Improving Match Rates in Dating Markets Through Assortment Optimization

Abstract

Problem definition: Motivated by our collaboration with an online dating company, we study how a platform should dynamically select the set of potential partners to show to each user in each period in order to maximize the expected number of matches in a time horizon, where a match is formed only after two users like each other, possibly in different periods. Academic/practical relevance: Increasing match rates is a prevalent objective of online platforms. We provide insights into how to leverage users’ preferences and behavior toward this end. Our proposed algorithm was piloted by our collaborator, a major online dating company in the United States. Methodology: Our work combines several methodologies. We model the platform’s problem as a dynamic optimization problem. We use econometric tools and exploit a change in the company’s algorithm in order to estimate the users’ preferences and the causal effect of previous matches on the like behavior of users, as well as other parameters of interest. Leveraging our data findings, we propose a family of heuristics to solve the platform’s problem and use simulations and field experiments to assess their benefits. Results: We find that the number of matches obtained in the recent past has a negative effect on the like behavior of users. We propose a family of heuristics to decide the profiles to show to each user on each day that accounts for this finding. Two field experiments show that our algorithm yields at least 27% more matches relative to our industry partner’s algorithm. Managerial implications: Our results highlight the importance of correctly accounting for the preferences, behavior, and activity metrics of users on both ends of a transaction to improve the operational efficiency of matching platforms. In addition, we propose a novel identification strategy to measure the effect of previous matches on the users’ preferences in a two-sided matching market, the result of which is leveraged by our algorithm. Our methodology may also be applied to online matching platforms in other domains. History: This paper has been accepted as part of the 2021 Manufacturing & Service Operations Management Practice-Based Research Competition. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2022.1107 .