Disentangling Guest Satisfaction in a Hotel Chain: Frequency Patterns and Semantic Networks in Online Reviews from Cinnamon Hotels in Sri Lanka

Guest satisfaction is a critical performance indicator in the hotel industry; however, existing research in Sri Lanka has predominantly emphasized internal operational practices and staff perspectives, offering limited insight into guests’ lived experiences. Addressing this gap, the present study investigates guest satisfaction at Cinnamon Hotels in Sri Lanka through a large-scale analysis of online customer reviews. Drawing on electronic word-of-mouth (eWOM) data, a total of 23,350 Google Reviews from 11 Cinnamon Hotels were collected and analyzed to capture authentic and unsolicited guest evaluations. A text mining approach was employed, combining word frequency analysis with semantic network analysis to identify both dominant attributes and their relational meaning structures within guest narratives. Following rigorous data preprocessing in the R programming environment, high frequency keywords were extracted and transformed into a co-occurrence network, which was subsequently analyzed using CONCOR analysis in UCINET. The results reveal four distinct yet interconnected dimensions of guest satisfaction: Location, Gastronomic Experience, Service Interaction, and Holistic Stay Experience. Among these, food and beverage experiences and interpersonal service interactions emerged as particularly central in structuring guests’ overall evaluations. By moving beyond isolated word counts to uncover the semantic organization of guest experiences, this study contributes a guest centric, data-driven perspective to hospitality literature from an underexplored South Asian context. The findings offer practical insights for hotel managers seeking to enhance experiential quality and strengthen customer satisfaction through targeted service strategies.