Churn rate


Churn prediction is one of the Artificial Intelligence applications with higher business impact. It looks for detecting customers who are likely to cancel a subscription to a service. Though originally used within the telecommunications industry, it has become common practice across banks, ISPs, insurance firms and other verticals.

We have developed a Terminus7 machine learning module focus on predicting churn rate  based on client behaviour patterns.


To predict custumers likely to cancel a service


We have collaborated with several of our customers of Retail and Food&Beverage sectors to test our Terminus7 Churn Rate prediction module.

Our solution is based on the hypothesis that similar user behaviours implies similar probability to leave the service. The behaviour is based on both demographic characteristics and customer purchase/usage patterns. Therefore, the solution link customers based on the past behaviours. Timeline should be defined in both sides: how far to go in the past data to predict future customer behaviour; how far to go in the future as a prediction borderline with a real business value.

This is a similar approach to Collaborative Filtering techniques used in Machine Learning personalization for e-commerce recommendation systems.  

We trained T7 Machine with past customers data and churn rates in order to correlate both information sources with Machine Learning techniques.

After Terminus7 training, Terminus7 demonstrates to be able to predict churn rate with high accuracy and low false positives. Our experience demonstrates that the final Machine Learning prediction quality strongly depends on data quality and volume.


Prediction based on behaviour patterns

Do you want integrate AI to your business?