“People Who Liked This Study Also Liked”: An Empirical Investigation of the Impact of Recommender Systems on Sales Volume and Diversity

We investigate the impact of collaborative filtering recommender algorithms (e.g., Amazon.com’s “Customers who bought this item also bought”), commonly used in e-commerce, on sales volume and diversity. We use data from a randomized field experiment on movie sales run by a top retailer in North America. For sales volume, we show that different algorithms have differential impacts. Purchase-based collaborative filtering (“Customers who bought this item also bought”) causes a 25% lift in views and a 35% lift in the number of items purchased over the control group (no recommender). In contrast, View-based collaborative filtering (“Customers who viewed this item also viewed”) shows only a 3% lift in views and a 9% lift in the number of items purchased, albeit not statistically significant. For sales diversity, we find that collaborative filtering algorithms cause individuals to discover and purchase a greater variety of products but push users to the same set of titles, leading to concentration bias at the aggregate level. We show that this differential impact on individual versus aggregate diversity is caused by users exploring into only a few ’pathway’ popular genres. That is, the recommenders were more effective in aiding discovery for a few popular genres rather than uniformly aiding discovery in all genres. For managers, our results inform personalization and recommender strategy in e-commerce. From an academic standpoint, we provide the first empirical evidence from a randomized field experiment to help reconcile opposing views on the impact of recommenders on sales diversity.