Analysis of the effect of environmental conditions on the performance of retailing stores

• Main Author: Vaz, Clara B.
• Publication Date: 2006
• Other Authors: Camanho, Ana
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10198/2815
• topic: Data envelopment Analysis; Retailing; Non-discretionary factors

The assessment of performance in retailing services has gained considerable attention in recent years. The increased competition motivated the organisations to strive for efficiency in order to cut costs and deliver better customer services. This paper develops a method based on Data Envelopment Analysis (DEA) for the efficiency assessment of retailing stores. The method enables the quantification of inefficiencies taking into account the effect of exogenously factors (i.e., non-discretionary inputs and outputs). The method developed starts with the identification of the factors that affect the DMUs’ performance, including those outside the decision makers’ control. The statistical significance of the effect of the exogenous factors on performance is tested using the Kolmogorov-Smirnov test, such that only the relevant factors are considered in the DEA assessment. An enhanced DEA model is run for each store, ensuring that the definition of the efficient frontier is based exclusively on the discretionary variables, and only comparable stores are allowed as peers, i.e., stores whose area of influence is identical or less favourable than the catchment area of the store under assessment. As a result, the shape of the production possibility set is adjusted for each DMU according to the exogenous conditions where it operates. To avoid the loss of discrimination power of the DEA analysis, the choice of the peers authorized in the assessments is done allowing for trade-offs between the nondiscretionary factors (e.g., a store with more population in the surrounding area than the DMU under assessment may be allowed as peer if its level of competition is also higher). The adjustments required to the standard DEA model imply the definition of a mixed integer DEA formulation. To disentangle technical inefficiency from the effect of environmental conditions, the results of the DEA model that takes into account the environmental conditions are compared with the results of a standard DEA model including only the discretionary inputs and outputs. This analysis showed that the average efficiency estimates of the stores analysed increases when the exogenous factors are taken into account, with the efficiency values for some stores increasing up to 15%. Finally, to explore in more detail the impact of each of the non-discretionary factors (population and competition) on the performance of individual stores, a “step by step” approach was used. This approach consists of adding to the DEA model the environmental factors, one at each time (like a stepwise procedure of regression analysis). The results obtained in the successive models are analysed for each DMU to quantify the effect of each exogenous variable on store performance. The applicability of the approach developed in this paper is illustrated in the context of a real-world efficiency assessment of grocery stores. The assessment adopted an output oriented perspective, consistent with the objective of sales maximization. The managerial implications of the results obtained are explored, comparing the insights gained with the DEA analysis with the results of the approaches for managerial planning and control currently used in the organisation.