Profiling satisfied and dissatisfied hotel visitors using publicly available data from a booking platform

Abstract

We develop a set of models for predicting hotel visitor satisfaction and the probability of complaints about various service aspects. Our empirical analysis is based on 3630 reviews from one of the Dubai hotels. We identify profiles of visitors who are likely to be dissatisfied with the hotel service and need special attention, as well as of visitors, who are likely to be satisfied with the service and, therefore, do not require extra attention. The predictions are based on observable characteristics of visitors, thus making it possible for hotel managers to apply the models in their everyday work. Using content analysis we also reveal specific problems that different groups of visitors encountered and infer which of the problems has the highest impact on the overall satisfaction with the hotel.