Practical skills are widely believed to be a key determinant of labor market success, yet credible empirical evidence remains scarce, because these skills are occupation-specific, acquired primarily through workplace experience, and notoriously difficult to measure. We address these challenges using an exceptional dataset covering over 170 occupations. Our measure of practical skills is based on high-stakes expert evaluations of apprentices’ performance in occupation-specific standardized practical examinations, conducted under authentic workplace conditions and lasting from several hours to several weeks. Combined with rich administrative data, this allows us to separate practical skills from general and vocational knowledge. We show that practical skills form a distinct dimension of human capital, only weakly correlated with traditional achievement measures. Practical skills are the strongest predictor of early labor market success, consistently associated with higher first-year earnings, lower NEET risk, greater retention by the training firm, and a higher likelihood of entering tertiary education. These relationships are robust across occupational task groups and by gender. Out-of-sample analyses show that practical skills substantially improve predictions beyond conventional educational achievement and background characteristics.