Prior Information and Consumer Search: Evidence from Eye Tracking

Do consumers search the brands they know more or less frequently than the brands with which they are unfamiliar? In this paper, we attempt to answer this question using data from an experiment with two novel features: (i) survey information on consumers’ prior brand ownership, familiarity with each brand, and prior experience using different product features and (ii) eye-tracking data capturing search behavior at a very granular level. We find that consumers are generally more likely to search and buy brands they own and with which they are familiar, highlighting the importance of accounting for prior information. For this reason, we develop a search model in which both the information obtained during the search process and the information possessed by consumers prior to search are allowed to influence search and purchase decisions. Our model contributes to prior work by modeling search at the brand and attribute levels within a Bayesian learning framework. Using this model, we then quantify the impact of prior information on consumer choices as well as document the estimation bias arising when prior information is absent from the model. Finally, through a series of counterfactuals, we explore the managerial value of prior information data. This paper was accepted by Matt Shum, marketing. Funding: Q. Wang acknowledges financial support from the National Science Foundation of China [Grant 72072145] and the Natural Science Basic Research Program of Shaanxi Province [Grant 2020JQ-226]. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2021.00611 .