Many platforms operating online marketplaces sell their own products alongside those of third-party sellers, and may use marketplace data on the demand for third-party products to inform the launch of their own products. To evaluate the anti-competitive implications of this practice, we model how a platform who can commit to a product introduction policy optimally uses marketplace data. An optimal policy trades off the ex post profitability of imitating successful third-party products against the ex ante reduction in innovation caused by imitation. We find that a regulation which bans the sharing of marketplace data stimulates innovation for “experimental” products with significant upside demand potential, but stifles it for “incremental” products with little upside potential.