“Selling on Platforms : Demand Boost versus Customer Migration”

abstract :
We study sellers of non-durable goods or services that can sell directly to consumers or via an online marketplace (platform) over multiple periods. We analyze how the quality of the platform’s recommendation algorithm as well as the platform’s market power affect market outcomes, in particular seller and consumer surplus. We find that the number of sellers joining the platform, a key parameter influencing the quality of recommendations that consumers receive, first increases and then decreases in the quality of the platform’s recommendation algorithm. As a result, consumer surplus is inverse U-shaped in said quality.