Detecting factors affecting customers’ contract termination in the electricity distribution network

A data set obtained from distribution system operators (DSOs) consisting of 41,574 individual customer contracts with termination information was combined with socio-economic and geographical data. A geospatial analysis was conducted and subsequently, multivariate statistical models were used to study the factors associated with contract termination. Independent validation data were used to replicate the results. Distance to a lake and lack of jobs in the area were established as risk factors for contract termination. The constructed model enables calculation of risk scores for individual customers. Risk scores ranged from 0.0% to 7.0%, whereas the baseline risk was 1.0% over a study period of 3.9 years. Thus, our model has discriminatory power.