Price customization and within-chain data do not mix!

Abstract

Purpose – Grocery retailers have access to detailed data on consumer purchases within their own chains. Previous research has used across-chain scanner panel data to develop optimal price cuts targeted to individual households but whether such a targeting strategy will work with only within-chain data is unknown. The purpose of this research is to address this specific question. Design/methodology/approach – The authors use scanner panel data from multiple categories to create across-chain and within-chain purchase histories for the same consumers. They then estimate models of purchase decisions on the two datasets and compare their performance. Findings – Within-chain data fares significantly worse on both fit and prediction criteria. Retailers' upside to customizing is minimal compared to those reported for manufacturers. Finally, customized prices based on the within-chain model significantly underperform the promise of across-chain data. Research limitations/implications – Store choice is not modelled. Research also needs to be replicated in other contexts. The authors conclude that limited purchase histories may not yield accurate enough estimates of marketing mix responsiveness, and that across-chain purchase histories are essential for effective targeted price cuts. Practical implications – Loyalty card data may be useful for other purposes, like experimenting with segment-specific discounts, but its value in customizing prices at individual level is limited without adding other sources of information. Originality/value – Previous research on price customization has been based almost exclusively on across-store data. However, retailers only have access to their own chain-specific data. This is the first study to comprehensively compare price customization based on within- and across-chain purchase data and show that the upside potential for price customization based on the former information set is quite limited.