Cloud properties and their evolution influence Earth's radiative balance. The cloud microphysical (CMP) processes that shape these properties are therefore important to represent in global climate models. Historically, parameterizations in these models have grown more detailed and complex. However, a simpler formulation of CMP processes may leave the model results mostly unchanged while enabling an easier interpretation of model results and helping to increase process understanding. This study employs sensitivity analysis of an emulated perturbed parameter ensemble of the global aerosol–climate model ECHAM-HAM to illuminate the impact of selected CMP cloud ice processes on model output. The response to the perturbation of a process serves as a proxy for the effect of a simplification. Autoconversion of ice crystals is found to be the dominant CMP process in influencing key variables such as the ice water path and cloud radiative effects, while riming of cloud droplets on snow has the most influence on the liquid phase. Accretion of ice and snow and self-collection of ice crystals have a negligible influence on model output and are therefore identified as suitable candidates for future simplifications. In turn, the dominating role of autoconversion suggests that this process has the greatest need to be represented correctly. A seasonal and spatially resolved analysis employing a spherical harmonics expansion of the data corroborates the results. This study introduces a new application for the combination of statistical emulation and sensitivity analysis to evaluate the sensitivity of a complex numerical model to a specific parameterized process. It paves the way for simplifications of CMP processes leading to more interpretable climate model results.