What determines the earnings of a worker relative to his peers in the same occupation? What makes a worker fail in one occupation but succeed in another? More broadly, what are the factors that determine the productivity of a worker occupation match? To help answer questions like these, we propose an empirical measure of multidimensional skill mismatch, which is based on the discrepancy between the portfolio of skills required by an occupation and the portfolio of abilities possessed by a worker for learning those skills. This measure arises naturally in a dynamic model of occupational choice and human capital accumulation with multidimensional skills and Bayesian learning about one’s ability to learn skills. Not only does mismatch depress wage growth in the current occupation, it also leaves a scarring effect—by stunting skill acquisition—that reduces wages in future occupations. Mismatch also predicts different aspects of occupational switching behavior. We construct the empirical analog of our skill mismatch measure from readily available US panel data on individuals and occupations and find empirical support for these implications. The magnitudes of these effects are large: moving from the worst- to the best-matched decile can improve wages by 11% per year for the rest of one’s career.