Evaluating performance of white goods services using DEA, beta regression, and cluster analysis

This study adopts an integrated quantitative framework that combines Data Envelopment Analysis (DEA), bootstrap methods, beta regression, and cluster analysis to assess the performance of authorized service centers in the white goods industry. Initially, traditional DEA was applied to estimate relative efficiency levels, after which a bootstrap-enhanced DEA was utilized to improve the robustness of the efficiency measures by mitigating sample sensitivity. Given that the efficiency outcomes followed a continuous distribution between 0 and 1, beta regression was employed as a suitable modeling technique to identify and evaluate the determinants of service performance, yielding statistically significant insights while effectively addressing issues such as heteroscedasticity and skewness that often challenge linear models. Building on these results, cluster analysis was conducted to classify service centers into groups with similar performance profiles, highlighting meaningful distinctions among clusters in areas such as strategic planning, resource utilization, and service quality. The findings from the clustering and determinant analysis offer actionable managerial recommendations, suggesting that high-performing clusters should be benchmarked for resource optimization and targeted quality improvement initiatives.