Impact of conflict delisting and relisting on remaining products in retail stores: Sales gains across products categories and spillovers to nearby stores

• Published in: Production and operations management Vol. 32; no. 7; pp. 2264 - 2282
• Main Authors: Li, H. Alice, Wan, Xiang
• Format: Journal Article
• Language: English
• Published: Los Angeles, CA SAGE Publications 01.07.2023
• Subjects: conflict delisting; difference‐in‐differences; generalized random forest; machine learning; relisting; spillover; stockout; spillover; stockout; conflict delisting; generalized random forest; difference‐in‐differences; relisting; machine learning
• ISSN: 1059-1478; 1937-5956
• DOI: 10.1111/poms.13972

Disputes between retailers and manufacturers often result in the retailer delisting the manufacturer's products, which dramatically alters competition in the retail market. When rival products are temporarily delisted, the products remaining on retail shelves are expected to gain sales due to substitution. However, it is unclear how much sales gain of remaining products can be expected, how long the sales gain persists after the rival products are relisted, whether the sales gain varies in different product categories, and whether the sales gain spills over from the delisting retailer to nearby retailers. In this research, we exploit a natural experiment enabled by a dispute between a retail chain and a beverage manufacturer that resulted in a 4‐week delisting of all products by the manufacturer (the delisted manufacturer) across all of the retailer's U.S. stores. We focus on a manufacturer (the focal manufacturer) that directly competes with the delisted manufacturer and quantify the sales gain for this focal manufacturer during the delisting and after the relisting event. Interestingly, in a near‐duopoly market, the focal manufacturer only gained 12.2% of sales during the delisting. After the dispute was settled and competition restored, the focal manufacturer's sales fluctuated for about 2 months before returning to the pre‐delisting level. Additionally, we find a spillover in sales gain to nearby stores during the delisting, which was mainly to small‐box retail stores rather than big‐box stores. Moreover, at both delisting and nearby stores, we find the focal manufacturer's sugar‐sweetened, caffeinated, and star products gained sales during the event while other products did not. We check the robustness of our results with a machine learning nonparametric method, among a few other alternative methods, to measure the average treatment effects. Understanding the size and duration of these heterogeneous effects can help manufacturers and retailers better respond to changes in market competition and evaluate their delisting decisions.