Global Financial Institution
Client churn probability calculation
Client retention dashboard
Our client retention dashboard predicts the probability of a client churn – also the probability, that a client cancels a subscription-based service. Our model is applicable in any subscription-based context, e.g. banking, insurance, telecommunication, newspaper subscriptions. Our references are in the banking industry.
The self-learning artificial intelligence based model uses ca. 200 attributes to predict the probability of the client attrition. A unique feature is the ability to automatically verbalise the detected churn reasons.
How does it work?
The goal is to drive the fluctuation of the customers in a way that maximizes profits:
Keep profitable customers
Replace loss-generating customers with profitable ones or transform them into profitable ones
Client Retention Dashboard – main features
Minimum 150-200 data points per client per month describing client behavior
A minimum of 12 months of historical data
|Expected churn probability for the coming 6 months per client|
|Displayed in an easy to interpret & actionable report|
How good is the prediction?
Prediction accuracy is above 90% (based on project experience)
Model Retraining Frequency
We recommend a retraining every 3-6 months