Targeted Marketing with Large Batches

Adaptive learning policies that guide how firms trade off acquiring new information to improve a current targeting policy, versus exploiting the current policy to harvest, typically focus on settings in which customers arrive individually, in a frequent sequence. However, in practice, firms often conduct marketing campaigns in batches, in which they target a large group of customers with personalized marketing actions together. This has an important implication for how firms resolve the tradeoff between acquiring new information and exploiting the current policy. The large number of customers in each batch (campaign) introduces an information externality: the incremental information contributed by a single customer depends upon the assignment decisions for other customers in the batch. We investigate how to optimally acquire and coordinate information in these settings. The algorithm we propose uses Gaussian Processes to estimate the value of incremental information, while accounting for the information externality between customers in the same batch. We validate our findings using data from a field experiment.