churn rate

Detect your clients "moments of life"

Target your offers

Retain your customers

What is the churn rate?

Churn rate helps to calculate the phenomenon of losing customers or subscribers in a year and usually concerns the banking and insurance sectors. This rate is often hard to calculate because many customers do not close their bank account before leaving.

Features show that it is the low quality and the cost of the services offered that will lead people to switch from their initial bank. Moreover, we can see that online banking has been taking steps over traditional banking.

Caisse d’Epargne Normandie banks on Saagie (video)

How can Saagie help you to reduce churn rate?

The data about customers are extracted, collected and gathered by Saagie in a data lake . Every customer’s transaction is analyzed (data mining) in order to predict what kind of situation might drive a client leaving the bank.

Predictive models are created by Data Scientists according to a various data : the number of transactions per month, the amount of these transactions, the means of payment, the number of persons in the family, the age, the SPC and so on.

Time-based schemes are designed to determine the “moments of life” of the customers in order to push the most accurate offers at the right time.

The case of Caisse d’Epargne Normandie

Saagie’s purpose in this project in collaboration with the CEN was to help understand the churn rate among young people aged 15-26 years old.

Several technologies of machine learning were used, including Random Forest a classification model based on descriptive statistics on all operations carried out by each client per month. The goal was to develop several predictive models by adding more data each time. The Python language was used by data scientists for digital analysis and visualization. Each model used is adapted to each problem faced: classification (Random Forest), regression and clustering.

Results came quickly, the first model was built within a day and the analysis of clients’ data permitted to segment them according to their socio-demographic characteristics to adapt the offers and attract clients.

The strategic challenge is to customize customer relationship. Predictive analysis aims to know the customers needs before they know them themselves. Our objective is to offer the right service to the right customer at the right time.

Guillaume Cordelier
Head of Innovation, Caisse d'Epargne Normandie (Bank)