We study the effect of incorporating heterogeneity into default rules by examining the choice between retirement plans at a firm that transitioned from a defined benefit (DB) to a defined contribution (DC) plan. The default plan for existing employees varied discontinuously depending on their age. Employing regression discontinuity techniques, we find that the default increased the probability of enrollment in the default plan by 60 percentage points. We develop a framework to solve for the optimal default rule analytically and numerically and find that considerable welfare gains are possible if defaults vary by observable characteristics. (Contains 8 tables, 7 figures, and 22 footnotes.)