The main losses in solar cells result from the incomplete utilization of the solar spectrum. Via the addition of an upconverting layer to the rear side of a solar cell, the otherwise-unused sub-bandgap photons can be utilized. In this paper, we demonstrate an efficiency enhancement of a silicon solar cell under real sunlight due to upconversion of sub-bandgap photons. Sunlight was concentrated geometrically with a lens with a factor of up to 50 suns onto upconverter silicon solar cell devices. The upconverter solar cell devices (UCSCDs) were also measured indoors using a solar simulator. To correct for differences in the spectral distribution between real sunlight and the solar simulator a spectral mismatch correction is required and is especially important to properly predict the performance when a non-linear response (e.g. upconversion) is involved. By applying a spectral mismatch correction, good agreement between the solar simulator measurements and the outdoor measurements using real sunlight was achieved. The method was tested on two different upconverter powders, β-NaYF4: 25% Er3+ and Gd2O2S: 10% Er3+, which were both embedded in a polymer. We determined additional photocurrents due to upconversion of 9.4 mA/cm2 with β-NaYF4 and 8.2 mA/cm2 with Gd2O2S under 94-suns concentration. Our results show i) the applicability of measurements using standard solar cell characterization equipment for predicting the performance of non-linear solar devices, and ii) underline the importance of applying proper mismatch corrections for accurate prediction of the performance of such non-linear devices.