To support researchers in questions on Research Data Management (RDM), a lot of universities have established centers that offer RDM counseling services. For aligning these services with the RDM requirements of their target groups, many of these centers have conducted quantitative surveys and qualitative interviews. Unfortunately, these methods are not entirely appropriate to analyze RDM requirements. This blind spot in RDM research was addressed in my PhD dissertation, which resulted in a formal description model for RDM requirements in the Humanities, their conditions, workflows and solutions as well as in an analyis of such model on a content level. The model consists of a hierarchical system of categories developed on the basis of a corpus of anonymized RDM counseling protocols. In my talk, I will present the formal system of categories, which have come to constitute a conceptual language for RDM emerging from actual praxis in the field, and the profiles of RDM requirements which resulted from the content analysis.