The Profan project focused on using process data to predict successful work on tasks in the domains of complex problem solving and reading of digital texts. To this end, data from a computer-based assessment for PISA 2012 was examined. Suitable indicators were built to demonstrate which particular processes of the two competence domains are not sufficiently mastered by students who failed or only partially succeeded in solving the tasks. The indicators were subsequently checked for their appropriateness regarding the explanation of differences in the achievement of students from distinct sociodemographic groups. Based on the information thus gained, a prototypical formative assessment item was finally developed for each of the two domains.