We argue that to support agriculture–environmental policy decision making, stakeholders need ‘quantitative back-of-the-envelope’ analysis that is timely and sufficiently accurate to make informed decisions. We apply this concept to the analysis of the supply of ecosystem services from agriculture. We present a spatially explicit production model and show how it can be used to derive the supply of ecosystem services in a region. This model shows that the supply of ecosystem services can be derived from the spatial distribution of opportunity cost of providing those services. We then show how this conceptual model can be used to develop a minimum-data (MD) approach to the analysis of the supply of ecosystem services from agriculture that can be implemented with the kinds of secondary data that are available in most parts of the world. We apply the MD approach to simulate the supply of carbon that could be sequestered in agricultural soils in the dryland grain-producing region of Montana. We find that the supply curve derived from the MD approach can approximate the supply curve obtained from a more elaborate model based on site-specific data, and can do so with sufficient accuracy for policy analysis.