“Collaboration” is generally portrayed as being beneficial to authorities, even if previous collaborative research shows that conflicts are common between authorities who are supposed to cooperate. What takes place when different actors in the collaboration meet in practice? And what is the best way to analyse this? In qualitative studies, it is often problematic to go from an exhaustive analysis of individual empirical instances to an overall picture of the context or phenomenon in which all instances taken together can be viewed as a case. Years of close engagement with the data may interfere with the analyst’s capacities and opportunities to contextualize a study more broadly and theoretically, and detailed knowledge about a range of situations in the field may make novel contextualizations difficult. This article discusses how to overcome such obstacles, using examples from a study about a “collaboration” project in Swedish youth care.