Going Back to Move Forward? How Search Revisits on a Website We Built, and in Field Data, Inform Us about Search Outcomes

Abstract

Why do consumers search some products more than once (revisit) before making a purchase decision? How are these revisits related to search outcomes, such as consumers' consideration sets and choices? In this paper, the authors build an online shopping website and run an incentive-aligned research study to find out. They show that most consumers revisit with the goal of comparing products, followed second by the desire to obtain more information, and third by forgetting. Search behavior differs across revisit motivations, with consumers revisiting to compare when products are similar and revisiting to obtain information when previous searches are short. Revisits to obtain more information, although not the most frequently occurring, are most likely to lead to a purchase. Interestingly, revisits also reveal information about consumers' consideration sets, which are typically unobserved: most consumers have eliminated a product from consideration if they don't revisit it. With these insights, the authors train a Random Forest model to show that revisits and other search patterns can be used to predict consumer consideration sets. They also demonstrate the external validity of the results on two datasets: hotel searches from Trivago and camera searches collected by comScore. Using the Random Forest model the authors show that choice predictions in these two datasets can be improved with information on revisits and predicted consideration sets. Finally, managerial implications of the results, as well as suggestions for modeling consumer search behavior are discussed.

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

JEL Classification: D83, L81, M31