The challenge
A global corporate-card business was leaking revenue two ways. Employees were routing spend onto personal cards to earn their own points, so corporate cards quietly went dormant. And large cardholders were cancelling — but only visibly, after the fact, once the fee and spend revenue had already gone.
What we did
Targeted dormancy by behavioural signal — spend slowdown, replacement cards not re-ordered, weak ramp-up on newly issued cards. Redesigned the value exchange rather than fighting employees, enabling personal use of corporate-card points to win the spend back. Built a predictive cancellation model that read each cardholder’s seasonal spend pattern to flag likely cancellations before they happened, feeding a monthly at-risk list for retention.
The outcome
- Saved 35% of attrition cases, protecting $5m+ in fees every month
- Reactivated 2,500 cardholders who still held a card but had stopped spending
- Retention shifted from reactive to predictive — acting before the customer left