This paper describes an approach to mine for competencies of users´ in a social software system. The system supports knowl¬edge sharing in a structured way supported by social tagging and an ontology. This meta-data is used to find relevant knowl¬edge sources, cluster these resources and to recommend certain resources to users. Our research assumption is, thatif a user enters knowledge about a certain domain, s/he is in a certain degree competent in this area. In our experiment we generate competence profiles of users and compare it with a competence profile obtained by a self-assessment of users. In this paper we describe the performed experiments at the university where students and university staff use the system to share knowledge.