We empirically estimate the expected basket-level demand effects and, therefore, the expected store profit effects, of three different types of retailer targeted coupon campaigns used by a national supermarket retail chain. The three types of campaigns differ in terms of household-level customization and personalization. One campaign employs a one-to-one approach by targeting a set of selected households with a unique bundle of coupons customized to that household’s preferences (high customization). Another promotes a bundle of brands around a theme (i.e., back-to-school, new baby, etc.) relevant to the household, but where all households receiving the promotion receive the same set of coupons (medium customization). The third provides a limited number of coupons for one brand in a single category relevant to the consumer, but where again all households targeted with the promotion receive the same promotional materials and coupons (low customization).
To accomplish the above, we build and estimate an econometric model of a household’s contemporaneous purchase incidence outcomes in 28 product categories, together with a household-level store choice model. Our basket-level demand model captures pair-wise cross-category dependencies in purchase incidence outcomes of a household in a flexible manner as a function of exposure to the three retailer promotion types described above. Such dependencies have not been modeled across such a large number of categories in previous research and doing so allows us to measure the effect of each of the three targeted campaign types on expected retailer profit after correctly accounting for cross-category spillovers within a basket of categories. We estimate our proposed multi-category purchase incidence model on purchase data of 800 households over 102 weeks, obtained from a retailer’s loyalty card database.
The results from our study provide insights to retailers about the value of investing in more customized promotional efforts, and with a detailed cross-category glimpse of where such value is gained.
This research attempts to advance the literature on multi-category demand models by modeling cross-category dependencies in households’ purchase incidence outcomes. Such dependencies are of enormous practical interest to retailers in managing shelf placements of products and co-promotion decisions across categories. These dependencies are currently estimated by practitioners using data mining techniques (such as affinity analysis). Marketing researchers have employed econometric techniques, which can simultaneously model the impact of marketing variables and unobserved heterogeneity across households, for the same purpose. Two such econometric models are the Multivariate Probit Model (Chib and Greenberg 1998, Chib, Seetharaman and Strijnev 2002) and the Multivariate Logit Model (Russell and Peterson 2000, Niraj, Padmanabhan and Seetharaman 2008). One limitation of these econometric models is that they incorporate cross-category dependencies only at the pair-wise level. However, such dependencies can be expected to manifest at higher order (third-order, fourth-order etc.) as well. The goal of this research is to explicitly model such higher order dependencies in households’ cross-category purchase incidence outcomes. We propose an extended version of the Multivariate Logit model of Russell and Peterson (2000) that enables the estimation of such higher-order cross-category dependencies. We investigate the relative magnitudes of such higher-order effects relative to the second-order effects, as well as the methodological and substantive consequences of ignoring such higher-order effects in the model.(Artwork by Gunilla Klingberg: “WHEEL OF EVERYDAY LIFE”)
We develop an analytical model describing the price negotiations between a manufacturer and a customer in cyclical business to business environments. Following a Generalized Nash Bargaining methodology our model allows for the explicit estimation of the manufacturer “bargaining power” for each dyadic interaction. This information can be used to develop an optimal personalized pricing strategy for each customer at each point in time.
We estimate our model on a unique dataset that was provided by a multi-national Chinese manufacturer of oil & gas equipment and represents 4 years of monthly mud pump sales in the United States. In addition to the sales data we obtain complete quote data from the negotiation process. These are crucial for the estimation of the “bargaining power” for each customer interaction.
Our dataset has sales prices for 2008, 2010-2012 spanning before and after the financial crisis in 2009. This change in demand provides us with a natural experiment to test our model when estimating the shift of “bargaining power“ from the manufacturer to the buyer. In figure 1 the demand shock is clearly marked by a sharp decline in the drilling rig count. In this study we use weekly rig count data (oil and natural gas drilling) as a proxy for the aggregate mud pump demand as every drilling rig is equipped with two mud pumps. The average rig count in the US between the years 1998-2004 was 700 but during 2005-2012 it more than doubled to 1500 (US Dept. of Energy).