Traditional marketing researchers have suggested various models for customer satisfaction/ dissatisfaction. Recently suggested models for eCRM have been developed from traditional cus- tomer satisfaction/dissatisfaction models. Those models have investigated how variables affect customer satisfaction/dissatisfaction. Traditional models for customer satisfaction/dissatisfaction include the expectation-disconfirmation model, perceived-performance model, norms-based model, multiple-process models, attribution models, affective model, equity model, the American Cus- tomer Satisfaction Index model (ACSI), and complaint behavior model.
Various researchers have measured the level of satisfaction/dissatisfaction and complaints by considering the difference between expectations and disconfirmation. Erevelles and Leavitt (1992) posit that the expectancy-disconfirmation (ED) paradigm has dominated consumer satisfaction/ dissatisfaction research since its emergence as a legitimate field of inquiry in the early 1970s. According to this paradigm, consumers are believed to form expectations about a product before they purchase it (Oliver 1980). The ED paradigm can be derived from expectancy theory (Tolman 1932), and, especially, the notion of expectations is generally defined as consumers’ beliefs that a product has certain desired attributes (Erevelles and Leavitt 1992). Bearden and Teel (1983) also considered expectations disconfirmation in the model to examine the antecedents and conse- quences of customer satisfaction/dissatisfaction. Oliver (1980) established a process to describe how satisfaction is produced in this expectation-disconfirmation framework. Before making a purchase, buyers form expectations of the products or service. Consumption of the product or service reveals a level of perceived quality (which itself can be influenced by expectations). The perceived quality either positively confirms expectations or negatively disconfirms them. Expec- tations serve, in Oliver’s model, as an anchor or baseline for satisfaction, the positive confirma- tion or negative disconfirmation either increasing or decreasing the customer’s resulting satisfaction (Vavra 1997).
A traditional model of satisfaction by Oliver (1980: Figure 1) is related to the The Ameri- can Customer Satisfaction Index (ACSI), which was developed by Fornell (1992), and roughly emulates a national measure conducted in Sweden, the Swedish Customer Satisfaction Ba- rometer. Fornell’s model expresses satisfaction as the result of three elements: perceived (ex-
Figure 3.1 The American Customer Satisfaction Index (ACSI) Model
Customization Perceived Quality
Customization ExpeCctuastitoonmser Expectations
Value S(AaCtiSsfIa) ction
Confirm/Disconfirm to Expectations
Source: American Society for Quality Research.
perienced) quality, expectations, and perceived value. Customer satisfaction models (Figure 3.1) have considered three components: antecedents (prepurchase), satisfaction process, and consequences (postpurchase). Prior experience is the most important antecedent of satisfac- tion. The model explains influences, such as demographics, word of mouth, personal exper- tise, evolution of technology, nature of competition, advertising and PR that affect customer expectations and performance. Further, the model explores how the satisfaction process sub- sequently influences complaining (or complementing) behavior as well as customer loyalty (Fornell 1992). Models are also embedded in the system of cause-and-effect relationships (as shown in Figure 3.1), which makes the model the centerpiece in a chain of relationships run- ning from the antecedents of overall customer satisfaction—voice and loyalty (Bateson and Hoffman 1999).
Table 3.2 summarizes variables applied for the study of eCRM. Cho, Im, Hiltz, and Fjermestad (2001b) examined how differences in degree of dissatisfaction may occur, for many reasons, between online and offline customers. The major reasons include problems associated with different customer service center approaches (e.g., lack of an information or help desk during the order process, slow feedback response time, poor after-sales support), general terms and conditions (e.g., guarantees, guidelines for returning products), delivery issues (e.g., late or no delivery, product damage during delivery), security and privacy issues, failure of information quality, and system performance (e.g., slow Web sites, broken links to other pages).
Studies by Schubert (2002–2003) and Gehrke and Turban (1999) measured Web site effec- tiveness and also considered it an important dependent variable for eCRM. Various studies ap- plied customer satisfaction as a dependent variable that affects eCRM (Cho and Ha 2004a). A suggested eCRM model by Lee, Kim and Moon (2000) applied customer loyalty as a dependent variable. The majority of studies examined Web site effectiveness including page loading speed, navigation efficiency (Gehrke and Turban 1999), information quality (Cho and Ha 2004a; Kim and Moon 2000; Schubert 2002–2003), product expectation (Cho, Im, Hiltz, and Fjermestad