EPCs in Sweden are based solely on the measured energy use in the building. The regulation requires that the measured energy use is corrected to normal use during a reference year, however, there is no standardised methodology to account for normal use in the EPC and this is therefore seldom done. This study is focused on investigating the cause of differences between the calculated and measured energy use in buildings, in order to detect compliance problems with measured EPCs. Analysis of two EPC databases for single-family houses and multi-family buildings is conducted based on interviews and complemented with energy use calculations, in order to identify and investigate potential general procedures and parameters that cause such differences. Suggestions on improvements to the Swedish EPC scheme are also presented.