Navy Federal Depends on TRIAD to Help Automate Human Decisions in Credit Card Management
MERRIFIELD, Va. -- An increasingly crucial part of the credit card portfolio management team at Navy Federal Credit Union is a piece of software that engages in something called adaptive analytics.
The technology is at the core of the TRIAD system used by the world's largest credit union to automate decision making about such things as credit limits, card re-issues and, coming soon, cross-sell offers for rewards and other incentive products.
TRIAD 8.2 is the latest version of the adaptive control system from Fair Isaac, the creator of the ubiquitous FICO scoring system. There are about 300 installed users of the TRIAD system, covering about 65% of the world's credit card accounts, says Linda Heger, business consulting director for the Minneapolis-based company.
"What this does is give our customers full control of their system, so they can develop their own strategies for their credit card portfolios. It lets them develop data-driven approaches to meet their own business objectives and modify their strategies as conditions evolve," Heger says.
At Navy, those evolving strategies affect about 780,000 credit card accounts that generate about 60 million transactions a year, says Mac Golden, the $28 billion CU's manager of credit card systems.
Credit union management is determined by two groups at Navy Federal, Golden says. "One is our strategy group that meets weekly and goes over things on a day-to-day basis," he says. "The second group is a steering committee that meets periodically to give guidance to the strategy group."
Those groups determine policies that are reflected in real-time decisions on the frontline that can have a big effect on the bottom line.
For instance, using TRIAD-generated data, the credit union decided to change a scoring parameter for approving authorizations at the point-of-sale.
"It was good for members, obviously, because they knew they would not be embarrassed at being declined at the POS, and it provided a certain level of trust in our product because of that," Golden says. "We saw an uptick in the neighborhood of one-half to 1% in authorization approvals, which doesn't sound like a whole lot, but for us, that was $40 million a year in additional balances coming in."
The TRIAD system allows such decisions to be made by human beings using vast stores of data to score and predict outcomes, as well as incorporates flexibility in such areas as automatic one-time waiving of credit limits for specific purposes.
Models that predict outcomes of strategy changes also can be run, helping to drive and refine decisions, Heger says.