The photovoltaic performance of a perovskite solar cell is evaluated as a function of its material properties, device geometry, and operating conditions. We conduct a Monte Carlo simulation based on a mechanistic model for meso-structured perovskite solar cell to correlate the device performance with the variability of input modeling parameters. The presented sensitivity analysis is statistically performed in two different scenarios: first by varying the modeling parameters individually, and second by varying all of them simultaneously. The stochastic parameters are ranked and quantified according to their contributions to the variation of the cell performance, thereby providing insights for optimum device performance. Furthermore, a reduced multiple linear regression model is derived for calculating the cell performance without having to solve the full physics-based model. The main finding is that the layer thickness of the hole and electron-transporting layers, and the hole mobility in the hole-transporting layer are the three most critical parameters influencing on the cell performance. When this result is applied to the world record perovskite solar cell of 23.2% efficiency with parameter cross-validation, it is predicted that this efficiency can be further improved by 1.8% to achieve 25%.