What Can Platforms Learn from Consumer Choices?

Abstract

Consumers search on a platform to learn how a product fits their preferences. Consumers’ value for the product has a common and an idiosyncratic component. The platform observes which products consumers inspect and what they eventually buy. Based on these observations on past consumer choices the platform ranks products. We find that monopoly platforms maximizing long-term consumer’s utility will first experiment with product rankings, providing future consumers with rankings that only list products that many consumers have bought before. This guarantees that consumers are more picky and only buy products they really like. The more important the idiosyncratic component, the better the platform is able to assist consumers in their search. In contrast, learning is much more restricted if platforms compete.