Anticipate market trends and make data-driven decisions with our powerful predictive analytics solutions
Insurance | Customer Retention | Predictive Analytics
Case Study Details
In an era where data-driven decision-making is critical, a leading local insurer has taken a significant step forward in utilising predictive analytics to improve customer retention for its life and disability insurance products. We are proud to share the success of our recent collaboration with a leading financial services company, which aimed to predict customer propensity to lapse, cancel, or not take up life and disability insurance policies.
Challenge
A leading local insurer faced a critical challenge with a substantial portion of its revenue being lost due to poor customer retention in the form of not taken ups, lapses, and cancellations. This posed a significant impact on their profitability and customer satisfaction.
Solution
To address this, we embarked on a project to determine the feasibility of implementing a predictive analytics model leveraging machine learning. The goal was to accurately predict which policies were at risk, allowing the local insurer to take proactive measures.
Approach
- Data Collection and Exploration: We gathered data from various sources including leads data from vehicle finance groups, the insurers own internal sales and collections data. This comprehensive data set, spanning several years, provided a robust foundation for our analysis.
- Feature Engineering: We performed extensive feature engineering, including handling missing values, categorical encoding, normalisation, and scaling. This ensured that the data was in optimal form for model training.
- Model Development: Using both traditional and auto machine learning techniques, we trained a predictive model. We focused on XGBoost, a powerful machine learning algorithm known for its accuracy in regression and classification tasks. Our model achieved an impressive 80% accuracy in predicting not taken ups, lapses, and cancellations.
- Implementation: We integrated the predictive model into the insurer’s customer management system. This enabled real-time risk assessment and allowed the insurer to engage at-risk customers through targeted interventions such as personalised communications via WhatsApp, social media, and telephone.
Results
The implementation of our predictive analytics model yielded remarkable results:
- Significant Reduction in Lapses: Within the first three months, the insurer experienced a noticeable reduction in policy lapses.
- Improved Customer Engagement: Proactive engagement strategies led to higher customer satisfaction and retention rates.
- Increased Profitability: By mitigating not taken ups, lapses, and cancellations, the insurer saw a significant positive impact on their bottom line achieving a 10X return on their initial investment.
Lessons Learnt
- AutoML: Today’s leading edge auto machine learning tools can greatly accelerate multiple aspects of the implementation journey and are a great complement to the more traditional machine learning tools and techniques.
- Data Augmentation: If you are not satisfied with your model accuracy or results in training we can augment your internal training data set with quality data such as credit scores sourced from third parties to tremendous effect.
- Cost Effectiveness: AutoML allows for rapid training of predictive models and slashes the cost & effort required of traditional machine learning techniques.
Conclusion
This project underscores the transformative power of predictive analytics in customer retention strategies. The ability to anticipate and address potential lapses and cancellations not only enhances customer satisfaction but also drives business growth.
We are thrilled to have partnered with this leading South African insurer in this innovative venture. If your organisation is facing similar challenges, let’s connect to explore how we can leverage data to drive better decision-making and outcomes for your business.