Online communities are pervasive in the modern Internet. They are now commonly used in business (for customer support and harnessing customer innovation), education (to foster peer-to-peer learning) and government (to induce civic participation). Many online communities employ professionals called online community managers to set the community’s tone, mediate conflict and foster growth. They operate by following protocols («policies») mandated by their employers. We present and discuss an agent-based network model of a growing online community with and without online community management policies. We propose four dimensions against which their work can be evaluated, and compare the effectiveness and efficiency of two fairly common policies against a baseline of no online community management at all.