Lore: A large-scale ofer recommendation engine with eligibility and capacity constraints

Abstract

Businesses, such as Amazon, department store chains, home furnishing store chains, Uber, and Lyft, frequently ofer deals, product discounts and incentives to drive sales, increase new product acceptance and engage with users. In order to appeal to diverse user groups, these businesses typically design more than one promotion ofer but market diferent ones to diferent users. For instance, Uber ofers a percentage discount in the rides to some users and a low fxed price to others. In this paper, we propose solutions to optimally recommend promotions and items to maximize user conversion constrained by user eligibility and item or ofer capacity (limited quantity of items or ofers) simultaneously. We achieve this through an ofer recommendation model based on Min-Cost Flow network optimization, which enables us to satisfy the constraints within the optimization itself and solve it in polynomial time. We present two approaches that can be used in various settings: single period solution and sequential time period ofering. We evaluate these approaches against competing methods using counterfactual evaluation in ofine mode. We also discuss three practical aspects that may afect the online performance of constrained optimization: capacity determination, trafc arrival pattern and clustering for large scale setting.