Hydro-Québec has been using CFD to analyze the performance of its existing turbines for many years. Most of those analyses are based on the measurement of a single runner blade. However, due to manufacturing techniques, in-situ modifications or repairs, there are often small differences between individual blades of the same runner. The impact of this non uniformity was not known thus far and was often assumed to be negligible given the size of the runner. This paper highlights the impact of such differences by presenting the CFD analysis of various blades measured on the same runner. Two different geometries are used for demonstration: the AxialT model propeller and a 50-MW Francis turbine. In both cases, about 50% of the blades could not be considered as representative of the whole turbine and using them could lead to wrong conclusions regarding the turbine performance.