The goal of this paper is to analyze the innovation space where firms and free collaborators meet. We considered it from the point of view of the intermediation function theorized in the open innovation paradigm and focused on the decisions firms can take in order to shape the in- and out-knowledge-flow for their own goals. Building on the case study of a player in the automotive industry, we examined how the structure and governance of interactions as well as the activity of community managers can influence the final output of collaborative innovation practices. Using network-, statistical and content analysis we argued that different strategies address different motivations and thus generate specific forms of engagement which could be relevant in specific phases of the innovation process.