Integrating Text Mining and Servqual Framework to Analyze Customer Feedback for Service Quality Improvement in the Hospitality Industry

Abstract

Service quality is a critical factor in the hospitality industry, where customer satisfaction directly influences business performance. With the growing use of online review platforms like TripAdvisor, vast amounts of customer feedback are available for analysis. This study aims to develop a data-driven framework that integrates text mining techniques with the SERVQUAL model to evaluate and improve service quality in the hospitality sector. Using 20,491 TripAdvisor reviews from Kimpton Hotel Monaco Seattle, the study applies a multi-stage methodology: data preprocessing, sentiment analysis using DistilBERT, topic modeling via Latent Dirichlet Allocation (LDA), keyword extraction with TF-IDF, and SERVQUAL dimension mapping using BERT embeddings and cosine similarity. The SERVQUAL model’s five dimensions Tangibles, Reliability, Responsiveness, Assurance, and Empathy are used to categorize and interpret customer concerns. The results show that Tangibles and Responsiveness are the most frequently mentioned dimensions, with both positive and negative sentiments, suggesting these areas are key to customer satisfaction. In contrast, Empathy and Assurance are less discussed, possibly indicating satisfactory performance or lower customer expectations. This research contributes to Industrial Engineering by combining natural language processing (NLP) with quality management frameworks to support data-driven service improvements. The proposed framework enables hospitality managers to identify service gaps, prioritize improvements, and make informed operational decisions based on customer sentiment.