In this paper authors analyzed credit risks of retail stores by using their payment history. For this reason. first of all the default risk of these firms using logit model was estimated. Then authors applied Multi Level Data Envelopment Analysis (DEA) on risky firms, grouping them according to their risk levels. Data of one of the main wholesaler of Turkey were used for the analysis. The authors' sample was comprised by approximately 6000 retailer costumers and 600.000 transactions. Given model provides a hybrid logistic regression and Multi-DEA toot that can be used by wholesalers to assess the credit risk of their retail customers and to cluster their risky customers by their risk levels. Additionally. the presented results underline the importance of payment history and some non financial factors data for predicting the creditworthiness of a firm.