The paper provides a theoretical framework for categorizing organized crime groups based on what they do – whether they produce, trade or govern – as well as their aims. This paper then tests whether the internal structure of a heroin distribution organization in New York City, a Sicilian mafia group and the Provisional Irish Republican Army differ. Applying Exponential Random Graph Models (ERGMs) methods to network data, we find the organizational structure of trade-type organized crime differs markedly from governance-type, as well as between financially-motivated and politically-motivated groups. Trade-type organized crime and financially-motivated groups display a high level of centralization, an even distribution of clustering values, short paths and low homophily. Governance-type organized crime and politically-motivated groups display the opposite features. We conclude that the core activity and aim of the group are crucial in understanding the organizational structure.