Displays, Sales, and In-Store Search in Retail Markets

This paper develops and estimates a structural model of differentiated-products demand in an environment with in-store displays and costly consumer price search. This investigation is motivated by two stylized facts on retail markets. First, the prevalence of retail “sales” means that relative prices in most product categories vary substantially from week to week. Second, quantities sold often respond at least as much to changes in relative locations (in-store displays) as to changes in relative prices. The model proposed here incorporates both effects: sale-induced price variation provides a reason to search, and displays convey information about prevailing prices. This model is then applied to store-level data on laundry detergent purchases, using short-run price fluctuation to recover preference parameters and short-run display fluctuations to recover search parameters. The resulting structural estimates suggest that information frictions have substantial effects on purchase outcomes, with roughly 52 percent of consumers having positive search costs and a mean search cost of roughly $1.68 among this sub-population. I further explore the potential relationship between consumer search and demand analysis, and find that accounting for displays and other promotions substantially lowers elasticity estimates