This study empirically estimates the expected basket-level demand effects, as well as the expected store profit effects, of three different types of retailer targeted promotions varying in their customization level (high, medium, low customization). Using data from a national grocery store retailer that targets households with different types of promotions, we build and estimate an econometric model of a household’s contemporaneous purchase incidence outcomes in 28 frequently shopped product categories. Estimating the cross-category dependencies in purchase incidence as a function of exposure to different levels of customized promotions, allows us to measure the effect of each campaign type on expected retailer profit and implement prescriptive analytics to identify the appropriate multi-level coupon mix for maximizing store profits. The findings reveal that all three levels of coupon customization result in per-customer returns, but that medium customization leads to the highest incremental expected profit, while high customization generates the highest expected profit. The results provide insights to retailers about the values of investing in more customized promotional efforts, with a detailed cross- category perspective into where such value is gained.
In sum, our research offers retailers both a computationally tractable predictive model to forecast demand in a large number of categories, as well as a prescriptive analytics solution to target customers with one-to-one coupon bundles that maximize retail profit.