In an attempt to conceptualise complex and multi-dimensional constructs in tourism and hospitality contexts, this paper proposes a combined-learning approach that incorporates machine and human learning techniques in analysing online reviews. The application of machine learning in online review analytics in tourism and hospitality is synthesised to address the existing gaps in the literature. The role of machine learning is also critically evaluated, and thereby proposing the complementary role of employing conventional research techniques to overcome drawbacks imposed by machine learning. This paper therefore offers crucial theoretical and methodological implications of employing combined-learning approaches in analysing online review texts to better understand complex consumer-based constructs. This paper also calls for greater attention to more rigorous and systematised procedures of processing and analysing user-generated content in tourism and hospitality.