Estimating Position and Social Influence Effects in Online Search
Ata Jameei Osgouei, Andrew T. Ching, Brian T. Ratchford, Shervin Shahrokhi Tehrani
Marketing Science · 2026 · Volume 45 · Issue 1
Abstract
This paper contributes to the literature on structural modeling of online consumer search by incorporating both product position and social influence effects (e.g., product popularity) while allowing consumers to select multiple items per search and learn adaptively about their match with the platform. Disentangling the product position and popularity effects is challenging because they are typically highly correlated in field data. To address this, we leverage a publicly available data set from a field experiment conducted on an online music platform. The experimental design motivates a two-step estimation procedure to aid identification. We also introduce a content-based filtering method to predict individual download probability for all products, including those not sampled in the experiment. Combining this download decision model with our structural search model, we conduct counterfactual experiments that show (i) revealing product popularity and sorting products accordingly improves consumer efficiency by reducing search effort and increasing downloads; (ii) random sorting combined with product popularity disclosure stimulates more search activity, potentially boosting advertising revenue; and (iii) sorting based on personalized ranking with displayed product popularity enhances engagement by increasing both downloads and search efficiency. History: Tat Chan served as the senior editor. Supplemental Material: The web appendix and data files are available at https://doi.org/10.1287/mksc.2023.0392 .
Keywords
multi-item selection in consumer search, product position, social influence, information disclosure, content-based filtering, generalized Weitzman model
Enable JavaScript for citation tools and additional paper interactions.