Covid-19 halted spends for a couple of years. And thus, CASA ratios of banks went up. They are now back to pre-covid levels. As a dominant regional bank navigates the change, a retention scorecard is now an integral part of their CASA strategy.
Context
With a 30,000 crore CASA book, a diverse tier-II geographic, mixed demographic customer base the newly formed data and analytics team of this bank looked to scale this mountain. They wanted to build models in the bank. And were kind to explore a partnership with us to get them started.
Challenge:
CASA balance movements for a savings base are nuanced. Driven by both macro and micro, external and internal they are not easy to predict. It’s like the calm and poise of ocean, that hides many turbulent undercurrents. Oftentimes an outcome is straightforward – someone buys a policy or not, clicks a link or not, makes the payment or not. Capturing the downward or upward trend of savings account balance needs business acumen. And then to predict which way it would go is both craft and science.
Solution
We worked with the bank’s fledgling technology and data science teams and cast a wide net to observe balance movements. Transactions data was complimented by demographics, account instructions, product holding and more. A super powerful gradient boosting algorithm achieved satisfactory and meaningful gain, with a bit of explanation too, and put a smile on the MD’s face.
Benefit
The first pilot of the model has kick started on a select priority customer base. This base is reachable both by relationship managers and the broader digital CRM system. For more on details please do not hesitate to write to us.
