Product Ranking on Online Platforms

On online platforms, consumers face an abundance of options that are displayed in the form a position ranking. Only products placed first few positions readily accessible to consumer, and she needs exert effort access more options. For such we develop two-stage sequential search model where, stage, consumer sequentially screens observe preference weight them forms consideration set. In second observes additional idiosyncratic utility can derive from each product chooses highest-utility within her this model, characterize optimal policy welfare-maximizing consumer. We then study how platforms with different objectives should rank products. focus on two objectives: (i) maximizing platform’s market share (ii) consumer’s welfare. Somewhat surprisingly, show ranking decreasing order their weights does not necessarily maximize or Such may shorten set due externality effect high-positioned low-positioned ones, leading insufficient screening. both problems—maximizing welfare—are NP-complete. novel near-optimal polynomial-time algorithms for objective. Further, that, even though is suboptimal, enjoys strong performance guarantees objectives. complement our theoretical developments numerical studies using synthetic data, which (1) heuristic versions do rely primitives perform well (2) be effectively estimated maximum likelihood estimator. This paper was accepted by Gabriel Weintraub, revenue management analytics.