A Competing Risk and Bayesian Learning Model of Retail Out-of-Stocks

• Activity Number: 240
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Marketing
• Author(s): Jason Duan*+ and Frenkel ter Hofstede
• Companies: The University of Texas at Austin and The University of Texas at Austin
• Keywords: competing risk ; choice model ; Bayesian learning ; out-of-stock ; promotion ; inventory

Abstract

In this study we investigate the effects of out-of-stocks on consumer purchase behavior at retail outlets. When a consumer faces an out-of-stock situation of the product of her choice, she may decide to postpone the purchase, choose a different package size of the same brand or switch to another brand. Switching due to out-of-stocks can help the consumer gain more information on different brands and package sizes, which in turn may change the consumer's future product preferences. Such changes have an impact on a retailer's category profit. Based on structural utility maximization, we develop a discrete-time competing risk hazard model which simultaneously explains consumers' purchase incidences, timing and product choices. To capture the long-term impact of out-of-stock, we allow consumers to learn from past experiences with the products, some of which result from out-of-stocks. We specify a Bayesian learning model in which consumers update their preferences. We apply our methodology to two year purchase histories of a panel of consumers for a product category characterized by frequent out-of-stocks. Our estimation uses individual-level data and complete information on product ava