Estimating Position and Social Influence Effects in Online Search: An Empirical Generalized Weitzman Model

More on Andrew Ching

Signup Sheet

Abstract:  In this paper, we contribute to the literature on the structural modeling of online consumer search by (i) separately identifying product position and social influence effects, and (ii) allowing consumers to purchase multiple items in each search session. We achieve these two goals by utilizing a publicly available dataset from a field experiment conducted on an online music platform and a new structural search model that allows multiple purchases in each search session. The field experiment, which manipulates presentation formats, product positions, and popularity information, provides us with the necessary exogenous data variation to address our identification challenge. We find that the social influence effect is twice as impactful in the column format compared to the matrix format. Our counterfactual experiments show that by revealing product popularity and sorting items accordingly, platforms can help customers locate and select products more efficiently, resulting in reduced search efforts but more purchases. Conversely, sorting products randomly with the disclosure of their popularity leads to the highest search activity by customers.

Joint with Ata Jameei Osgouei, Brian Ratchford and Shervin Shahrokhi Tehrani