Where hearts return in Yunnan: How UGC-driven homestay attractiveness shapes social evaluation

Homestays are increasingly marketed as locally embedded experience products, yet their attractiveness is often explained through destination- or hotel-derived attribute frameworks. This study contextually identifies and examines a user-generated content (UGC)-driven model of homestay attractiveness in Yunnan, China, within the cognition-affect-conation framework. Study 1 analyses 30,076 Ctrip reviews through latent Dirichlet allocation topic modelling, TF-IDF analysis, and grounded-theory coding, identifying seven dimensions of homestay attractiveness. Study 2 uses 307 valid questionnaires and partial least squares structural equation modelling (PLS-SEM) to assess the measurement and structural models. The findings show that Accessibility, Leisure Activity, Landscape Image, and Room Physical and Aesthetic Attractiveness significantly enhance Service and Affective Attractiveness, which in turn improves social evaluation attractiveness. Service and Affective Attractiveness also mediates the effects of these four cognitive dimensions on social evaluation attractiveness. Safety operational process management attractiveness has no significant direct or indirect effect, suggesting that it mainly functions as a basic assurance attribute in highly rated, frequently reviewed, and highly visible Yunnan homestays. This study contributes a context-sensitive explanation of homestay attractiveness and clarifies how UGC-derived attractiveness cues can inform market positioning, platform communication, experience design, and online reputation management.