Gaining insights for service improvement through unstructured text from online reviews

This paper aims to address the challenges of dealing with information redundancy in the context of service quality improvement by proposing an innovative online review-based service evaluation method. The proposed method identifies key service attributes that significantly influence consumer satisfaction and assist service providers in prioritizing service improvement. Specifically, this paper presents four identification dimensions and corresponding indicators of service attributes, followed by the application of Proportional Marginal Variance Decomposition (PMVD) method to determine the key service attributes related to consumer satisfaction. Taking 96,322 online reviews of 1398 hotels across four types in Beijing as the empirical validation, this paper highlights the noteworthy discrepancies in key service attributes across different hotel types and reveals the significant dimensions associated with these key service attributes. Furthermore, the calculated PMVD scores serve as valuable references for service managers in determining the ultimate improvement priorities. Overall, our findings emphasize the importance of identifying key service attribute through online review opinion mining, which help avoid wasting resources on service attributes that are repeatedly mentioned in online reviews but are not actually related to consumer satisfaction. The proposed method offers a comprehensive and effective measurement of service quality from various perspectives, providing practical insights for service providers seeking to enhance their performance.