A network approach helps us better specify and model collaboration among people involved in organized crime. The focus on collaboration raises the boundary specification problem: Where do criminal organizations start, where do they end, and who is involved? Traditional approaches sometimes assume the existence of simple, rigid structures when complexity and fluidity are the norms. A network approach embraces this complexity conceptually and provides methodological guidelines for clarifying boundaries. Boundary specification in organized crime helps solve four puzzles. First, social boundaries: a network approach reduces confusion about social boundaries as criminal entrepreneurs interact with criminals and noncriminals in diverse contexts, only some of them illicit. Second, boundaries of group membership: network data and methods obviate the need for formal membership attributions. Third, ethnic boundaries network analyses reveal that the effective boundaries of criminal organizations are based on social relations, not attributes such as ethnicity. Fourth, recruitment: attending to the larger social environments in which organizations are embedded provides a clearer view of how mechanisms of recruitment cross seemingly rigid boundaries between members and prospective members.