We demonstrate that empirical evidence of employer learning is sensitive to how one defines the career start date and, in turn, measures cumulative work experience. Arcidiacono, Bayer, and Hizmo (2010) find evidence of employer learning for high school graduates but not for college graduates, and conclude that high levels of schooling reveal true productivity. We show that their choice of start date – based on first-observed school exit and often triggered by school vacations – systematically overstates experience and biases learning estimates towards zero for college-educated workers. Using career start dates tied to a more systematic definition of school exit, we find that employer learning is equally evident for high school and college graduates.