Individual contributions to public good investments are subject to the problem of free riding. We investigate the possibility of overcoming the free-riding problem through creating social capital via communication. Using data from a public goods experiment, we empirically test the effectiveness of two commonly used types of communication interventions in various organizations—structured, goal-oriented communication and unstructured, free-form communication—in creating social capital. Although both types of communication are found to reduce free-riding, when players stay in the same groups before and after communication, unstructured communication brings voluntary contributions closer to the efficient level persistently; structured communication is less successful. In contrast, structured communication is more successful when players are allocated to different groups after communication; unstructured communication has a smaller impact on voluntary contributions in this case.