Customer churn factors detection in Indonesian postpaid telecommunication services as a supporting medium for preventing waste of IT components

Abstract

Abstract One of the most important tasks in the telecommunication company is to maintain the customer churn rate and customer satisfaction level. These things can be maintained by evaluating the churn customer factors. The results of this evaluation can certainly be used to improve the services of the company, so that customers will be stopped from churn and customer satisfaction level will be increased. This condition indirectly provided positive effect on the environment by preventing e-waste of IT components. In this research study, customer churn factors from customer feedback were evaluated by using text mining, by targeting the postpaid customers in one of telecommunication company in Indonesia. In this research we used 3 classification methodology, Fast Large-Margin, Naïve Bayes, and Support Vector Machine. Fast Large-Margin classification methodology was the most accurate model in this research, was used to classify the customer feedback for five following customer churn factors in telecommunication industry: poor customer service, poor network, high cost of services, billing errors, and promotion seeking. The result shows that promotion seeking was the most factor that caused customer to churn during January and December 2019 period.