Deriving aerosol-type-dependent parameterizations for ice nucleation processes remains challenging due to large uncertainties associated with laboratory studies and field measurements. One source of uncertainty is a lack of knowledge about the magnitude of particle-to-particle differences in freezing efficiency associated with particles' chemical composition and surface features. In most experimental setups, it is challenging to disentangle the contributions of inherent experimental uncertainties, variability in aerosol surface area and the particle-to-particle differences in freezing efficiency. Therefore, in this study, we use a Monte Carlo approach to simulate synthetic ice nucleation experiments to better understand the impact of variability in heterogeneous ice nucleation propensities on simulated frozen fractions. We represent this variability by differently-shaped distributions of the heterogeneous ice nucleation rate coefficients J_het. Distributions spanning one order of magnitude result in simulated median frozen fractions that are up to a factor of two higher compared to simulations with narrow Gaussian distributions, in the case of small frozen fractions. For the assumed range of variability in J_het, impacts on cloud variables (e.g., ice water path based on Hawker et al. (2021)) calculated from hypothetical parameterizations based on our simulated frozen fractions seem to be relatively small, with only up to 25% difference between lognormal and constant J_het distributions. One of the major unknowns, however, is the interparticle variability in J_het, which depends on particle composition and surface features. Resolving this uncertainty requires a combination of laboratory studies and field experiments relating physicochemical particle features and ice nucleation propensity in a size-resolved manner.