Researchers have developed a new algorithm called Matching for Retention (MRet) to optimize user retention on two-sided matching platforms. Unlike previous methods that focused on maximizing the number of matches or ensuring fairness, MRet directly models and learns personalized retention curves for each user. This approach dynamically adapts recommendations by considering the retention gains for both parties involved in a potential match, aiming to allocate limited matching opportunities more effectively to improve overall user retention. Empirical evaluations on synthetic and real-world data from an online dating platform demonstrated MRet's superior performance in achieving higher user retention compared to conventional algorithms. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT Introduces a novel algorithm for optimizing user retention on matching platforms, potentially improving engagement and revenue for services like dating apps.
RANK_REASON This is a research paper introducing a new algorithm for two-sided matching platforms.