Using an example from a recent cluster-randomized controlled trial, we discuss the benefits of the methodological approach but also important aspects that need to be considered when implementing and evaluating complex interventions. The intervention case to be used is a skill training program for social workers, the Comprehensive, Methodological and Principle-based Approach. The program focuses on systematizing and improving the follow-up work for longer-term welfare recipients taking part in an activation program. The skill training program for social workers is organized within three essential areas: encountering the user, coordinating services around the user, and administrative work. Within all areas of intervention, social workers learn how to apply various tools to make the follow-up work more systematic and efficient. This is a complex intervention as it contains several components, is implemented at both the professional and organizational level, and because the effect of the intervention depends on interactions between components at different levels. By this case, we exemplify why a cluster-randomized design is applicable for evaluations of complex interventions, however, we also discuss problematic aspects and elements to acknowledge when utilizing cluster-randomized designs.