Advances in estimating heterogeneous treatment effects enable firms to personalize marketing mix elements and target individuals at an unmatched level of granularity, but feasibility constraints limit such personalization. In practice, firms choose which unique treatments to offer and which individuals to offer these treatments with the goal of maximizing profits: we call this the coarse personalization problem. We propose a two-step solution that makes segmentation and targeting decisions in concert. First, the firm personalizes by estimating conditional average treatment effects. Second, the firm discretizes by utilizing treatment effects to choose which unique treatments to offer and who to assign to these treatments. We show that a combination of available machine learning tools for estimating heterogeneous treatment effects and a novel application of optimal transport methods provides a viable and efficient solution. With data from a large-scale field experiment for promotions management, we find that our methodology outperforms extant approaches that segment on consumer characteristics or preferences and those that only search over a prespecified grid. Using our procedure, the firm recoups over 99.5% of its expected incremental profits under fully granular personalization while offering only five unique treatments. We conclude by discussing how coarse personalization arises in other domains.

Research in the area of information acquisition has been based on two key assumptions, namely, nonsatiety (implying more is preferable to less) and transitivity (implying that if A is preferred to B and B to C then A will be preferred to C). While at first glance these assumptions do not seem altogether unjustifiable, they do, in practice, constrain the empirical application of the research models. The primary purpose of the paper is to test the validity of these two key assumptions underlying most theoretical research in information acquisition and resource allocation. We begin by introducing four partial order structures and highlight the assumptions upon which they have been based. We then discuss how these structures could be used in estimating consumer preferences. We also provide a statistical methodology to implement such estimation. In later sections, we implement the empirical model on two data sets and discuss the results and propose generalizations.