CO-OP’s COOPER Fraud Score Enables Real-Time Decisioning at WestStar CU
By Roy Urrico
Rancho Cucamonga, Calif.-based payments and financial technology company CO-OP Financial Services added a new instrument to its expanding fraud detection and prevention toolbelt with COOPER Fraud Score. The real-time predictive machine learning model, fully focused on the credit union industry, helped Las Vegas’ WestStar Credit Union address two critical problems: velocity attacks and fraudulent transactions.
“We help in the fight against fraud at the most fundamental level. Fraudsters are hard at work, figuring out how to exploit vulnerabilities,” said Patrice Alexander-Lee, product management director for CO-OP Financial Services, who is responsible for the COOPER line including Cooper Fraud Score. She described today’s bad actors as organized and sophisticated in their use of tools and technology. “Many leaders in the credit union space and in financial services are definitely interested in how we use technology as a new layer in the fight against fraud.”
Speed is a critical capability in today’s dynamic threat environment, said Bruce Dragt, chief product officer for CO-OP. “The beauty of real-time transaction data and machine learning technology is immediacy. The technology generates cost savings from reduced false positives, fraud chargebacks and fraud losses. Just as importantly, though, is the increased trust and reliability members gain when they experience fewer hiccups in the day-to-day movement of money.”
How COOPER Works
COOPER Fraud Score addresses problematic vulnerabilities, such as identity crimes and bank identification number (BIN) attacks, which involves a fraudster taking a card’s first six numbers.
“We're using machine learning to build out a predictive scoring model, where we can use that score in the transaction authorization stream,” said Alexander-Lee. The system, she noted, then delivers this score to a range of decisioning tools within CO-OP’s fraud prevention ecosystem covering credit, signature/PIN debit and ATM transactions. CO-OP fraud prevention consultants set custom strategies for individual credit unions.
She added, “As that card activity is coming in, we have the ability to look at it, generate a score, and then use that score as part of our real time decisioning. So, if it is looking highly suspicious, then we can stop it.”
Alexander-Lee explained when COOPER Fraud Score red-flags a transaction as potentially fraudulent, the credit union receives an explanation that provides context for a transaction decline with “reason codes” to explain the scores received, such as whether it was related to the amount or to the merchant.
Credit unions may choose to relay this level of intelligence to members, helping them understand the rejection of a particular transaction, improving the payments experience while adding to the trust members have in their credit unions.
A key differentiator for COOPER Fraud Score is the integrated team of credit union-centric experts working alongside the technology. Said Dragt, “CO-OP’s fraud team consists of data scientists, prevention consultants and detection analysts, all working to monitor COOPER Fraud Score and apply its use in fraud-fighting strategies. Continually learning from the solution’s data feedback loop, as well as emerging fraud trends and use cases, the team is highly focused on outcomes and model efficacy. Because they understand credit unions as much as they do fraud risk, the member experience is always a top priority.”
WestStar Beta Tests COOPER Fraud Score
The state-charted $260-million Las Vegas-based WestStar Credit Union revealed it was losing, at times, up to 20% of its net income to fraud losses. “Our revenue was being hit hard,” said WestStar President and CEO Rick Schmidt, who saw a greater volume and variety of fraud in his first year at WestStar than he had seen in 15 years prior, during which he led operations for a national, federally chartered credit union.
As a result, WestStar turned to COOPER Fraud Score by joining a handful of other credit unions serving as beta test sites. WestStar’s beta experience also involved more than engagement with the technology. The integrated CO-OP team of credit union-centric experts who work alongside the technology benefited from the credit union’s insights, perspectives and experiences.
Alexander-Lee described working collaboratively with WestStar credit union, sharing the performance and data analysis observed and discussing the financial institution’s pain points and objectives. “Credit unions have different goals when it comes to managing risk; their appetites maybe are a little more conservative. Some are not. We really wanted to hear from (WestStar).”
CO-OP learned among WestStar’s concerning pain points: velocity attacks, when a fraudster repeatedly submits a debit or credit card to make unauthorized changes or identify valid accounts, and fraudulent transactions that get through.
“It was really a collaborative session with them, with our fraud team; we were able to show we had an opportunity with COOPER for both of their pain points. We incrementally improve the detection on their activity,” Alexander-Lee said. Not only catching more bad transactions but also with a more accurate score, impacting less of the good transactions. “Because as you start to knock down good transactions it erodes some of the trust with members. We were able to demonstrate when we go live, we have the opportunity to make those decisions right. Those were probably the two biggest outcomes and learnings from the beta period with WestStar.”
COOPER outperformed WestStar’s existing fraud solution on all transactions where both systems scored suspicious transactions, according to Schmidt. “There was a lift of up to 30% in the set of velocity checking data where the other system tried and failed,” said Schmidt. “We didn’t know we could engage with something as sophisticated as COOPER.”
Schmidt suggested solving the problem of fraud is as much about meeting member expectations as it is about maintaining a healthy income statement. “It is a massive frustration when a member calls up and asks how we missed an obvious fraud, and we have no good answer for them. They lose faith. They expect us to be better at protecting our members than others.”
Fraudsters, however, are creative in their fraud schemes. “You close one door; they open a window. You are never going to catch it all and that is not the expectation,” Schmidt continued. “But we should be able to catch the low-hanging fruit, which are the misses that are impossible to defend with members. Throughout the beta experience, we saw that COOPER Fraud Score is smarter about preventing that kind of fraud.”
Knowledge is key. Schmidt maintained credit unions must figure out how to make its institutions – at the very least – the ones the fraudsters know to stay away from. Schmidt believes COOPER has the potential to improve the entire industry’s stance against fraudsters. He said the previous fraud solution caught roughly 40% of the credit union’s fraud. “So, it’s missing more than half of our fraud.”
COOPER Fraud Score went live on Nov. 16, 2021 for general availability to CO-OP clients, with WestStar now taking advantage of it in production, according to Alexander-Lee.
Nonstop Learning Helps in Fraud Detection
Part of machine learning’s role is understanding patterns and behaviors. “What looks normal for Roy could be completely different than what looks normal for Bill,” noted Alexander-Lee. “We understand those patterns and you are going to get a score output that ranges anywhere from 1 to 999, the higher the score, the greater likelihood of fraud.”
Alexander-Lee added, “With machine learning, one of the great things is it is continuously learning,” said Alexander-Lee. She added over a period of time the analytics gathers historical transaction patterns. “Being able to have that understanding is really key in being able to react quickly.”
Another aspect contributing heavily to CO-OP’s fraud prevention is the people component — the team working directly with credit unions. Alexander-Lee summed up the company’s involvement: “They are managing the strategies, they are understanding what is happening within the industry, what is happening within CO-OP, so that we can update the strategies based on what is happening in the market. We're actively looking at trends, working closely with our data science team to make sure that we understand any new use cases that come up to make sure our model is effective.”