Abstract Households’ heating energy consumption is one of the most important sources of a nation's total CO2 emissions. But there remains uncertainty about the impact of occupant behaviour on heating energy consumption, especially in multi-apartment buildings. In bringing together apartment specific data of occupant behaviour and physical building properties of multi-apartment buildings, the quantitative influence of different factors determining apartments’ heating energy consumption can be separated by means of OLS-regression. For multi-apartment buildings the heating energy consumption of a single apartment seems to be determined not only by the behaviour of households in opening windows and using radiators but also by internal heat transfers, triggered by temperature differences between neighbouring apartments. This study shows some limits of refurbishment policies and consumer education and points to the importance of considering both data on occupant behaviour and building physics at the apartment level together to understand households’ heating energy consumption.