Understanding consumers’ needs and wants has been a major source of success for hotel organizations. Notwithstanding, investigating the valence of online reviews and modeling hotel attributes and performance is still a rather novel approach. Using partial least squares path modeling, Swiss country-level data for online reviews from 68 online platforms, together with data from 442 hotels, we test 11 hypotheses. Our research model includes three distinctive areas of the hotel: physical aspects, quality of food and drink, and human aspects of service provision. RevPAR and occupancy are employed as performance metrics. We also test for mediation effects. Results indicate that hotel attributes, including the quality of rooms, Internet provision and building show the highest impact on hotel performance, and that positive comments have the highest impact on customer demand. This study contributes to theories of valence on hotel performance and presents salient implications for practitioners to enhance performance.