Building science commonly studies comfort—a subjective concept—through quantitative methods. These methods are often not the most appropriate to study subjective concepts because they struggle to consider non-quantifiable factors that are sometimes relevant in the determination of people’s comfort. Complementing the already used quantitative methods with qualitative ones can help illuminate some of the areas where the former struggle, but this is an uncommon practice in building science. This paper presents the results of a qualitative study that aimed at understanding comfort without the constraints of it being quantifiable. It introduces the feeling of comfort model, which summarises what was found. This model suggests that most building science comfort models ignore more than two out of the three factors that determine people’s comfort. Additionally, it shows that it is potentially possible to make sense of the psychology and subjectivity of comfort in an organised and structured manner. This example of how qualitative methods can be a powerful addition to building science’s comfort research. Practice relevance This paper reports a project that sought to understand an occupant-centred meaning of ‘comfort’ (constrained to what has conventionally been described as thermal, visual and acoustic domains). A qualitative model of the feeling of comfort was developed from the empirical data collated in the study. This model demonstrates that it was possible to make sense of the subjective nature of comfort in a coherent and ordered manner. This study argues that the qualitative and subjective nature of comfort should be embraced within building design and building science practice. Although this model was not meant to predict comfort quantitatively, it can serve as a theoretical basis for informing policymaking, building performance analysis and comfort research. For instance, it can help evaluate the assumptions behind building performance simulation (e.g. questioning whether people will truly open windows in certain scenarios).