To ensure good thermal performance is delivered consistently and at scale, there is a need to measure and understand the as-built heat loss of dwellings. Co-heating is a steady state, linear regression method, used to measure whole building heat transfer coefficients. This paper assesses the uncertainties in such outdoor, in situ, measurements due to the presence and treatment of solar gains. Uncertainties relating to solar gains are explored through both a number of field test results and simulated co-heating tests. Results demonstrate the potential for fractions of solar gains received on one day to be re-emitted on subsequent days. This dynamic behaviour can lead the steady state analysis to underestimate heat loss. Furthermore, inappropriate measurements of on-site solar radiation are shown to lead to bias in heat lOoss measurements. In particular, horizontal on-site solar radiation measurements are shown to significantly overestimate heat loss in buildings experiencing high proportions of direct gains through vertical openings. Both forms of uncertainty are dependent upon both the environmental test conditions and the characteristics of a test dwelling. Highly glazed, low heat loss and heavyweight buildings prove to be the most susceptible to such uncertainties, which ultimately limit both when tests can be successfully performed and which buildings can be tested. (C) 2017 The Authors. Published by Elsevier E.V.