Customer churn analysis for mobile network operators in Zimbabwe. Case study (Network Company in Zimbabwe)

Abstract

Customer churn analysis helps to understand the behaviour and trend of those who churn so that network operator always ahead of them. The research is hereby analysing the effect of age group on churning intention and an algorithm for prediction of churners to ensure steering of retention strategies to the pertinent group at churning risk. The researcher distributed hundred questionnaires to mobile phone subscribers using systematic random sampling technique. SPSS was used to generate inferential output with a statistical test called One-way ANOVA. The results revealed that not all age groups have significant customers who agitate network provider because of poor network quality, poor network reliability and poor customer handling with the exception of poor network coverage. However, poor network coverage significantly agitates clientele base despite of their age group. SPSS Statistics 20 employed to fit a binary logistic regression model in which seven independent variables were attested with 90% accuracy of predicting potential churners. Sample of 400 mobile phone subscribers randomly selected from the customer base of network company in Zimbabwe. The researcher also used Eviews to perform unit root test, which was part of model building. Explanatory variables in the logit model found to have a significant association with customer churn probability. Recommendations mainly grounded on the effect of each logit coefficient. The researcher, finally suggested telecoms in Zimbabwe to enhance customer loyalty through provision of quality services, lower tariffs, elongate validation period for its services to assure that subscribers are able to meet their communication needs anytime, anywhere and to anyone. The research also recommended provision of lower tariffs to upsurge recharges. Keywords: Customer churn; One-way ANOVA; Binary logistic regression.