In this paper, we empirically study whether firms improve their quality based on reviews left by their customers in a dynamic quality environment. We do so by analyzing the US hotel industry using data from two major online review platforms: TripAdvisor and Expedia. Using management response as a proxy for whether hotels read and listen to consumer reviews, and a difference-in-differences strategy, we demonstrate that hotels make improvements in quality by paying attention to the reviews they receive. Moreover, we show that these improvements are primarily made by low-rated hotels that have more room for improvement, and by chain hotels likely because of their lower operational marginal cost. To pin down the underlying mechanism, we analyze the text of reviews and responses using novel tools for natural language processing, and show that: (i) hotels that listen to customers improve on issues that are frequently mentioned in their reviews, and (ii) hotels that use canned responses are less likely to see improvements in quality. Overall, our results suggest that online user-generated reviews form a feedback mechanism through which consumers make themselves heard by businesses and contribute to changes in the quality of those businesses.