Optimal Information Design of Online Marketplaces

Customer data increasingly enables online marketplace to identify buyers' preferences and provide individualized product information. I characterize the impact of such ex-ante information on buyer surplus and seller surplus. In my model, a marketplace provides an ex-ante signal about buyers' match value for a product. Buyers fully learn their product value only after contracting. The seller sets prices and refund conditions in response to the ex-ante information. I show that efficient trade and an arbitrary split of the surplus can be achieved. For the buyer optimal signal buyers remain partially uninformed, since such an information structure induces the seller to set low prices