A majority of studies around efficiency assessment just estimate it. Indeed, they don’t address the roots of inefficiency of DMUs. These shortcomings could be resulted from silo approaches conducted by researchers. Our research aims address this gap. We propose a two-stage procedure in which first, the efficiency of bancassurance desks are estimated by DEA (Data Envelopment Analysis) and then, roots of inefficiency are identified by a survey designed based on enabler section of European Foundation for Quality Management model (EFQM) and conducted among bancassurance desks employees. This study is applied in terms of purpose, and mathematical as well as statistical in terms of data analysis. The population of this study consisted of all bancassurance desks of BANK MELLAT in 2015. After efficient and inefficient branches of bancassurance have been estimated and ranked, in order to identify factors affecting efficiency, a questionnaire designed according with criteria of EFQM. Reliability of questionnaire was tested by Cronbach's alpha coefficient, and normal distribution of data was tested by using the Kolmogorov-Smirnov test. The effects of variables (factors) on efficiency were evaluated via regression analysis, then obtained answers were tested for research questions. The results of ranking the impact of factors on the efficiency of bancassurance desks showed that employees, partnerships and resources, products and services, strategy, and leadership are the efficient factors, respectively.