With increasing price-competition in retail business, marketing becomes the critical success factor. All activities have to be oriented towards customer needs. Within marketing research, the individual act of purchase plays a decisive role and necessitates a holistic analysis of the market basket. Moreover, progress in information technology has enabled us to store huge amounts of data, including a valuable pool of business experiences. It is not possible to make this potential knowledge completely available with conventional statistical approaches. This is where Data Mining provides an ideal approach. It is an analysis process for extracting information from large databases. The aim of the study, submitted as diploma thesis at the ETH Zurich, is to show a possible Data Mining Process for analyzing sales data in retail business. Thereby the analysis is focused on the efficiency of the process model and the identification of professional needs. According to Hippner, the Data Mining Process is applied systematically on the basis of a Market Basket Analysis. The Association Analysis, which detects relevant correlations between different products of an assortment, is the core of the process. Problems of interpreting identified association rules make a transformation into marketing recommendations very difficult. Nevertheless, the Data Mining Process turned out to be very efficient.