Repeated Product Searches and Choice Elimination: Evidence from a Lab Study

Abstract

An important assumption of empirical search models is that search can only expand the set of products finally considered for a purchase (e.g., more clicks while searching online implies larger consideration sets). Additionally, most such models posit that consumers would only search a product again (revisit) in order to purchase it. In this paper, we test both of these assumptions using a controlled incentive-aligned lab study that facilitates queries of consumer actions as they go through the search process. Our results show that consumer behavior is not always consistent with search models frequently used in empirical work. Instead, we find that the search process results in both addition and elimination of a product from the set considered for purchase. Importantly, 90% of eliminated products are those that have never been revisited. This result implies that revisits are important for choice predictions, since they can predict eliminated products which will never be purchased. Further, revisits are quite prevalent (27% of all product searches) and they occur for a variety of motivations that differ from a desire to make a purchase, e.g., obtaining more information, comparing options, forgetting, etc. We discuss implications for theory, as well as managerial implications related to consideration set predictions and targeting.

Keywords: consumer search, search revisits, incentive-aligned research study, consideration set formation, choice modeling

JEL Classification: D83, L81, M31