Diverse Assortments in Online Recommendations

This paper studies the business case for incorporating diverse content in recommendation systems. As online platforms and marketplaces have grown in recent years, they have increasingly relied on recommendation and ranking algorithms. These algorithms promote popular content, leading to a “rich get richer” effect. On the surface, this appears to be a favorable outcome for the platform as it would improve user engagement in the short term; however, undermining diverse (and new) items may affect the long-term health of the marketplace. In this paper, we make a connection between customer engagement and satiation. Satiation examines how past consumption patterns determine consumers’ willingness to pay for goods/services. In this setting, we find conditions on parameters under which the optimal policy always alternates and favors a more diverse recommendation. Under less restrictive conditions, the optimal policy favors diversity in earlier periods, and after some time, the optimal policy keeps offering the product with the largest value. This implies that a more diverse recommendation set is preferred over longer time horizons. Moroever, we show that we have a satiation threshold policy. This threshold assigns more area to the myopically better action as time passes and we get closer to the end of the horizon.

Recommended citation: Hosseini, M., Sekar, S., Baron, O., Malekian, A. (2022). Diverse Assortments in Online Recommendations. Rotman School of Management Working Paper, (3774324).