Churn Forecasting Model for South African Pre-Paid Service Providers
Churn Forecasting Model for South African Pre-Paid Service Providers

Author(s): Olusola Gbenga Olufemi, Trudie Strydom
Subject(s): Business Economy / Management
Published by: Scientia Moralitas Research Institute
Keywords: Churn; Data Mining; Decision Support System; Customer Relationship Management; Bayesian Networks
Summary/Abstract: Telecommunication companies globally confront with rising problems of customer agitations. Inadequacy of telecoms’ rendered services, delivered products and many other causes, result in the difficult moments telecoms face. These problems have further degenerated to customers leaving from one network provider to the other, in quest for improved satisfaction. Churn is the term used to describe this customers’ resultant movement, due to agitation caused by inadequate operations. The Republic of South Africa (RSA) telecoms presently face this social problem called churn. To understand what causes customer churn, deep studies on varied literature on customer churn revealed the reasons behind this movement, i.e. churn factors. A developed customer experience questionnaire from these studied factors identified the main churn causing factors in RSA telecoms. This questionnaire eased the obtaining of data-records from respondents in South Africa, used in creating varied datasets. Using the varied datasets, a Bayesian networks’ model developed detected and evaluated churn likelihood in these different telecoms. This model proved to have more predicting potentials and relevance in our present days. Three factors revealed to impact more on customer churn in South Africa, by way of the predictions carried out by the derived model. These factors are: Friends & Family Deals on Networks (FFD), Customer Care Service (CCS), and Offers & Promotions (OP).

• Details
• Contents