Enthusiasm for adaptive management has outrun evaluation of its usefulness as a natural resource management tool. Policymakers routinely endorse, and frequently require, it. Managers and academic observers alike have tended to assume that adaptive management is uniformly the best strategy. Little has been said, particularly in the policy literature, about how to decide whether an adaptive management approach makes sense. Looking at adaptive management as an information problem, this Article argues that adaptive management should be used only when it promises to improve management outcomes sufficiently to justify the additional costs it imposes. An explicit formal analysis of the prospects for learning and the value of learning for management should precede any decision to engage in adaptive management. For large-scale, long-term, or high-profile adaptive management programs, that analysis should be reviewed by outside experts and periodically reexamined. The type of analysis recommended here would help limit the use of adaptive management to appropriate circumstances, improve implementation when adaptive management is adopted, and enhance accountability. It would also highlight situations in which learning would be valuable for managers but appears too costly or difficult. The analysis should highlight barriers to learning. Many will be context specific, but others are systematic. This Article offers suggestions for addressing some of the most common systematic impediments to learning.