Nowadays, understanding the actual performance of building components is one of the key factor to achieve energy savings. For this reason, on-site measurements are essential but the boundary conditions during surveys can affect the final results. This can occur during heat flow meter measurements, when the thermal transmittance value of a wall can be influenced by disturbing factors, such as the heating system power-on and off. Due to this, the aim of this study is to investigate the influence of these disturbing factors, moving away from steady-state conditions. This research is divided in two main steps: a first critical analysis of data obtained by in-situ measurements and an investigation of how the mentioned disturbing factors can affect the final results, employing a FEM code, where stationary conditions are not respected; a second analysis related to the data post-processing procedures, proposing a new supplementary approach able to exclude heat flow distortions and able to obtain measured U-values closer to the calculated ones, according to ISO 6946. Starting from simulations and on-site measurements, the proposed method was preliminary validated, analyzing actual case studies characterized by heating systems with radiators and obtaining preliminary satisfying results. The simulations allowed to assess a reduction in the difference between the measured and the calculated U-value that goes from +22.1% to +0.7%. Post processing of experimental data with the proposed methodology allowed to significantly reduce the difference between measured and calculated U-values (from +36.9% to −7.6% in the best case study). Starting from the preliminary results, the proposed approach seems to be promising with U-value corrections in accordance with the theoretical ones.