Executive Summary

When planning and implementing their price-promotions strategy, retail chain managers face the typical dilemma of “thinking globally, but acting locally.” In other words, they must plan their strategy keeping in mind the global chain-level impact of their promotions, to deliver on the commitments made to manufacturers. At the same time, managers need to make sure that the implementation of such strategy takes into account the fact that each store caters to a different market with different needs and responses to marketing programs. Moreover, the retail chain manager must consider not only how the promotion of a brand affects competing brands and total category sales, but also how it could affect sales in other categories.

In this study, we propose a methodology that addresses these two important aspects of chain-wide and store-level cross-category analysis.  First, our proposed factor regression model takes store differences and longitudinal market shifts into account, thereby providing the retail chain manager with unbiased global, chain-level estimates. It also provides stable local estimates of cross-category promotion effects at the store level, taking advantage of all available data, not only for the store in question but from all other stores in the retail chain.  Second, while allowing this flexibility, our methodology is parsimonious enough relative to existing alternatives, making it particularly useful for chain-wide and store-level cross-category analysis. 

Aside from providing retail-chain managers with chain-wide and store-level estimates of price-promotion effects within and across product categories, our proposed factor-regression model produces a graphical portrayal of sales-promotion response across brands, product categories and stores.  The proposed model produces a cross-elasticity map which shows how sales response to price-promotion by one brand in one category correlates with response to promotions by competing brands in the same and in related product categories.  A mapping of individual stores in the same cross-elasticity map shows stores that have higher-than-average sales response to the price-promotions of each brand, helping the manager identify stores where the price-promotion of a brand in one category is likely to have the most sales impact. 

To illustrate these features of our proposed methodology, we apply it to store-level data from one retail chain, comparing it with competing approaches.  Our empirical results demonstrate that this methodology provides the best balance between flexibility and parsimony.  Most importantly, we show that the proposed model provides useful insights regarding cross-category effects at the chain-level, for individual stores, and their patterns across stores.   For example, our results show that even though the two product categories we studied are not necessarily complementary in purchase (they are complementary in consumption), there are clear cross-category price-promotion effects for the leading brands in both product categories; when a leading brand promotes in one category, there is an increase in sales for the same brand in the other product category, suggesting that the two categories are complementary in purchase.   One unexpected result was the strong positive sales response by the store brand in response to a price-promotion by the leading brands in the same category, leading to an increase in category sales.  We suspected that this negative price-promotion cross-elasticity might be due to simultaneous promotions by the private label, but did not see strong price correlations between leading national brands and the store brand.

Even though we applied our model to two product categories, it is applicable to cross-category analyses involving multiple product categories.  In fact, the parsimony accrued by our factor-analytic framework will be greater as the number of brands and categories increase, while competing approaches become even less feasible.