Aim Because the field of information systems (IS) research is vast and diverse, structuring it is a necessary precondition for any further analysis of artefacts. To structure research fields, taxonomies are a useful tool. Approaches aiming to develop sound taxonomies exist, but they do not focus on empirical development. We aimed to close this gap by providing the CAFE methodology, which is based on quantitative content analysis. Subject and methods Existing taxonomies are used to build a coding scheme, which is then validated on an IS project database. After describing the methodology, it is applied to develop a telemedicine taxonomy. Results The CAFE methodology consists of four steps, including applicable methods. It helps in producing quantitative data for statistical analysis to empirically ground any newly developed taxonomy. By applying the methodology, a taxonomy for telemedicine is presented, including, e.g. application types, settings or the technology involved in telemedicine initiatives. Conclusion Taxonomies can serve in identifying both components and outcomes to analyse. As such, our empirically sound methodology for deriving those is a contribution not only to evaluation research but also to the development of future successful telemedicine or other digital applications.