Service quality dimensions in digital guest reviews: evidence-based insights for African hospitality management

This study examines how service quality dimensions expressed in digital guest reviews are associated with guest satisfaction in African hospitality markets. Drawing on SERVQUAL and using the expectation-confirmation theory (ECT) as an interpretive lens, the study analyses a secondary TripAdvisor hotel-review corpus associated with Alam et al. (2016). The original corpus contained 20 186 English-language reviews; after filtering the property-level country metadata to retain hotels located in Morocco, Kenya, Tanzania, South Africa and Zimbabwe, the final African analytical sample consisted of 15 139 reviews. Exploratory latent Dirichlet allocation (LDA) topic modelling showed that unrestricted topics were dominated by geographic and destination-experience terms; therefore, the final service quality indicators were operationalised through a SERVQUAL-informed aspect dictionary and examined through descriptive mention analysis, partial correlations, multiple regression and multicollinearity diagnostics. Six service quality dimensions produced estimable variation in the final sample. Regression results indicate that staff service is the strongest positive predictor of guest satisfaction (β = 0.218, p < 0.001), while value-for-money mentions (β = –0.132, p < 0.001) and cleanliness-related mentions (β = –0.350, p < 0.001) are associated with lower ratings, suggesting that these dimensions operate primarily as dissatisfaction-sensitive signals in the review corpus. VIF values ranged from 1.007 to 1.159, indicating that multicollinearity was not a concern. The recalculated results do not show a substantive sign reversal for cleanliness or room quality. The study contributes to digital hospitality analytics by showing how SERVQUAL-informed aspect coding can convert unstructured review data into actionable evidence for service improvement in African hotel markets.