The distributions of microorganisms in foods impact the likelihood that a foodstuff will cause illness and therefore also impact the consequential public health burden. As part of food safety management systems, food is sampled and microbiologically tested. The effectiveness of the sampling programme is related to the spatial distribution of the microorganisms that are being sampled for. However, detailed information about the spatial distributions of the microorganisms in food is scarce. The impact of microbial clustering and different types of statistical distributions on public health, performance of sampling and performance objectives are discussed. Examples with moderate levels of Listeria monocytogenes and low levels of Salmonella spp. both distributed with various degree of clustering, show the impact of microbial clustering and the impact of different distributions of exposure on the probability of illness. It can be concluded that the risk to get ill can be heavily influenced by variability in doses, as caused by clustering, as well as average dose. This risk is often largely determined by infrequent high doses represented by the right hand tail of the frequency distribution. These infrequent high values are the most important contributors to the arithmetic mean in a batch and thus it is the arithmetic mean (mean of counts), which is more relevant to the assessment of risk than the geometric mean (mean of logs), which has been the most commonly used parameter to represent average microbial counts. Furthermore, a more sophisticated definition of performance objectives that includes consideration of clustering might be needed. Both clustering and the choice of statistical distributions have a substantial effect on the acceptance probability of microbiological criteria.