Predicting Churn Before Customers Leave
Vodafone Ziggo knew bad connections cost them customers, but not which customers were about to leave. We built models that flag unhappy households and the lines due for maintenance, so support reaches them first.
Vodafone Ziggo
- 4Mhomes
- Vodafone Ziggo serves with television, internet and telephone
- 2model families
- one predicting unhappy customers, one predicting lines due for maintenance
About the client
Vodafone Ziggo is the largest cable operator in the Netherlands, providing television, internet, and telephone services to approximately 4 million homes.
Why were we there?
DataChef helped Vodafone Ziggo by providing an ML solution to the "unhappy customers" problem. They knew that a bad internet connection leads to unhappy customers, which increases the churn rates, but it was unclear exactly what caused the bad connection. On the other hand, long waiting times for mechanics added to the customers' unhappiness.
How we helped
DataChef provided Machine Learning models to predict the unhappiness of the customers. We also implemented predictive models to foresee which lines needed maintenance.
This allowed Ziggo to proactively tend to customers who were unsatisfied with their Wi-Fi and keep them from turning to other internet providers.
The impact we made
Perform preventive maintenance
Reach out proactively to offer support in probable unhappy cases
Increase customer happiness and therefore reduce churn rate
Technologies we used
This allowed Ziggo to proactively tend to customers who were unsatisfied with their Wi-Fi and keep them from turning to other internet providers.
More case studies
