The article introduces Data Envelopment Analysis (DEA) and its applicability as mathematical programming technique. It evaluates the performance of homogeneous operating decision-making units. DEA has been proven as valuable performance evaluation method in situations when decision-making units under consideration have multiple inputs and outputs and operate under similar conditions. For dealing with situations when units operate under different conditions, we have proposed categorical approach and analysed the influence of unit’s environment on relative efficiency results by applying categorical model on real data of 57 shops within one retailing organization. DEA identified good operating practices as members of efficient frontier (benchmark members) and those under efficient frontier that should be analysed as candidates for reorganization or even closure. Relative efficiency results obtained by non-controllable BCC model and categorical BCC model were significantly different so our conclusion is that business environment greatly influences the performance evaluation for several units and should be additionally investigated.