In digital markets with peer-to-peer reviews, new products encounter the so-called “coldstart” problem: Little-known products are bought too rarely and remain little known. While consumers benefit from observing reviews written by others, they do not account for the benefit they generate from trying a new product and revealing the product’s quality to everyone else. In this paper, we examine the inefficiency linked to social learning on Airbnb, including its implications for hosts’ price, entry and exit decisions. We estimate a dynamic structural model of demand and supply for Airbnbs in Manhattan, New York, from 2016 to 2019. We then introduce a counterfactual tax-subsidy scheme aimed at changing the relative price of listings with few compared to many reviews, thereby shifting demand and addressing the cold-start problem. In our main counterfactual, we find that the average price decreases by around 17% for listings with at most one review, but increases by almost 10% for listings with more than 15 reviews, which leads to 14% higher demand for the former. Furthermore, the number of listings with at most one review declines by 18%, although the total number of listings increases by 5%. Based on our conservative estimates, widening the price gap between new and established listings leads to a welfare increase amounting to roughly 8.5% of total host revenue on Airbnb in Manhattan.