In the Netherlands ‘Safety Houses’ have been established, in which partner organizations in the field of criminal justice, crime prevention, law enforcement, public administration and social services collaborate in order to reduce crime and recidivism, and to increase public safety. This article examines why some Safety Houses are better in achieving these goals than others. The effectiveness of 39 Safety Houses is analyzed by means of QCA (qualitative comparative analysis). Results show two different paths leading to effective outcomes. Effective Safety Houses have been in existence for at least three years, show a high degree of stability and a centrally integrated collaboration structure. In addition, they either have considerable resources at their disposal or have been set up with a network administrative organization, where a neutral coordinator governs the network.