"Speed to decision is so key": A CU Fireside With Lending Expert Don Arkell
"We've been telling credit unions for 15 years that these days are coming. It was gradual but now it's right in everybody's face: Speed to decision is so key."
This is Don Arkell talking. He's the owner of CU Lending Advice and a veteran credit union executive who's now a busy consultant to underwriting teams across the country. We knew we had to have him on for a CU Fireside and indeed it was a blast having him bring us up to speed on CU lending today and into 2022.
Don's a big believer in innovation and a champion of those leading the change to make lending more digital, fast, and efficient. He says real loan growth starts at the top, with leaders who want more loans and want to say yes to more members. "There's a whole generation of leaders coming in that have different expectations than folks in the past and they're helping driving change. They are consumers first and just happen to work at a credit union."
"The role of machine learning in consumer underwriting is the ability to extend your credit risk appetite across space and time."
A big part of that change is knowing who your competition is when it comes to digital experiences. That means shifting from thinking that your CU is competing with banks to realizing that the real comparison should be with Amazon and Uber: fast, efficient, informative, and 100% digital. Says Don, "I've got a great credit union client down in Florida, and the CEO's mantra is to that he wants his members' last experience at Amazon to be their first experience with his credit union."
What does that mean? "It means understanding that members are shopping first online and in this industry we've romanced this idea that people want to come in and sit across the a table for a $3,000 loan from us. And it's just not happening that way," says Don, who reminds us that even before the pandemic 85% of loans were originated by a first point of contact online or on the phone.
Being more digital and fast means getting more confidence in your scoring in order to boost automated decisions. "There are a lot of good solutions out there for automated approvals and Zest is great one," says Don.
What's keeping CU lenders of all sizes back from deciding faster? "The existing processes that we use to auto-approve loans today misses a lot of great loans," says Don. "With the standard rules we use, we may be able to tweak them, but they can't catch the edge cases. And lending is an edge case in many circumstances. So applications that miss our cutoffs end up falling [into human judgmental review] which takes time and employee time is a cost. This idea that our members have to wait for a loan decision is what makes us not competitive."
To decision faster and with more confidence, says Don, lenders are going to have go beyond the standard 20 - 50 variables they look at today. The limits of current scoring tools breaking down beyond that number of data points. As a result, technologies like machine learning will become necessary for CUs to harness all that data and approve more loans. "More importantly," says Don, "the role of machine learning in consumer underwriting is the ability to extend your credit risk appetite across space and time." In a way, the use of more data and deeper insight replicates the traditional intuition of the human underwriter. Trusting the algorithm is like going back to the future.
Don's Tesla can change lanes and navigate around corners all by itself. "Certainly those same algorithms can decide if we can give a 10-year member a $5,000 loan?"
Watch the whole interview if you want more great insight. We touched on many other topics, including:
• The outlook for auto lending and the recapture game, aka "Grand Theft Auto"
• The role of technology at lifting approval rates
• How to kick the tires of any machine learning lending solution
• Why going down the credit spectrum is key to loan growth
