The choices that smallholder farmers are able to make are strongly conditioned by the geographic conditions in which they live. The importance of this fact for rural development strategy is not lost on policy makers. For example, the government of Ethiopia frequently frames policy discussions by broadly different geographical conditions of moisture availability, recognizing moisture reliable, drought prone and pastoralist areas. These conditions are seen as important criteria for determining the nature, extent and priority of development interventions for different parts of the country. There is considerable evidence, however, that other geographical factors also have important implications for rural development options. This paper uses agroecology, access to markets, and population density to define development domains: geographical locations sharing broadly similar rural development constraints and opportunities. Unlike similar efforts conducted elsewhere, this work is unique in that it seeks to move away from a subjective mapping of factors of theorized importance to a more rigorous definition of development domains on the basis of quantitative data on smallholder livelihood strategies. After selecting variables for mapping, we calibrate our definition for domains in such a way that their explanatory power is maximized across a range of livelihood strategies that figure in the current Ethiopian rural development discourse (market engagement, dependence upon agriculture, etc.).