When
Where
Wenjun Zhou, Amazon Distinguished Professor, Martin Lee & Carol Fri Robinson Faculty Fellow, Haslam College of Business, University of Tennessee, Knoxville
When the Best Items Disappear First: Availability Bias in Recommender Systems
Abstract: Many digital platforms, such as crowdfunding, job, and house marketplaces, recommend items from a demand-depleted catalog, where the most appealing items attract the most demand and close first. Standard negative sampling treats closed items as rejected by later users, causing a phenomenon we call availability bias in recommender systems. This bias concentrates on the best items. Treating them as user-rejected items misguides model training and selection, leading to inverted rankings and irrelevant recommendations. We propose the active-set estimator, which scores each choice only over the reconstructed set of alternatives open at that moment. We prove that it recovers the true utilities even though closure is caused by the choices being modeled. On a large education-crowdfunding marketplace, training on the whole catalog instead of the active set costs the strongest methods about three-quarters of their accuracy. Under the correct protocol, the best-performing method captures 87% of donation dollars in a donor's top ten recommendations, whereas a pipeline that ignores availability in both training and evaluation selects a method that captures only 19%. The size of the bias follows the share of closed items in the candidate set: on a second demand-depleted platform, where sales close fewer than one listing in a hundred, it is negligible, and on a third platform, where demand sustains availability, the bias reverses sign. Our results give platform operators an estimator with a consistency guarantee, the measured cost of ignoring availability, and design principles for any market in which demand removes the most attractive items.