Despite the centrality of harm to crime and criminalization and increasing interest in harm as a basis for crime-control policy, the criminological community has yet done little to systematically reflect on criminal harms or their identification, evaluation and comparison. This paper presents a newly developed framework with which to systematize the empirical assessment of such harms and address at least some of the attendant conceptual and technical challenges. It also suggests several roles for the framework in policy making. Our conclusions are twofold: it is possible to reliably evaluate the harms of criminal activities, as our examples suggest, but it is not possible—for both conceptual and technical reasons—to develop an encompassing estimate of the total harms of these activities.