How Prophesy Helped Predict Customer Churn To Retain At-Risk Customers

This Merchant Credit Card services division of a major US bank provides secure and innovative payment solutions to power more than 1 million growing businesses worldwide. The company needed an analytics solution that would help them leverage their existing data, use advanced analysis to identify key retention opportunities within the small merchant segment, and provide fast, accurate, and actionable insights.

Challenge

As an innovative player in the category that processes 3 billion transactions annually around the world, the company needed help harvesting useful intelligence from the massive amounts of data related to their services. With the potential to save millions in lost revenue and understanding the cost of new customer acquisition, they wanted to know who was planning to leave them and what could be done to prevent it.

For large clients in this category, this responsibility rests upon a dedicated account team that maintains healthy client relationships, manages operating procedures and fulfills contractual obligations. However, this is not the case for small merchants.

An early warning system that provides an opportunity to deliver a higher level of service would not only help retain merchants with a high likelihood of defecting, but would also become a competitive advantage.

Solution

Scintel was engaged to help identify these at-risk merchants so that a service recovery program could be implemented for course correction and retention. Using Prophesy, a predictive model for merchant churn was created with the goal of accurately predicting the specific merchants and calculating the potential business value across a segment of the company’s merchant population. Instead of relying on hunches – or overlooking important factors altogether, Prophesy identified actionable patterns in the data that indicated when a particular merchant would be likely to lapse.

The analysis revealed that churn was likely for merchants paying specific fees, those recruited through a specific promotion, and also according to the types of payment they accepted. But Prophesy uncovered something else that was previously undiscovered: a strong correlation between the type of software used by merchant and their churn. Through its analysis, Prophesy was able to predict the vulnerable merchants with a hit rate 7X higher than achieved prior to the predictive model.

Results

Close collaboration with Scintel Prophesy led to tremendous improvements in the accuracy of identification and potential to retain relationships with merchants.

$350k

From the initial segment, Prophesy identified over $350,000 in potential revenue loss.

700%

Prophesy produced a 700% increase in accuracy of predicting churn.

$6M

Applying Prophesy to the entire merchant population would result in potential recovery of over $6 million in revenue loss.

10x

Rolling out Prophesy to the entire merchant population would result in a projected hit rate 10X higher than original process.