The idea of institutional diagnostics has sparked considerable interest over the last twenty years, especially among those seeking to assess how architectural choices can affect the performance of institutions established to address environmental issues. But so far, this appealing idea has not developed into a practice that is useful in making choices regarding the architecture of regimes addressing needs for governance arising at various levels of social organization. I argue that this is attributable to a dramatic expansion in the range of concerns we now include in thinking about needs for governance and how to respond to them. Diagnosis cannot deal with the full range of these concerns in one step. To make diagnostic procedures useful, we must proceed in stages, identifying the basic nature of a need for governance and then proceeding in a stepwise fashion to identify and evaluate institutional options. I introduce the idea of a diagnostic tree as a tool to guide this process and explore its usefulness in a generalized account of key issues arising in the realm of fisheries management. I conclude with a discussion of points to bear in mind in any effort to use this tool to sort out the complexities of institutional design in a variety of real-world settings.