Due to an increasing diffusion of information technologies (IT), the labor market requires more and more so-called “IT skills”. Recent studies confirm that IT skills are relevant for individual labor market outcomes. However, so far researchers do not use a consistent definition and measure of IT skills. In our paper, we distinguish different types of IT skills that may lead to structurally different labor market outcomes like wages. To measure these different types of IT skills, we propose an innovative way of measuring skills based on training curricula of apprenticeship occupations. We use modern computational linguistics methods, i.e. a topic modelling algorithm called Non-Negative Matrix Factorization. By doing so, we identify different types of IT skills like e.g. implementing ICT (Information and Communications Technologies), developing applications, designing webpages or installing software, handling system technology, CNC (Computerized Numerical Control), CAD (Computer-Aided Design), and handling control technology. Our results show that although IT skills in general have a positive effect on wages, different types of IT skills are associated with differing labor market returns, e.g. general digital skills like ICT and developing applications relate to higher wages than technology-specific IT skills like handling system technology.