Recommender systems have a long tradition of reducing users' search costs by proposing items on the basis of users' preferences and aggregated information about other users. In e-commerce scenarios, different types of user preferences—implicitly collected ratings as well as explicitly formulated requirements—are available. The authors perform a comparative evaluation across different recommendation techniques, such as knowledge-based sales advisory and collaborative filtering, on a commercial data set. By making this data set publicly available, the authors hope to foster research efforts on the specific requirements of commercial shopping platforms.