Designing Pandora's Box: Product Rankings and Marketplace Participation

Abstract

With the rapid growth of e-commerce, product rankings have emerged as a primary lever used by two-sided platforms to match supply and demand. Although product rankings have been studied from a variety of perspectives, existing research typically overlooks the interaction between ranking policies and seller behavior. In this paper, we present a model in which both sides of the market respond strategically to the platform's ranking policy: on the supply side, sellers participate only if their earnings exceed the outside option, while on the demand side, consumers adjust their search behavior to maximize their utility. We illustrate that in such settings, greedy policies that simply direct consumers to the sellers that best match their preferences are often suboptimal due to poor marketplace participation. To reduce this inefficiency, we develop a new optimization framework and propose an approach for designing randomized ranking policies which are scalable, incentive-compatible, and near-optimal in many practically relevant cases.