Non-profit institutions must make many kinds of decisions about allocations of resources. The theory is clear: allocate so that the marginal utility is equal in all uses. But this is more easily said than done. This paper suggests a series of heuristic steps by which real-life non-profit allocation situations can be coordinated to the marginal analysis. It discusses the assumptions that must and should be made and sets forth a conceptual framework for the analytic decisions. The process is illustrated with an example from library research--the decision about how many books should be placed into storage libraries to maximize the welfare of a university community.