Quantifying peer effects in large branded networks: The role of distance, ownership, experience, and market size

• 成果类型: Article; Early Access
• 署名作者: Mankad, Shawn; Shunko, Masha; Yu, Qiuping
• 署名单位: North Carolina State University; University of Washington; University of Washington Seattle; Georgetown University
• 刊物名称: PRODUCTION AND OPERATIONS MANAGEMENT
• ISSN/ISSBN: 1059-1478
• DOI: 10.1177/10591478261472683
• 发表日期: 2026
• 关键词: identification; Cannibalization; PERSPECTIVE; performance; diffusion; industry; models; driven; work

摘要

We study how one store's performance affects others within a large branded network. Stores influence each other both negatively, through cannibalization, and positively, via agglomeration as well as knowledge and reputation spillovers. This study causally quantifies these peer effects at the dyadic level. To address the complexity of estimating peer effects across a large number of stores, we develop a semi-parametric model that balances parsimony with the flexibility to capture heterogeneity between store pairs. To address endogeneity, we use instrumental variables based on local weather, nearby large events, and staffing policies. Using data from 792 stores of a national fast-food chain clustered around Chicago, we find a non-linear effect of distance on peer effects. On average, within 2 miles, cannibalization dominates, reducing nearby store sales by $0.154 per $1 increase at a focal store. Between 2 and 3 miles, competition and positive spillovers offset each other, yielding no significant net effect. From 3 to 7 miles, positive spillovers dominate, boosting nearby sales by $0.050 per $1, but beyond 7 miles, peer effects disappear. We also find that stores under the same ownership experience stronger positive peer effects, and urban stores benefit more from spillovers due to higher population density. Moreover, stores located near younger stores experience reduced cannibalization. Finally, we calculate each store's network contribution-the total change in network sales per $1 increase at a focal store-which ranges from -$2.763 to $9.540, with a mean of $2.691. Only 3.8% of stores exhibit negative network contributions, primarily in dense urban centers. Incorporating factors such as shared ownership, store experience, and market size further reduces these negative contributions. These findings provide actionable guidance for optimizing store network performance by amplifying positive spillovers and mitigating cannibalization.