Rangeland ecological research is generally driven by interest in a particular pattern, process or organism, by a desire to maximize a particular ecosystem service (such as livestock production) at a specific location, or by an interest in the effects of a particular management system on soils, vegetation or wildlife. Government managers and policymakers responsible for making the most effective use of funding to improve rangeland management are often required to base their decisions on science, but existing science is often inaccessible or irrelevant. It is also often presented in a form that makes it difficult to determine its relevance to management of a particular location. We propose a framework for increasing the relevance of science to policy and management, and the role of scientists in providing feedback to policymakers on how to increase the effectiveness of government programs that are designed to improve rangeland management. The framework consists of two components. The first is an ecological site-based system for prioritizing rangeland research, and for increasing the ability of managers to access and interpret relevant research results. This component includes a proposal for the adoption of a minimum set of required metadata for journal publications. The second component is designed to increase the quality and quantity of monitoring data collected in association with the implementation of conservation programs, and to increase the involvement of scientists in the interpretation of these data. We will conclude with a discussion of the benefits, costs and potential risks associated with adopting this framework.