Analýza a modelování platebních preferencí zákazníků společnosti s cílem optimalizace nákladů

This bachelor’s thesis focuses on the analysis and modeling of customer payment behavior at a major supplier operating in the Czech energy market. The main objective of the research was to identify customer segments that, from the company’s perspective, use costly payment methods. SIPO payments were identified as a disadvantageous method due to high transaction costs, which are more than double those of other offered methods. To effectively cluster customers, cluster analysis using the non-hierarchical k-prototypes method was selected. This method was applied to a dataset containing 102,564 observations and 28 variables. Due to the specific nature of the data, the dataset was divided into four subsets by account type and commodity prior to clustering. A total of 14 distinct customer segments were identified. Cost-intensive segments were found only in the subsets for the household account type, for both commodities. In the household–electricity sub-set, the identified segment had a 55.43% share of SIPO payments, and in the household–gas sub-set, one segment had a 65.80% share of customers using SIPO. Both segments were characterized by low technology adoption rates and a high median age (70 and 73 years), which significantly exceeds the median of the entire dataset (53 years). Another common feature of these segments is the low median advance payment amount (CZK 940 and CZK 350) compared to the overall sample median of CZK 1,500. The SIPO method was also present in the remaining segments, particularly among customers in older age groups. However, compared to the segments mentioned above, the percentage of customers using this payment method is lower in these groups. Based on these findings, the conclusion of the study formulated specific recommendations for the company targeting cost-inefficient customer segments.