Case Study
Building a Predictive Model for Customer Retention
The Situation
A direct-to-consumer health and wellness brand with a global presence needed to identify potential high-value customers who had taken a self-assessment survey on the client’s website but did not buy a product. Merilytics (an Accordion Company) partnered with the client to identify key features of existing high-value customers and build a predictive model to identify prospective high-value customers.
Services
Reporting & Analytics
The Execution
- Built multiple classification models to identify key pre-conversion features of current high-value customers.
- Determined the best-performing model for each subset of customers using various modeling & optimization techniques.
- Recognized potential high-value customers and enabled the client’s sales team to customize their marketing pitch to acquire these customers.
- Identified a list of customers to retarget based on their prior website interactions and questionnaire responses.
The Results
The customers targeted by the predictive model had 10% higher retention rates than the regular customer base, leading to a ~5% increase in revenues. Overall, the client was able to improve customer acquisition and retention, leading to a ~10% higher Customer Lifetime Value.