Selling on Recommender Platforms: Demand Boost versus Customer Migration

Résumé

We develop a model to study the pricing and welfare implications of recom- mender platforms—i.e., platforms that provide recommendations about products that match a consumer’s preference. Sellers decide whether to be active on the platform, and consumers can buy via seller’s direct channel or via the platform. Consumers can also search for additional products on the platform. We find novel trade-offs that occur because better recommendations do not only lead to better matches between consumers and sellers, but also reduce the number of consumers buying via the seller’s direct channel. The effect of better recommendations on product prices depends on the platform’s commission fee, which sellers pay when- ever consumers buy via the platform. We find that the price level decreases as recommendations improve if search costs are high (in which case the price level is already high), and increases if search cots are small. We also show that regulating data collection by the platform to reduce the quality of recommendations raises consumer surplus if search costs are small, but can lower consumer surplus other- wise. Finally, we identify adverse effects of capping platform fees and of facilitating dis-intermediation.