Attraction and commercial success of web sites depend heavily on the additional values visitors may find. Here, in- dividual, automatically obtained and maintained user pro- files are the key for user satisfaction. This contribution shows for the example of a cooking information site how user profiles might be obtained using category information provided by cooking recipes. It is shown that metrical dis- tance functions and standard clustering procedures lead to erroneous results. Instead, we propose a new mutual infor- mation based clustering approach and outline its implica- tions for the example of user profiling.