In this study of optimal organizational performance, we explore how the extent of interactions, both within and among other organizations, affects group performance and stability in a stochastic environment. We have refined a modeling framework (Kauffman and Johnsen's NKC model) so that group size and connections among groups (externalities) can be finely tuned. The search for improved group configurations is modeled as a random walk on a space of possible configurations whereby agents in a group periodically have the opportunity to accept or reject random changes in their characteristics. By controlling which groups have external connections with which other groups, and the magnitude of such connections, we can manipulate the topology of the problem—the web of interactions within and between groups. We present numerical results showing that optimal group size relates to the magnitude of externalities and the accumulated number of random trials. Our main result suggests that for short periods with few trials, large organizations perform best, while for longer time horizons, the advantage accrues to small sized groups with a small number of externalities. However, over these long time horizons, as the extent of external connections increases, modest increases in group size enhances their performance. Under all circumstances, organizations that perform best in the long run fall into a regime of largely stable responses to perturbations, which however, borders on a region of instability.