Market segmentation of an online auction site (www.ebay.de) is studied using a novel clustering technique via community detection. A network of bidders connected by common interest in individual articles is constructed, whose community structure corresponds to the main user groups according to common interest. A key feature of the analysis is its independence of any kind of measure of similarity between the bidders or articles offered, or any kind of dimensionality reduction possibly biasing the analysis. Further, the method works on the sparse raw data directly, is scalable to large systems and can be used to discover both hierarchical and overlapping cluster structures. Results are compared to null models based on random networks and clusters are validated and interpreted using the taxonomic classifications of eBay categories. We find clear-cut and coherent interest profiles for the bidders in each cluster. The interest profiles of bidder groups are compared to the classification of articles actually bought by these users during the time span 6–9 months after the initial grouping. Their temporal stability indicates typical interest profiles in society. Our results shed some light on typical characteristics of online markets and their segmentation. They show how network theory can be applied successfully to problems of cluster analysis in economic and sociological milieu studies with large, sparse and high dimensional data.