Thermal building simulations serve to optimize the energy consumption of a building during its design phase. Because modern building design offers occupants increasing opportunities to interact with the building in order to adjust internal conditions according to their individual needs, the prediction of these interactions needs to be included in optimization calculations. The way to an adequate consideration of human interaction in building simulations requires that two questions be addressed: First, which contextual factors are impactful to this process, and how are they related to each other? Second, which methods are best suited to predicting energy-relevant interaction based on the knowledge of these contextual factors? This article addresses the first question through the development of a data acquisition and analysis method. The technique is based on “protocol analysis,” which is a method originally developed to generate knowledge about the cognitive processes of humans during task implementation. The proposed method is applied in a pilot-study, which produced valuable, qualitative insights into the nature of the interaction and its relation to context. The results serve to address the second question in that they point to psychological theories that may prove significant for the future conceptualization of a cognitive theory of energy-relevant interaction.