An Online Review Analytics for Quality Evaluation and Diagnosis in Hotel Services: A Perspective of Benchmarking through MTGS

This study developed a customer perception-oriented and benchmarking method for firm-level quality evaluation and diagnosis in hotel services. Text mining techniques and Latent Dirichlet Allocation (LDA) were used to extract customer-concerned hotel features from online reviews. The performance of individual hotels on these hotel features was then quantified and vectorized from reviews. Mahalanobis–Taguchi–Gram–Schmidt System (MTGS) was employed to construct a quality measurement scale with a benchmark base for hotel quality evaluation and diagnosis and to assess the importance of each feature in quality discrimination. By the measurement scale, the overall quality evaluation and the directional quality diagnosis of hotel features can be achieved for a particular hotel. The results indicate how much each feature performs better or worse than the benchmark. By considering the feature importance, the diagnosed hotel can identify its relative strengths and weaknesses and prioritize improvements accordingly. The data from Booking.com were used to show the effectiveness of the proposed method.