A saturated riparian buffer (SRB) is an edge‐of‐field conservation practice that reduces nitrate export from agricultural lands by redistributing tile drainage as shallow groundwater and allowing for denitrification and plant uptake. We propose an approach to improve the design of SRBs by analyzing a tradeoff in choosing the SRB width, and we apply the approach to six sites with SRBs in central Iowa. A larger width allows for more residence time, which increases the opportunity for removing nitrate that enters the buffer. However, because the SRBs considered here treat only a portion of the tile flow when it is large, for the same difference in hydraulic head, a smaller width allows more of the total tile flow to enter the buffer and therefore treats more of the drainage. By maximizing the effectiveness of nitrate removal—defined as the ratio of total nitrate removed by the SRB to total nitrate leaving the field in tile drainage, an equation for the optimal width was derived in terms of soil properties, denitrification rates, and head difference. All six sites with existing SRBs considered here have optimal widths smaller than the current width, and two are below the minimum width listed in current design standards. In terms of uncertainty, the main challenges in computing the optimal width for a site are estimating the removal coefficient for nitrate and determining the saturated hydraulic conductivity. Nevertheless, including a width that accounts for site conditions in the design standards would improve water quality, locally and regionally.