Personalization is a new system development approach for designing information systems that change configurations based on each user's needs and preferences. Although personalization capabilities are present throughout a large number of commercial software packages, they are just beginning to be incorporated into electronic commerce. Most of this personalization has been done in an ad hoc fashion. In this paper, we present a categorization framework for organizing the various types of personalization that have been attempted on Web sites. We develop an algorithm for classifying Web sites into high, medium and low personalization support and apply it to a set of well-known Web sites such as amazon.com. Finally, we discuss why various Web sites have high or low degrees of personalization depending on the organization's goals and product mix and also how personalization might have little or large impact on a Web site's achieving its intended goal (e.g., sales, customer service, information dissemination, etc.) We also present future research that needs to be done to support our theories on how personalization impacts a Web site's success.