PREDICTING CUSTOMER CHURN
project name

PREDICTING CUSTOMER CHURN

description

RESULT – Predictive Model Identifies ‘High Churn-Risk’ With 69% Accuracy Leading to Improved Customer Retention Through Optimal Call Center Spends and Better Policy Management
Predicting Customer Churn Based on
Demographics and Payment History

Insurance
Life Insurance
CONTEXT

Client is a life insurance provider
with over 100,000 active policies.


CHALLENGES
Client was facing very high customer churn leading to loss of long-term business.
Client was unable to anticipate customers that have high churn risk.


Client had low retention rate and high customer dissatisfaction due to misdirected collection efforts that were based on thumb rules – leading to further churn.

TOOLS & METHODS
Quantiphi built a predictive model in R to understand payment behavior, identify high risk cases and assign a persistency score to each policy.

In order to boost the accuracy of the model, Quantiphi integrated data from multiple systems – CRM, Policy Administration, Call Center and Client Demographics.

The various tools and methods used to build the model include Multicollinearity, Adaptive Boosting, Random Forest, Cramer’s V etc.

Quantiphi industrialized the process for one-click deployment of the model every month.




RESULT – Predictive Model identifies ‘high churn-risk’ With 69% accuracy leading to Improved Customer Retention through Optimal Call Center Spends and Better Policy Management

You may also be interested in

contact_small.png