I Hear You: Does Quality Improve with Customer Voice?

Abstract

In a static quality environment, online customer reviews help consumers separate high- and low-quality firms. However, in a dynamic quality environment, online reviews can help firms---in particular low-quality ones---to improve over time, reducing the gap between low- and high-quality firms. In this paper, we empirically test this hypothesis by analyzing the U.S. hotel industry using data from two major online 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 find that hotels that pay attention to reviews increase their ratings more than hotels that do not do so. Moreover, we show that low-rated hotels experience larger gains as they have more margin for improvement than high-rated hotels. Finally, to explain the mechanism driving the increase in ratings, we use natural language processing tools to show that (i) hotels that pay attention to reviews improve on issues that are frequently mentioned in their reviews, and (ii) hotels that write more canned responses and therefore are less likely to have paid attention to reviews experience smaller improvement in ratings.

Keywords: Online reviews, Reputation management, TripAdvisor, Expedia, Hotel industry, Sentiment analysis, Sentence embeddings