The lifetime of nitrous oxide, the third-most-important human-emitted greenhouse gas, is based to date primarily on model studies or scaling to other gases. This work calculates a semiempirical lifetime based on Microwave Limb Sounder satellite measurements of stratospheric profiles of nitrous oxide, ozone, and temperature; laboratory cross-section data for ozone and molecular oxygen plus kinetics for O(1D); the observed solar spectrum; and a simple radiative transfer model. The result is 116 ± 9 years. The observed monthly-to-biennial variations in lifetime and tropical abundance are well matched by four independent chemistry-transport models driven by reanalysis meteorological fields for the period of observation (2005–2010), but all these models overestimate the lifetime due to lower abundances in the critical loss region near 32 km in the tropics. These models plus a chemistry-climate model agree on the nitrous oxide feedback factor on its own lifetime of 0.94 ± 0.01, giving N2O perturbations an effective residence time of 109 years. Combining this new empirical lifetime with model estimates of residence time and preindustrial lifetime (123 years) adjusts our best estimates of the human-natural balance of emissions today and improves the accuracy of projected nitrous oxide increases over this century.