The present study aims to analyze the competitive strategies adopted by Brazilian retail companies and the impact on their results, generating a taxonomy that allows a better understanding of this sector. The selected companies are the largest in the segment in Brazil, representing more than 60 percent of revenue, from 2017 to 2021. Secondary data from Economática and from the investor information of each company were used to select strategic and performance variables. Such data were analyzed using SPSS (Statistical Package for Social Science) algorithms. After preparing the database, consisting of 37 companies, with treatment of outliers and missing values, the relevant factors for defining the clustering variables were extracted from the strategic space. The application, in sequence, of hierarchical cluster analysis and k-means made it possible to determine the optimal number of clusters and generate groups of companies with defined strategic profiles, so that those in the same group adopt similar strategies, although different from those employed by companies in other groups. MANOVA tests were performed to validate this conclusion. Finally, the performance vectors of each cluster were calculated and the appropriate statistical tests, MANOVA and ANOVA, including post hoc analysis, were performed to create a hierarchy that would lend practical meaning to the taxonomy. The conclusion is that the taxonomy with three clusters has better adherence to the reality of the sector and that companies that practice broad differentiation strategies perform better than those that practice cost or any type of approach.