors (i.e. after-sales service, transaction efficiency, security, convenience, and cost savings). The effect of perceived usefulness on behavioural intention (such as repurchase) has been consistently reported in most empirical studies in various contexts, including information technology and electronic commerce (see a summary in Gefen & Straub, 2000). In this accord, Bhattacherjee (2001a) also suggested and empirically demonstrated that an individual is more likely to intend for continued usage when such usage is perceived to be useful. Applying TAM in the context of online shopping, individuals will be more likely to intend to repurchase online when they believe such behaviour to be able to confer positive values (Davis et al., 1989). For example, one may be more willing to shop again (i.e., repurchase intention) from an online store if one perceives it to be cheaper than other physical or online stores or providing better after-sales services. Consistent with prior studies, we hypothesise that: H1: Perceived Usefulness has a significant positive effect on Repurchase Intention. Satisfaction is another major determinant of repurchase intention. It has been widely defined as a postevaluative judgement over a particular purchase (Oliver, 1979, 1980; Churchill & Suprenant, 1982; Bearden & Teel, 1983; Oliver & DeSarbo, 1988). It is an attitudinal variable, and as such is an important determinant of intention as per TAM (Davis et al., 1989). We rely on the transaction cycle to characterise the customer s satisfaction with his/her online shopping experience. Consistent with Feinberg et al. (2002), Khalifa et al. (2002), and Lu (2003), our research model identifies three types of satisfaction (i.e., pre-purchase, at-purchase, and post-purchase satisfaction). The pre-purchase satisfaction is related to the customer s experience with online activities performed prior to placing an order, for example, consumer education, product search, and product comparison. The purchase satisfaction is formed based on the customer s experience with the order placement and payment processes. The post-purchase satisfaction, on the other hand, is associated with the customer s experience with after-sale services, for example, online customer support, helpdesks handling returns/refunds, and online installation manuals, etc. Judgement of satisfaction is entirely subjective and individual (Oliver, 1997). Perceived usefulness represents salient beliefs that are developed based on realistic consideration of actual situational factors (Davis et al., 1989). Previous TAM studies found that perceived usefulness significantly, consistently, and substantively drives the formation of user acceptance (e.g., Davis et al., 1989; Mathieson, 1991; Taylor & Todd, 1995; Karahanna et al., 1999) and attitudes in general, such as satisfaction. Similarly, the disconfirmation theory posits that satisfaction is an evaluative outcome of the gap between performance and expectations (Oliver & DeSarbo, 1988), with higher performance levels associated with more positive disconfirmation and increased satisfaction. Perceived usefulness refers to beliefs regarding performance (e.g., time savings, cost savings). An individual is more likely to be satisfied with an offering or experience when he or she perceives better performance therefore greater usefulness. We accordingly hypothesise that: H2: Perceived Usefulness has a significant positive effect on Shopping Satisfaction. The direct relationship between satisfaction and continuance intention is at the core of the IS continuance model and is validated empirically (Bhattacherjee, 2001a). In the online shopping context, a customer is more likely to intend to return to a particular store if he or she is satisfied with the previous purchases from that store. Unpleasant buying experience easily leads to dissatisfaction, discouraging the customer from coming back (Oliver, 1999). There is ample empirical evidence from the marketing literature that customer satisfaction is positively linked to repurchase intention (e.g., see Rust et al., 1995; Hallowell, 1996; Anderson & Srinivasan, 2003). We therefore hypothesise that: H3: Shopping Satisfaction has a significant positive effect on Repurchase Intention. shopping habit The effects of online shopping habit on online repurchase intention are two-fold: (1) mediated through satisfaction and (2) moderating the relationship between satisfaction and repurchase intention. Habit refers to situation-behaviour sequences that are or have become automaticy the individual is usually not conscious of these sequences (Triandis, 1980, p. 204). It is a behavioural tendency resulting from prior experience (Gefen, 2003). It represents behavioural disposition to repeat previous action, developed through frequent performance in a stable context (Ouellete & Wood, 1998). It can be viewed as an automatic behavioural response triggered by a situational stimulus without being preceded by a cognitive analysis process (Aarts et al., 1998). The individual is not necessarily aware of the behaviour nor is required to devote thought or rational evaluations prior to engaging in such behaviour (Ouellete & Wood, 1998). In other words, habitual behaviours

5 784 consumer retention Mohamed Khalifa and Vanessa Liu require minimal conscious thoughts (Aarts & Dijksterhuis, 2000), enabling individuals to conserve their limited mental resources (Bargh & Ferguson, 2000). Such automaticity allows habitual behaviours to be performed easily and quickly (Quinn & Wood, 2005). In the context of online shopping, individuals with online shopping habits would turn to the online channel (i.e., the behavioural response) instead of a physical outlet automatically without further consideration when they feel shopping needs (i.e., situational stimulus). Past research reported that habit is a major driver of affect (Limayem & Hirt, 2003). Affect is defined as feelings of joy, elation, or pleasure, or depression, disgust, displeasure, or hate associated by an individual with a particular act (Triandis, 1971, p. 211) and an emotional response to the thought of the behaviour (Limayem & Hirt, 2003, p. 69). Consistent with Bhattacherjee (2001a), we argue that satisfaction is also an affect. By giving rise to a favourable feeling towards a behaviour (Triandis, 1971), habit affects satisfaction directly (Limayem & Hirt, 2003). In other words, a customer is likely to be more satisfied with an internet store if he or she has acquired the habit of shopping online. We therefore hypothesise that: H4: Shopping Habit has a significant positive effect on Shopping Satisfaction. In addition to its mediated effect, habit is also supposed to positively moderate the relationship between satisfaction and repurchase intention. According to the theory of attitude and attitude change, behavioural intention is determined by the interaction of habit and affect among other constructs (i.e., social norms and attitude) (Triandis, 1971). With the same level of affect (i.e., satisfaction), an individual who has acquired a habit of a behaviour is more likely to intend to repeat the behaviour in the future than others without such habit. In other words, given particular level of satisfaction with an online store, a customer who has acquired a stronger online shopping habit is more likely to intend to repurchase from that store. A customer who rarely shops online may not return despite that he or she is satisfied. Accordingly, we hypothesise that: H5: Shopping Habit positively moderates the relationship between Shopping Satisfaction and Repurchase Intention. 1996). Other researchers, however, have argued that such operationalisation of habit is not optimal (e.g., Ajzen, 1991; Eagly & Chaiken, 1993). Although experience is a precursor of habit (Triandis, 1971; Chaudhuri, 1999), the two constructs are conceptually different. Unlike experience, the formation of a habit is partly subject to the ability of the individual to convert/absorb the behaviour into cognitive schemata (Limayem et al., 2001). The development of experience, on the other hand, solely requires actual usage behaviour. A prolonged usage experience does not necessarily imply that an individual has acquired the usage habit. We therefore develop two separate measurement scales for habit and experience, hypothesising the conceptual discrimination between the two constructs. Although conceptually different, experience and habit seem to have similar effects. This may explain the usage of experience as surrogate of habit by some researchers. In the context of our study, experience is expected to have positive effects on satisfaction and moderating effects on the relationship between satisfaction and repurchase intention. Prior studies showed that repeated behaviour (i.e., experience) enhances the affective component of one s attitude (e.g., satisfaction) towards such behaviour (Verplanken et al., 1998). Accordingly, we hypothesise that: H6: Shopping Experience has a significant positive effect on Shopping Satisfaction. Also, experience is shown to strengthen the accessibility of the affective component of attitude, for example, satisfaction (Verplanken et al., 1998). Accessibility refers to the speed of retrieving affect from memory. Enhanced accessibility, in turn, improves affect behaviour consistency (Fazio et al., 1982). This is consistent with studies reporting moderating effects of experience on the relationship between attitude and behavioural intention (Zanna et al., 1980; Fazio & Zanna, 1981). Therefore, we also argue that for the same level of satisfaction, higher experience leads to better accessibility of the satisfaction in memory, strengthening in this way the effect of satisfaction on intention. Accordingly, we hypothesise that: H7: Shopping Experience positively moderates the relationship between Shopping Satisfaction and Repurchase Intention. shopping experience Several previous studies have used experience as a surrogate measure for habit (e.g., Mittal, 1988; Montano & Taplin, 1991; East et al., 1994; Thompson et al., 1994; Verplanken et al., 1994; Bergeron et al., 1995; Verplanken et al., 1996; Verplanken & Svenson, 1997; Murray & Haubl, 2005). Accordingly, habit is often operationalised based on self-reported frequency of past behaviour (Aarts, Methodology Research procedure The research model was tested by conducting a survey study, which presents the advantages of enabling replicability, strengthening statistical power and serving as a foundation for building generalisability (Teo et al., 2003). The invitation to participate in the survey was distributed

6 consumer retention Mohamed Khalifa and Vanessa Liu 785 by through an internet service provider that identified a pool of online shoppers. The survey instrument was then electronically administered to respondents. The response rate could not be estimated as we were not granted permission to access the master mailing list. Participants Respondents of the survey composed of 122 online customers who had previously shopped from various internet stores. Participation was entirely voluntary. Table 1 presents the demographics of the respondents. Measurement In line with our definition of repurchase intention, the respondents were asked to think of a specific online store from which they had previously shopped in answering all questions in the survey. We measured online shopping habit, online shopping experience, and online repurchase intention using reflective items that were adapted from the literature (e.g., see Limayem & Hirt (2003) for habit, and Limayem et al. (2000) for online repurchase intention) and validated through the card sorting procedure (Moore & Benbasat, 1991). To identify specific usefulness factors for all phases of online shopping, we relied on belief elicitation to develop a formative measurement model for the construct of perceived usefulness. We wanted to distinguish between the different phases of online shopping (i.e. pre-purchase, purchase, and post-purchase experiences). Therefore, we also operationalised satisfaction as a formative, emergent construct formed with items based on these phases. Although the items of both perceived usefulness and satisfaction measure the perceptions towards similar offerings, for example, after-sale service, the major distinction lies in the contextual difference. Perceived usefulness refers to the evaluation of a specific offering, while satisfaction measures the general attitudes towards a stage in the purchase cycle as a whole. We developed an initial pool of formative items from a review of the literature. Twenty online shoppers were invited to participate in focused group discussions where they were asked to identify important usefulness factors, whereby assessing the face validity of the factors. Appendix A presents the formative items developed based on the belief elicitation results as well as all other measures in the survey. As respondents with more online shopping experience are likely to be more satisfied, hence introducing potential response bias, we examined the response distribution for satisfaction. An even distribution of the responses indicated that the possibility of having such a bias was remote. Although we rely on the same respondents for the dependent and independent variables, the structure of the research model (i.e. moderating effects) minimises potential common method bias. The moderating effects, if found to be significant, will provide a strong indication of the lack of common method bias (Evans, 1985). While Table 1 Demographics of respondents Demographic Category Percentage % Gender Male 62 Female 38 Age Less than Over 50 2 Marital Status Single 79 Married 21 Education High school 20 Undergraduate degree 38 Postgraduate degree 42 Experience with internet usage Less than 1 year years years 33 Over 4 years 58 shopping frequency Less than once a month 70 A few times a month 19 A few times a week 7 About once a day 4 Residing countries United states 100 the respondents may anticipate the linear relationships and answer accordingly, they are not likely to predict the moderating relationships. Existence of a common method bias increases statistically the covariance among the independent variables, impairing the likelihood of detecting a significant moderating effect. Therefore, with the verification of moderating effects, the presence of a common method bias becomes highly unlikely (Brockner et al., 1997). To check for possible non-response bias, we compared early respondents with late respondents (Armstrong & Overton, 1977). The last 25% to submit their response were considered to be late responses and were deemed to be representative of online shoppers who ultimately did not respond to the survey (Li & Calantone, 1998). We then compared the means of all items for the two groups and could not detect any significant differences. To assess the representativeness of our sample, we compared the demographics with those of global internet users (TNS-Infratest, 2005) and detected no significant differences. In formulating and testing the moderating effect of online shopping habit (experience) on the relationship between online shopping satisfaction and online repurchase intention, we followed a hierarchical process similar to multiple regressions where we compared the results of two models. That is, one with and one without

7 786 consumer retention Mohamed Khalifa and Vanessa Liu the interaction construct: online shopping habit (experience) online shopping satisfaction. The interaction construct is developed by multiplying the factor score for the satisfaction construct (with formative indicators) with each of the standardised items of the habit (experience) construct (with reflective indicators). The moderating effects are tested and interpreted according to the guidelines recommended by Carte & Russell (2003). More specifically, we assessed the significance of the change in the R 2, using the following formula (Carte & Russell, 2003): DR 2 =ðdf mul df add Þ ð1 R 2 mul Þ=ðN df mul 1Þ Measurement scale The survey instrument consisted of both reflective and formative indicators measured with a semantic differential scale in the form of a slider that was used to record the respondents answers (see Appendix A). The slider is a graphical scale with two anchors only appearing at both ends ( strongly disagree and strongly agree ). The neutral point is labelled as Indifferent. With a resolution ranging from 1 to 100, the slider provides 100-scale steps. According to numerous psychometric studies, the reliability of individual rating scales is a monotonically increasing function of the number of steps (Nunnally, 1978). Graphical scales are reported to be superior to numeric scales as people usually think of quantities as represented by degrees of physical extensions (e.g., the yardstick). Graphical scales can also help to convey the idea of a rating continuum and lessen clerical errors in making ratings (Nunnally, 1978). To mitigate potential threats to the reliability of the scale measurement resulting from different screen sizes and operativeness of input devices of the respondents, the exact ratings selected were indicated next to the slider. In this way, respondents were also able to re-adjust the slider when they mishandled the mouse devices. Data analysis Separate analyses were performed with research models using online shopping habit and online shopping experience respectively as the moderator. This allows us to compare the moderating and mediated effects of these constructs. From a methodological perspective, simultaneous testing of test effects may also be problematic. Although experience and habit may plausibly have independent effects on online repurchase intention, these two constructs may potentially be highly correlated as experience constitutes one of the pre-requisites for the formation of habit. The analysis of the data was done in a holistic manner with the partial least squares (PLS) procedure using PLS graph (Chin, 1998). The PLS procedure (Wold, 1989) allows the researcher to both specify the relationships among the conceptual factors of interest and the measures underlying each construct. Specifically, reflective indicators are invoked to account for the observed variances and covariance while formative indicators are weighted based on their relative importance in the formation of the construct (Chin, 1998; Law et al., 1998). The result of such a procedure is a simultaneous analysis of (1) how well the measures relate to each construct and (2) whether the hypothesised relationships at the theoretical level are empirically confirmed. Furthermore, due to non-normality of the data, structural equation modelling was not appropriate (Chin & Gopal, 1995; Bhattacherjee, 2004). PLS, on the other hand, is considered a better fit as it adopts a component-based approach and does not require multivariate normal distributions or a large sample size (Fornell & Bookstein, 1982). Compared to structural equation modelling, PLS is more focused on predictability by maximising the variance explained in constructs rather than overall fitness of the model (Barclay et al., 1995). With the prediction-oriented nature of the research questions, PLS is therefore considered preferable for this study. Tests of significance for all paths were conducted using the bootstrap re-sampling procedure (Cotterman & Senn, 1992). For reflective measures, we adopt the standard approach for evaluation, where all path loadings from construct to measures are expected to be strong (i.e., 0.70 or higher). The internal consistency of the reflective items was estimated using the composite reliability formula (Chin, 1998). We did not rely on the Cronbach s alpha, which represents a lower bound estimate of internal consistency due to its assumption of equal weightings of items. In the case of formative measures, we assess the contribution of each item to the pertaining construct based on their weights (Chin, 1998). Results and discussion Measurement model validation Tables 2, 3 and 4 present the results of the test of the measurement model. As shown in Table 2, the loadings of all reflective items are high (above 0.8) with significance at 1% level, demonstrating convergent validity. Composite reliability scores and average variance extracted (AVE) were computed to further assess convergent validity as well as internal consistency. Results (see Table 3) showed that for all constructs the composite reliability scores are higher than the recommended benchmark of 0.80 (Nunnally, 1978) whereas the AVEs exceeded the threshold of 0.50 (Hair et al., 1998), verifying internal consistency. As for discriminant validity, which denotes the extent to which the measures for each construct are distinct from each other, we performed the assessment by verifying whether the correlations between a referent construct with others are substantially different from the square roots of the AVE scores of that construct (Anderson, 1987). Such a comparison provided support for discriminant validity (correlation between repurchase

8 consumer retention Mohamed Khalifa and Vanessa Liu 787 Table 2 Reflective measurements Construct items Loadings Std. error t-values repurchase intention Anticipate to repurchase in the near future Likely to repurchase in the near future Expect to repurchase in the near future shopping habit Natural behaviour First thing that comes to mind Spontaneous shopping experience Used extensively Used for a long time Used frequently Table 3 Assessment of internal consistency, convergent validity, and discriminant validity Constructs No. of items Composite reliability (r) AVE (sq. root) Correlations with online repurchase intention Correlations with online shopping habit repurchase intention (0.93) 1.00 shopping habit (0.87) 0.50 shopping experience (0.84) Table 4 Formative measurements Constructs Items Weights Std. error t-values shopping satisfaction Post-purchase experience 0.59** Pre-purchase experience 0.34** Purchase experience 0.26** Perceived usefulness After-sale service 0.42** Transaction efficiency 0.42** Security 0.24** Convenience 0.23** Cost savings 0.20** **Significance at 1% level. and habit ¼ 0.50; correlation between repurchase and experience ¼ 0.61; see Table 3). We performed two tests to verify the conceptual distinction between online shopping habit and online shopping experience. First, we relied on the card sorting procedure. Habit and experience were classified into two separate constructs by the panel judges involved in the procedure, indicating face validity and hence discriminant validity. Furthermore, the square root of the AVE of the experience construct (0.84) is higher than the correlation of experience with habit (0.74) (see Table 3), therefore demonstrating discriminant validity. Based on these results, we conclude that online shopping habit and online shopping experience are conceptually distinct constructs. The higher correlation value, however, suggests potential multicollinearity problem if both constructs are included in the same model. The weights and the significance of all formative measures indicate that the items contribute significantly to the formation of the intended construct (Table 4). After-sale service and transaction efficiency emerged as the two most important drivers of perceived usefulness with identical weights of 0.42, both significant at the 1% level. The effects of security, convenience and cost savings are comparatively small, but are nevertheless significant. To check for the potential effects of demographics, we included age and gender in the model, but their path coefficients (links to online repurchase intention) were not significant. We therefore concluded that age and

9 788 consumer retention Mohamed Khalifa and Vanessa Liu gender differences did not provide additional explanation for intention formation. Results of PLS analysis Figure 2 provides the results of the PLS analysis of full model, including the habit construct. The estimated path effects are given along with their significance. All path coefficients are significant providing strong support for all the hypothesised relationships. Both perceived usefulness (b ¼ 0.31; P ¼ 0.01) and online shopping satisfaction (b ¼ 0.44; P ¼ 0.01) have significant positive effects on online repurchase intention, confirming H1 and H3. As hypothesised (H2), perceived usefulness also has a significant positive effect on satisfaction (b ¼ 0.65; P ¼ 0.01). These results confirm the applicability of the IS continuance model (Bhattacherjee, 2001a) to the online shopping context. The reduced model (without habit) explains about 66% of the variance in repurchase intention. Interestingly, this R 2 is identical to the one reported by Bhattacherjee (2001a). The full model (including habit), on the other hand, explains about 70% of the variance. Our results show strong support for both the mediated and moderating effects of online shopping habit. As hypothesised in H4, habit has a significant effect on satisfaction (b ¼ 0.18; P ¼ 0.01). The moderating effect of habit is also verified, as indicated by a significant change in R 2 (F ¼ 15.4; P ¼ 0.001) 2, confirming H5. In other words, as online shopping becomes more habitual, the effect of satisfaction on repurchase intention becomes stronger. We replicated the analysis using online shopping experience to compare its effects with those of online shopping habit (see Figure 3). The empirical results are quite similar. The full model with online shopping experience explains a comparable variance of online repurchase intention (R 2 ¼ 71%) to that of the full model with habit (R 2 ¼ 70%). Similar to habit, experience also exhibits mediated and moderating effects on online repurchase intention. As hypothesised in H6, experience has a significant effect on satisfaction (b ¼ 0.28; P ¼ 0.01). Confirmed by the significant change in R 2 (F ¼ 19.3; P ¼ 0.001), 3 experience also positively moderates the relationship between satisfaction and repurchase intention, verifying H7. These findings imply that experienced online customers are more likely to intend to repurchase when satisfied. The significance of H7 also qualifies online shopping experience as a valid segmentation criterion for customer retention. Managers may opt to collect data on customers online shopping experience in place of habit, which is relatively more difficult to measure. 2 F ¼ [R 2 /(df mul df add )]/[(1 R 2 mul)/(n df mul 1)] ¼ [( )/ (5 4)]/[(1 0.7)/( )] ¼ F ¼ [R 2 /(df mul df add )]/[(1 R 2 mul)/(n df mul 1)] ¼ [( )/ (5 4)]/[(1 0.71)/( )] ¼ Figure 2 Figure 3 Perceived Usefulness 0.65** R 2 : 0.56 Shopping Satisfaction ** p < ** 0.31** 0.44** R 2 (F = 15.4**) Shopping Habit R 2 : 0.70 Repurchase Intention Results of PLS analysis full model with habit. Perceived Usefulness 0.59** Shopping Satisfaction ** p < 0.01 R 2 : ** 0.30** 0.40** R 2 (F = 19.3**) Shopping Experience R 2 : 0.71 Repurchase Intention Results of PLS analysis full model with experience. Discussion In this study, we develop, operationalise, and empirically test a model that explains online consumer retention as measured by repurchase intention. Earlier studies in the marketing literature modelled repurchase as a direct effect of satisfaction. Recent IS research suggested continuance intention to be an outcome of satisfaction and perceived usefulness. Our findings demonstrated that such models may not be fully sufficient for explaining online shopping and that a contingency theory was needed (i.e. moderating factors of the link between satisfaction and repurchase intention should be included). Theoretical implications The empirical evidence provides strong support for our model, as demonstrated by the significant effects of habit and experience and the improved explanatory power (R 2 ) of the model. These results confirm our argument for the development of a contingency theory to account for the novelty of the online channel. Our findings also verify the conceptual distinction of the habit and experience constructs, reinforcing the argument of previous studies that experience is necessary but not sufficient for the formation of habit. Although conceptually distinct, experience and habit seem to have comparable effects.

10 consumer retention Mohamed Khalifa and Vanessa Liu 789 The similarity of their effects may be due to the fact that experience is a pre-requisite for habit formation. When we tried to include both constructs in the model, their high correlation created multicollinearity effects and only one of them (i.e., experience) remained significant. As to which construct should be considered, we suspect that for cases where online shopping habit is strong, its inclusion may provide a better explanation of online repurchase intention. On the other hand, when consumers have not acquired an online shopping habit yet, the inclusion of online shopping experience may provide a better explanatory power. The interplay of experience and habit should be investigated further in future research to shed more light on their relative importance in explaining online repurchase intention. The significant moderating effects of online shopping experience and habit provide important implications. The formation of online shopping habit may not necessarily involve a prolonged usage experience. Our results show that the effect of satisfaction on repurchase intention is strengthened with both habitual and experienced online shoppers. In other words, customers with online shopping habits are as likely to return as those experienced others when satisfied. Often times, customer retention programs focus extensively on experienced online shoppers. Our study demonstrates that habitual customers are worth similar attention despite they may not have a long history of online shopping. Managerial implications On the practical side, our study offers guidelines on customer profiling and prioritisation of customer retention programs for online retailers. Many companies grouped customers based on their satisfaction level. As our results show that experienced and habitual customers, when satisfied, are more likely to be retained, another useful market segmentation criterion may be the level of prior online shopping experience and online shopping habit. Customers may be segmented into four major profiles: (1) satisfied shoppers with online shopping experience or habit; (2) satisfied shoppers with less or without online shopping experience or habit; and (3) dissatisfied shoppers with online shopping experience or habit; and (4) dissatisfied shoppers without online shopping experience or habit. Each of these profiles should be assigned different marketing objectives and priority of retention efforts. For both the first and second profiles where customers are satisfied, the primary marketing objective should be the continuous enhancement of customer satisfaction in order to increase the likelihood of their repeat purchases. But for the second category, customer retention requires extra efforts on reinforcing the development of online shopping habit/ experience, converting the customers to more habitual online shoppers and hence increasing their life time profitability to the company. Managerial influence in this case is, however, limited. Internet merchants may exert little direct influence on such general habit/experience as the formation of which is subject to a number of individual factors, such as personality traits and general attitudes towards online shopping. Alternatively, managers may enhance the development of specific habit/ experience with buying from their particular stores. The online shopping channel may be useful in serving this purpose. For example, multimedia design may be incorporated in the website to improve the virtual shopping experience. retailers may also provide free online previews or trials of products to encourage the acquisition of usage experience. Furthermore, abolishment of traditional channels may be effective in enhancing the development of online shopping habit as in the case of banking services. Closures of bank branches have significantly contributed to the habit formation of using e-banking services. Product improvement may be the top priority for marketing managers serving the remaining two profiles in order to restore satisfaction from the dissatisfied customers. As our results also show the effects of habit and experience to be similar, practitioners may rely on prior experience in profiling customers who have not acquired the online shopping habit yet. Future research directions The role of online shopping habit may vary across industry contexts. For products or services where the majority of customers may already be accustomed to buying online (e.g., books and airline tickets), the role of habit in enhancing repurchase intention may not be as salient as it does in contexts where online shopping remains in its infancy stage (e.g., grocery shopping). Future research is needed to inform these variations by replicating our study in industry contexts varying in terms of the innovativeness of the online channel will shed more light on the relative importance of the moderating effects of habit and experience. Another limitation of this research is that its scope is limited to the online factors. Several empirical studies (e.g., Parasuraman et al., 1985) reported that the quality of how a product and service is actually delivered played a key role in satisfaction formation. Such studies suggest that the success of the online channel remains dependent on the effectiveness and efficiency of the physical channel, for example, delivery. Future research should consider online and contingent offline factors simultaneously. As our research model incorporates variables mainly from the TAM, future researchers should also expand our study by investigating the relative importance of other predictors of intention. For example, earlier empirical evidence demonstrates that online customers personality traits have a significant impact on the effectiveness of interaction with online retailers (Jahng et al., 2002). These variables have not been controlled for in the present study. Other potentially relevant factors include perceived usability and system quality (Juan et al., 2006). As well, testing the applicability of our research model to online customer communities may be of interest to internet retailers

11 790 consumer retention Mohamed Khalifa and Vanessa Liu given the recent proliferation of such e-commerce initiatives. Such examination may require additional variables like trust (e.g., see van der Heijden et al., 2003), which are important factors to consider when interpersonal relationships among customers are involved (Lankton & McKnight, 2006). Another contextual direction for future research is to extend and re-test our conceptual model in the context of actual purchase behaviour to assess its adequacy and the need for additional explanatory variables. From a methodological perspective, future research may also adopt a longitudinal design, investigating further the interplay of experience and habit and their relative importance in explaining continuance behaviour. Conclusion In conclusion, our study highlights the importance of the roles of online shopping habit and online shopping experience in the achievement of online customer retention. As verified by our data, online repurchase intention is not only an outcome of a rational analysis of satisfaction with and the perceived usefulness of previous transactions. Both habit and experience are also significant driving forces of online shopping behaviour. About the authors Professor Mohamed Khalifa received the degrees of M.A. in Decision sciences and Ph.D. in Information Systems from the Wharton Business School of the University of Pennsylvania. He is currently Professor at the Department of Information Systems, City University of Hong Kong. His current research interests include electronic business and knowledge management. His work appeared in journals such as Communications of the ACM, OMEGA, Journal of the Association of Information Systems, IEEE Transactions on Engineering Management, IEEE Transactions on Systems Man and Cybernetics, Decision Support Systems, Data Base and Information and Management. Vanessa Liu is currently Assistant Professor at the School of Management with joint appointment with the Department of Information systems at the New Jersey Institute of Technology. She received the degrees of M.Phil. and Ph.D. in information systems at the City University of Hong Kong. Her work has been published in some IS academic journals including Journal of the Association for Information Systems, OMEGA the International Journal of Management Science and IEEE Transactions on Professional Communication. She was one of the authors of the paper that was awarded the Best Paper in Conference Theme at ICIS Her research interests include electronic commerce and knowledge management. References AARTS H (1996) Habit and decision making: the case of travel model choice. Unpublished doctoral dissertation, University of Nijmegen, The Netherlands. AARTS H and DIJKSTERHUIS A (2000) Habits as knowledge structures: automaticity in goal-directed behaviour. Journal of Personality and Social Psychology 78(1), AARTS H, VERPLANKEN B and VAN KNIPPENBERG A (1998) Predicting behavior from actions in the past: repeated decision making or a matter of habit? Journal of Applied Social Psychology 28(15), AJZEN I (1991) The theory of planned behavior. Organizational Behavior and Human Decision Processes 50(2), AJZEN I (1988) Attitudes, Personality and Behaviour. The Dorsey Press, Chicago, IL. ANDERSON JC (1987) An approach for confirmatory measurement and structural equation modelling of organizational properties. Management Science 33(4), ANDERSON RE and SRINIVASAN SS (2003) E-satisfaction and E-loyalty: a contingency framework. Psychology & Marketing 20(2), ARMSTRONG JS and OVERTON TS (1977) Estimating nonresponse bias in mail surveys. Journal of Marketing Research 14, BARCLAY DC, HIGGINS CandTHOMPSON R (1995) The partial least squares approach to causal modelling: personal computer adoption and use as an illustration. Technology Studies 2(2), BARGH JA and FERGUSON MJ (2000) Beyond behaviourism: on the automaticity of higher mental processes. Psychological Bulletin 126(6), BEARDEN WO and TEEL JE (1983) Selected determinants of consumer satisfaction and complaint reports. Journal of Marketing Research 20, BERGERON F, RAYMOND L, RIVARD SandGARA MF (1995) Determinants of EIS use: testing a behavioural model. Decision Support Systems 14, BHATTACHERJEE A (2001a) Understanding information systems continuance: an expectation-confirmation model. MIS Quarterly 25(3), BHATTACHERJEE A (2001b) An empirical analysis of the antecedents of electronic commerce service continuance. Decision Support Systems 32(2), BHATTACHERJEE A (2004) Understanding changes in belief and attitude toward information technology usage: a theoretical model and longitudinal test. MIS Quarterly 28(2), BROCKNER J, SIEGEL PA, DALY JP, TYLER TandMARTIN C (1997) When trust matters: the moderating effect of outcome favourability. Administrative Science Quarterly 42(3), CAPRARO A, BRONIARCZYK J and SRIVASTAVA RK (2003) Factors influencing the likelihood of customer defection: the role of consumer knowledge. Journal of the Academy of Marketing Science 31(2), CARTE TA and RUSSELL CJ (2003) In pursuit of moderation: nine common errors and their solutions. MIS Quarterly 27(3), CHAUDHURI A (1999) Does brand loyalty mediate brand equity outcomes? Journal of Marketing Theory and Practice 7(2),

12 consumer retention Mohamed Khalifa and Vanessa Liu 791 CHILDERS TL, CARR CL, PECK JandCARSON S (2001) Hedonic and utilitarian motivations for online retail shopping behavior. Journal of Retailing 77, CHIN WW (1998) The partial least squares approach for structural equation modelling. In Modern Methods For Business Research (MARCOULIDES GA, Ed), pp , Lawrence Erlbaum Associates, Hillsdale, NJ. CHIN WW and GOPAL A (1995) Adoption intention in GSS: importance of beliefs. Data Base Advance 26(2 & 3), CHURCHILL G and SUPRENANT C (1982) An investigation into the determinants of customer satisfaction. Journal of Marketing Research 19, CISCO SYSTEMS (November/December 2003) Consumers Develop an Habit, archives/nov_dec_2003/departments/net_index/online_habit.html. COPELAND MT (1923) Relation of consumer s buying habits to marketing methods. Harvard Business Review 1, COTTERMAN WandSENN J (1992) Challenges and Strategies for Research in Information Systems Development. John Wiley & Sons, Chichester. DAVIS FD, BAGOZZI RP and WARSHAW PR (1989) User acceptance of computer technology: A comparison of two theoretical models. Management Science 35(8), DOYLE P (2003) Marketing management & strategy. marketing.co.uk/marketing-focus.htm (Current: November 28, 2003). EAST R, LOMAX W, WILSON GandHARRIS P (1994) Decision making and habit in shopping times. European Journal of Marketing 28(4), EAGLY AH and CHAIKEN S (1993) The Psychology of Attitudes. Harcourt Brace Jovanovich, Inc., Orlando. EVANS MG (1985) A Monte Carlo study of the effects of correlated method variance in moderated multiple regression analysis. Organizational Behaviour and Human Decision Processes 36(3), FAZIO RH, CHEN J, MCDONEL EC and SHERMAN SJ (1982) Attitude accessibility, attitude behaviour consistency and the strength of the object-evaluation association. Journal of Experimental Social Psychology 18, FAZIO RandZANNA MP (1981) Direct experience and attitude behaviour consistency. Advances in Experimental Social Psychology 14(1), FEINBERG RA, KADAM R, HOKAMA LandKIM I (2002) The state of electronic customer relationship management in retailing. International Journal of Retail & Distribution Management 30(10), FESTINGER LA (1957) A Theory of Cognitive Dissonance. Row and Peterson, Evanston. FORNELL CandBOOKSTEIN FL (1982) Two structural equation models: LISREL and PLS applied to consumer exit-voice theory. Journal of Marketing Research 19(3), GEFEN D (2003) TAM or just plan habit: a look at experienced online shoppers. Journal of End User Computing 15(3), GEFEN DandSTRAUB D (2000) The relative importance of perceived ease of use in IS adoption: a study of E-commerce adoption. Journal of Association for Information Systems 1(8), HAIR JF, ANDERSON RE, TATHAM RL and BLACK WC (1998) Multivariate Data Analysis with Readings, 5th edn, Macmillan, New York. HALLOWELL R (1996) The relationships of customer satisfaction, customer loyalty, and profitability: an empirical study. International Journal of Service Industry Management 7, HO CS and WU WH (1999) Antecedents of customer satisfaction on the internet: an empirical study of online shopping. Proceedings of the 32nd Hawaii International Conference on System Sciences (SPRAGUE RH, Ed), IEEE Computer Society, Maui, Hawaii, January, 1 9. JACOBY J and CHESTNUT RW (1978) Brand Loyalty Measurement and Management. John Wiley & Sons, New York. JAHNG J, JAIN H and RAMAMURTHY K (2002) Personality traits and effectiveness of presentation of product information in e-business systems. 11, JONES TandSASSER EW (1995) Why satisfied customers defect. Harvard Business Review 73, JONES MA and SUH J (2000) Transaction-specific satisfaction and overall satisfaction: an empirical analysis. Journal of Services Marketing 14(2), JUAN C, CHIU CM and MARTINEZ FJ (2006) Understanding e-learning continuance intention: an extension of the technology acceptance model. International Journal of Human-Computer Studies 64(8), KARAHANNA E, STRAUB DW and CHERVANY ML (1999) Information technology adoption across time: a cross-sectional comparison of pre-adoption and post-adoption beliefs. MIS Quarterly 23(2), KHALIFA M, LIMAYEM MandLIU V (2002) consumer stickiness: a longitudinal study. Journal of Global Information Management 10(3), KHARE AandINMAN JJ (2005) Habit regimes in consumption. Advances in Consumer Research 32, KOUFARIS M (2002) Applying the technology acceptance model and flow theory to online consumer behaviour. Information Systems Research 13(2), LANKTON NandMCKNIGHT DH (2006) Using expectation-disconfirmation theory to predict technology trust and continuance usage intentions. Working Paper, Workshops of University of Minnesota. LAW KS, WONG CandMOBLEY WH (1998) Toward of a taxonomy of multidimensional constructs. Academy of Management Review 23(4), LIMAYEM M and HIRT SG (2003) Force of habit and information systems usage: theory and initial validation. Journal of the Association for Information Systems (4), LIMAYEM M, HIRT SG and CHIN WW (2001) Intention does not always matter: the contingent role of habit on it usage behaviour. Proceedings of the 9th European Conference on Information Systems Global Co-operation in the New Millennium, ECLS 2001, Bled, Slovenia, June 27 29, LIMAYEM M, KHALIFA MandFRINI A (2000) What makes consumers buy from internet? A longitudinal study of online shopping. IEEE Transactions on Systems, Man, and Cybernetics Part A: Systems and Humans 30(4), LI T and CALANTONE RJ (1998) The impact of market knowledge competence on new product advantage: conceptualization and empirical examination. Journal of Marketing 62(3), LU J (2003) A model for evaluating e-commerce based on cost/benefit and customer satisfaction. Information Systems Frontiers 5(3), MATHIESON K (1991) Predicting user intentions: comparing the technology acceptance model with the theory of planned behaviour. Information Systems Research 2(3), MITTAL B (1988) Achieving higher seat belt usage: the role of habit in bridging the attitude behaviour gap. Journal of Applied Social Psychology 18, MONTANO DE and TAPLIN SH (1991) A test of an expanded theory of reasoned action to predict mammography participation. Social Science and Medicine 32, MOORE GC and BENBASAT I (1991) Development of an instrument to measure the perceptions of adopting an information technology innovation. Information Systems Research 2(3), MURRAY KB and HAUBL G (2005) Skill-based habits of use and consumer choices. Advances in Consumer Research 32, NEMZOW M (1999) E-commerce Stickiness for customer retention. Journal of Internet Banking and Commerce 8(3) com/commerce/jibc/ htm. NIJSSEN E, SINGH J, SIRDESHMUKH D and HOLZMUELLER H (2003) Investigating industry context effects in consumer firm relationships: preliminary results from a dispositional approach. Journal of the Academy of Marketing Science 31, NUNNALLY JC (1978) Psychometric Theory. McGraw-Hill, New York. OLIVER RL (1979) Product satisfaction as a function of prior expectation and subsequent disconfirmation: new evidence. In New Dimensions Of Consumer Satisfaction And Complaining Behaviour (DAY RandHUNT HK, Eds), pp 66 71, Indiana University, Bloomington. OLIVER RL (1980) A cognitive model of the antecedents and consequences of satisfaction decisions. Journal of Marketing Research 17, OLIVER RL (1997) Satisfaction: A Behavioural Perspective on the Consumer. McGraw-Hill, New York. OLIVER RL (1999) Whence consumer loyalty. Journal of Marketing 63,

13 792 consumer retention Mohamed Khalifa and Vanessa Liu OLIVER RL and DESARBO WS (1988) Response determinants in satisfaction judgments. Journal of Consumer Research 14, OUELLETE JA and WOOD W (1998) Habit and intention in everyday life: the multiple processes by which past behavior predicts future behavior. Psychological Bulletin 124(1), PARASURAMAN A, ZEITHAML VA and BERRY LL (1985) A conceptual model of service quality and its implications for future research. Journal of Marketing 49(4), PATELI A and GIAGLIS G (2004) A research framework for analysing ebusiness models. 13(4), QUINN J and WOOD W (2005) Everyday habits and self-regulation. Advances in Consumer Research 32, REICHHELD FF (1996) The Loyalty Effect: The Hidden Force Behind Growth, Profits, and Lasting Value. Harvard Business School Press, Boston. RUST RT and ZAHORIK AJ (1993) Customer satisfaction, customer retention, and market share. Journal of Retailing 69, RUST RT, ZAHORIK AJ and KEININGHAM TL (1995) Return on quality (ROQ): making service quality financially accountable. Journal of Marketing 59, SODERLUND M, VILGON MandGUNNARSSON J (2001) Predicting purchasing behaviour on business-to-business markets. European Journal of Marketing 35(1/2), TAYLOR SandTODD PA (1995) Understanding information technology usage: a test of competing models. Information Systems Research 6(2), TEO HH, WEI KK and BENBASAT I (2003) Predicting intention to adopt interorganizational linkages: an institutional perspective. MIS Quarterly 27(1), THOMPSON R, HIGGINS CA and HOWELL JM (1994) Influence of experience on personal computer utilization: testing a conceptual model. Journal of Management Information Systems 11(1), TNS-INFRATEST (2005) Monitoring the Information Economy 8th Factual Report June TNS-Infratest Business Intelligence, Munich. TRIANDIS HC (1971) Attitude and Attitude Change. John Wiley & Sons, New York. TRIANDIS HC (1980) Values, attitudes, and interpersonal behaviour. In Beliefs, Attitudes and Values (HOWE HE, Ed), pp , Nebraska Symposium on Motivation, University of Nebraska Press, Lincoln. VAN DER HEIJDEN H, VEHAGEN TandCREEMERS M (2003) Understanding online purchase intentions: contributions from technology and trust perspectives. 12, VERPLANKEN B, AARTS H, VAN KNIPPENBERG A and VAN KNIPPENBERG C (1994) Attitude vs general habit: antecedents of travel model choice. Journal of Applied Social Psychology 24, VERPLANKEN B, AARTS H, VAN KNIPPENBERG A and MOONEN A (1996) Habit vs planned behaviour: a field experiment. Unpublished manuscript, University of Nijmegen. VERPLANKEN B, AARTS H, VAN KNIPPENBERG A and MOONEN A (1998) Habit, information acquisition, and the process of making travel mode choices. British Journal of Social Psychology 37, VERPLANKEN BandSVENSON O. (1997) Personal involvement in human decision-making: on conceptualizations and effects on decision processes. In Decision Making: Cognitive Models and Explanations (RANYARD R, CROZIER WR and SVENSON O, Eds), pp , Routledge, London. WOLD H (1989) Introduction to the second generation of multivariate analysis. In Theoretical Empiricism (WOLD H, Ed), pp vii xl, Paragon House, New York. ZANNA MP, OLSON JM and FAZIO RH (1980) Attitude-behaviour consistency: an individual difference perspective. Journal of Personality and Social Psychology 38, Appendix A survey instrument Strongly Disagree mid-point of the slider Indifferent Click to display the pointer Strongly Agree Items measuring Repurchase Intention (Limayem et al., 2000) I anticipate to repurchase from this internet store in the near future It is likely that I will repurchase from this internet store in the near future I expect to repurchase from this internet store in the near future Items measuring Shopping Habit (Limayem & Hirt, 2003) Shopping online has become a natural act for me Whenever I think of shopping, the internet comes to my mind shopping has become spontaneous for me Items measuring Shopping Experience (Limayem & Hirt, 2003) I have shopped online extensively I have used the internet to shop for a long time I shop online frequently Items measuring Shopping Satisfaction (Bhattacherjee, 2001a) I am satisfied with the pre-purchase experience of this internet store (e.g., consumer education, product search, quality of information about products, product comparison) I am satisfied with the purchase experience of this internet store (e.g., ordering, payment procedure) I am satisfied with the post-purchase experience of this internet store (e.g., customer support and after sale support, handling of returns/refunds, delivery care) Items measuring Perceived Usefulness (Bhattacherjee, 2001a) I get good after-sale service from this internet store Transaction processing in this internet store is efficient (e.g., fast retrieval of information, ordering, payment processing and scheduling delivery) I am not too concerned about the security of this internet store Shopping from this internet store is convenient Purchasing from this internet store allows me to save money