In the construction sector, strong efforts are being made to reduce energy consumption in buildings, particularly regarding thermal insulation in northern countries. However, the objectives set for the thermal insulation performance of the building envelopes are rarely achieved in practice. For this reason, there is increasing interest in taking onsite measurements. This paper presents the new achievements obtained by the ISABELE (In Situ Assessment of the Building EnveLope pErformances) method for data treatment to evaluate the global transmission heat transfer coefficient of a building with a consolidated uncertainty and within a short unoccupied period. Most of the work relates to the quantification and propagation of systematic errors and to the adaptation of the thermal model used for the inverse method. Practical applications on a test cell and a demonstration on a low-energy house in real outdoor conditions are presented, as are comparisons with other measurement methods (coheating, Quick U-Value of Buildings (QUB)) and the resulting calculated parameters, which show coherent outcomes. The test cell results reveal that a stabilized, repeatable measure with an acceptable uncertainty range (+/- 10 % in average) can be obtained within 2 days. The house results are slightly less precise (+/- 15-20 %) but can be obtained just as quickly, assuming that both configurations are insulated from the inside. (C) 2018 Elsevier B.V. All rights reserved.