Building integrated photovoltaic (BIPV) is a key concept for the realisation of sustainable buildings. Despite the progress in BIPV modelling, the use of sensitivity analysis (SA) is still scarce in the BIPV literature. SA can help the modeller to identify which model inputs influence the model outputs the most. This paper presents a simulation framework that combines global SA methods with a multi-physics BIPV model. The analysis focuses on the performance of naturally ventilated BIPV facade elements (cell temperature and power). Building performance indicators, such as the total heat flux to the building interior and the building wall temperature, are also analysed. Inputs to the SA include convective heat transfer coefficients, cavity airflow rate, and weather conditions. As expected, the SA results were found to be highly dependent on the range selected for the inputs. For a narrow variation in weather conditions, the exterior convective heat transfer coefficient was identified as the input with the strongest influence on the BIPV performance. Results also showed that cavity ventilation becomes more important as the exterior convective heat transfer decreases. These findings indicate the need for accurate models to represent exterior convective heat transfer in BIPV facades and corroborate the importance of cavity ventilation. (C) 2020 Elsevier Ltd. All rights reserved.