Service Quality Using Text Mining: Measurement and Consequences

Abstract

Problem description: Measuring quality in the service industry remains a challenge. Existing methodologies are often costly and unscalable. Furthermore, understanding how elements of service quality contribute to the performance of service providers continues to be a concern in the service industry. In this paper, we address these challenges in the restaurant sector, a vital component of the service industry. Academic/practical relevance: Our work provides a scalable methodology for measuring the quality of service providers using the vast amount of text in social media. The quality metrics proposed are associated with economic outcomes for restaurants and can help predict future restaurant performance. Methodology: We use text present in online reviews on Yelp.com to identify and extract service dimensions using nonnegative matrix factorization for a large set of restaurants located in a major city in the United States. We subsequently validate these service dimensions as proxies for service quality using external data sources and a series of laboratory experiments. Finally, we use econometrics to test the relationship between these dimensions and restaurant survival as additional validation. Results: We find that our proposed service quality dimensions are scalable, match industry standards, and are correctly identified by subjects in a controlled setting. Furthermore, we show that specific service dimensions are significantly correlated with the survival of merchants, even after controlling for competition and other factors. Managerial implications: This work has implications for the strategic use of text analytics in the context of service operations, where an increasingly large text corpus is available. We discuss the benefits of this work for service providers and platforms, such as Yelp and OpenTable.