Conservation buffers, such as filter strips and riparian forest buffers, are widely prescribed to improve and protect water quality in agricultural landscapes. These buffers intercept field runoff and retain some of its pollutant load before it reaches a waterway. A buffer typically is designed to have uniform width along a field margin and to intercept runoff that flows uniformly to it. However, spatial analysis of field conditions and runoff patterns indicate that more runoff is likely to flow to some locations along a field margin than to others which can substantially limit a buffer's effectiveness. We propose that precision conservation, the use of precision spatial information, technologies, and procedures to implement conservation practices, can be used to improve the design of buffers and ensure their effectiveness. Precision conservation can integrate detailed landscape data with mathematical models in a geographic information system. We can then analyze spatial patterns of runoff and design variable-width buffers that precisely match the needs of every location along a waterway. Greater cost of precision conservation is offset partly by greater water-quality benefit from each acre of buffer. Many of the required data sources and modeling components already exist, substantial improvements are possible that can produce even greater conservation efficiency.