Predictive Content - Case Study: Customer Churn

For a provider of fuel and toll invoices, biX Consulting created a customer churn analysis in which it was determined which customers would churn at what point in time. With the help of the biX A.I. tools, it was possible to recognise patterns in the historical movement and master data of the customers and to classify them. Different algorithms were applied to different combinations of customer master data. The algorithm with the best test results in the conception phase was then used for classification.
Customer Churn Analysis

Figure 1: Graphical representation of churn prediction using decision tree

The biX solution enables a detailed view into the classification process of each customer. This allows the critical customer master data to be identified and influenced in order to avert customer churn in advance. The focus on customers at risk of churn reduces the effort of churn prevention and prevents the sleeper effect. The algorithms run natively in SAP BW, which is why no additional costs were incurred for licences, hardware or software. As the biX A.I. tools are based directly on SAP technology, existing data structures can be used and integrated into regular operations to ensure that the forecast models are up-to-date. Feel free to contact us.

Figure 2: Detailed view of the classification rules in the decison tree

Customer-churn prevention describes the identification of customers who are willing to cancel in order to prevent cancellation. For this purpose, historical data with terminations are analysed in order to be able to make predictions about the future behaviour of current customers. This allows us to focus on customers at risk and the critical factors.

Contact Person

Oliver Ossenbrink

Management of sales and HR