Effectiveness of Website Personalization: Does the Presence of Personalized Recommendations Cannibalize Sampling of Other Items?
David Bodoff and Shuk Ying Ho
International Journal of Electronic Commerce,
Volume 20, Number 2, Winter 2015-16, pp. 208-235.
Abstract:
With advances in personalization technologies, websites are increasingly able to customize Web content and provide users with a unique experience. Prior research has focused more on the appeal of the personalized recommendations themselves and examined user behavior toward the personalized recommendations. However, prior research has not compared the total volume of users’ sampling activity on websites that offer personalization, compared to those that do not. This motivates our interest in an extended study of the totality of users’ sampling behavior on the website when personalized recommendations are present. Where prior work has focused on the amount of sampling of the personalized items themselves, we define two additional indicators that characterize a consumer’s sampling behavior on a website that has both stock and personalized items: (1) the total volume of sampling to include both stock and personalized items, and (2) the pattern of sampling between personalized and stock items. Based on consumer search theory (CST) and a simulation analysis, we propose hypotheses about consumers’ behavior on these expanded outcomes, and test them in a longitudinal field study. We find that subjects sample personalized items before stock items, and sample fewer items in total when personalized recommendations are available. In addition, subjects engage in less sampling as time goes on, across sessions. Theoretically, this research brings the CST perspective to the literature on website personalization. Practically, this research sheds light on the precise nature of the benefit and possible drawbacks of personalization from the merchant’s perspective.
Key Words and Phrases: Consumer search theory, field experiment, longitudinal study, sampling pattern, sequential search, website personalization.