A DEA-Based Approach to Evaluate the Efficiency of Non-Homogeneous Service Locations

This study aims at evaluating the performance of a company, 'XYZ Company', that has 115 service locations. Because of its ability of handling large numbers of inputs and outputs, and removing the need of predefining the factors' weights, Data Envelopment Analysis (DEA) is used. DEA is benchmark tool that measures the efficiency of entities with respect to each other by assessing their performance of utilizing inputs to produce outputs. Researchers have developed several DEA models, all of which have different characteristics.A main assumption of DEA is that the entities are homogeneous – i.e. operating under similar conditions, which is not applicable sometimes. Thus, various approaches have been introduced to relax the homogeneity assumption. In this study, we propose an approach that estimates the efficiency over some stages, obtains efficiency scores from each stage, and then calculates the final weighted score by assigning a higher weight to the stage that represents the actual conditions of the entity more clearly.We apply three DEA models, utilizing the proposed approach to overcome the entities' heterogeneity, to the data set of XYZ Company. Then, we compare the results of the three models, analyze the efficiency scores of the 115 service locations, and provide some major findings.