Crime is distributed unevenly, a phenomenon commonplace at both areal and individual (personal or household) level. This is of core importance in the allocation of crime reduction resources. However, this importance is not reflected in the choice of measures used to target such resources. This paper distinguishes between alternative crime counts (incidence, prevalence and concentration) and elaborates on their deployment implications, using data from the Sandwell area of the West Midlands. The importance of the concentration measure (victimisations per victim) across units of count is stressed, and it is argued that necessary steps should be taken to incorporate concentration measures routinely into area crime profiles. The three-yearly audit- strategy sequence required under the Crime and Disorder Act 1998 should routinely incorporate such measures.