We study the incentives of a streaming platform to bias consumption when products are vertically differentiated. The platform offers mixed bundles of content to monetize consumers’ interest in variety and pays royalties to sellers based on the effective consumption of the content they produce. When products are not vertically differentiated, the platform has no incentive to bias consumption in equilibrium: the platform being active represents a Pareto-improvement compared to the case in which she is not. With vertical differentiation, royalties can differ; the platform always biases recommendations in favor of the cheapest content, which hurts consumers and the high-quality seller. Biased recommendation always diminishes the incentives of a seller to increase the quality of her content for a given demand. If a significant share of the users is ex-ante unaware of the existence of the sellers the platform can bias recommendations more freely, but joining the platform encourages investment in quality. The bias, however, can lead to inefficient allocation of R&D efforts. From a policy perspective, we propose this as a novel rationale for regulating algorithmic recommendations in streaming platforms.
Gambato, Jacopo und Luca Sandrini (2023), Not as Good as it Used to Be: Do Streaming Platforms Penalize Quality?, ZEW Discussion Paper Nr. 23-032, Mannheim.