During the last decade online retail sales have been growing constantly. This growth has been possible due to different factors like online reviews. Online reviews have been proven successful in predicting different variables like trust and sales in online settings; however, the impact of online reviews on other variables like customer satisfaction has not been widely studied. Based on expectation-disconfirmation theory, this study analyzes the effect of online reviews on customer satisfaction. A set of six hypotheses were proposed and tested by using a controlled experiment. A total of 278 usable responses were obtained from a sample of college students from a major Midwest US university. Five of the six hypotheses were supported. Results indicated that expectations and perceived performance are significant predictors of disconfirmation. They also indicate that disconfirmation is a significant predictor of satisfaction. Regarding the effect of online reviews on satisfaction, results showed an indirect effect of online reviews on satisfaction mediated by expectations and disconfirmation. Results have implications for research and practice. For research, results help to increase the understanding of customer expectations formation in online settings. For practice, results give advice to sellers about how to increase customer satisfaction.

Herzberg (1959) proposed a two-factor model, often named as two-factor theory when adopted by the following researches. Through measuring the level of dissatisfaction and satisfaction, he classified antecedents of work satisfaction into hygiene and motivator two type. While hygiene factors are variables which lead to dissatisfaction, motivators are variables correlated with satisfaction. In marketing area, Kano (1984) also proposed a similar concept, named two-way quality model. This model adopts both satisfaction and dissatisfaction concepts and, based on these concepts, he further classified quality properties into Must-be, Attractive and One-dimension three types. This implies that the separation of satisfaction with two dimensions allow researchers to understand how each antecedent functions. However, in past information system (IS) area, IS satisfaction was treated as a linear concept and measured with Likert or semantic differential scale. This approach lead to a direct consequence that identifying possible antecedents is possible but understanding the way they function is difficult. The purpose of this study is to develop a better approach to understand how different antecedents affect IS satisfaction. We attempt to adopt the confirmation concept in expectation-disconfirmation theory (EDT). Through measuring the level of confirmation (or disconfirmation) and satisfaction (or dissatisfaction), this study attempts to clarify how each antecedent generate effect under different contexts. This study attempts to achieve its purpose through a two-stage approach. We first used modified Delphi obtain various antecedents of IS satisfaction. In this stage, opinions from 29 scholars and experts resulted in 6 dimensions, contain 28 variables in total. In the next stage, we collected empirical data through survey to classify 28 variables into different types. We find the factors about the relation with user and IT are hygiene factor, e.g. system quality etc. The factors about the relation with user, IT and task are motivator, e.g. information quality etc. The factors about the relation with user, IT and organization are motivator and performance factor, e.g. management support and computer training are performance factors, MIS¡¦s service quality etc. are motivator. The result should provide insightful implications for both IS academia and practitioners.