Changes in Service Quality of Sharing Accommodation: Evidence from Airbnb

Abstract

As consumers' perceptions of quality change over time, service providers should track the dynamic changes in service quality and adjust their services to adapt to these changes. In the context of sharing accommodation, although many studies have focused on evaluating service quality, the dynamic changes in service quality over time remain unexplored. This study aims to explore the dynamic changes in the service quality of sharing accommodation. We propose a novel framework that uses deep learning and SERVQUAL to analyze online reviews of sharing accommodation services, thus contributing to the literature from a dynamic perspective. Using a 10-year-span longitudinal dataset of Airbnb's online reviews of San Francisco's listings (n1 = 366,643), we construct a weakly supervised topic model that extracts service quality topics from online reviews and then classifies reviews into irrelevant-topics and relevant-topics. Each relevant-topic review is mapped to one of the SERVQUAL dimensions, combined with its sentiment analysis score, which constitutes the output of text mining. We then analyze the dynamic changes in the service quality. The results show that both the overall service quality and that in each dimension of SERVQUAL exhibit a slight downward trend. We obtain the similar results for the Beijing longitudinal dataset (n2 = 251,081), which confirms that the downward trend in service quality is not unique to San Francisco. We discuss the reasons for this trend and provide managerial guidance for the platform and its hosts.