“Being that we're an insurance company, we have analytics in our DNA,” said Tim Peterson, AdvantEdge Analytics' president. “We have 80 years of history knowing how to assess risk and price risk, and 30 years' experience with direct marketing.”

Providing the analytical models and insights from the models is at the base of their work, but the most valuable step is helping credit unions use the information to form and execute business plans, Peterson said.

“There are lots of tools out there,” Peterson said. “But helping credit unions act upon the insights — that's where we see a pretty significant gap, a pretty significant opportunity.”

First Service Credit Union ($671.7 million in assets, 56,669 members) is in the early stages of using big data to target individuals for product offerings. The Houston credit union expects the data to provide advantages that will allow the credit union to gain market share so it can remain viable in a competitive market, EVP Mike McWethy said.

“Survival in today's market requires achieving economies of scale through growth,” McWethy said.

Achieving growth requires more effort because consumers want personalized offers timed to their needs and offering them the means to act whether online, on the phone or in person. Typical “shot-gun” blasting of advertising does not bring the same return on investment as targeting messages to candidates likely to qualify, McWethy said.

The credit union discovered this several years ago when it started pre-screening for auto refinance candidates and driving those consumers to a specific, single-product-focused microsite.

“We still love being the relational, fun-loving, hand-shaking credit union,” he said, “but we have to realize selling loan and deposit products is what allows our survival, so we must become more sophisticated at it.”

Michael Cochrum, CU Direct's vice president of analytics and advisory services, said credit unions have been developing their abilities over the decade to handle data past, present and future:

Telling the story of the past. “Credit unions have gotten good at analyzing past performance of loans,” Cochrum said.

Measuring the present condition. Credit unions are increasing deploying software and hiring consulting services to track aspects of their business, such as loan portfolio risks.

Predicting the future. Predictive analytics are especially useful for pricing loan rates and targeting customers, but very few credit unions are using these tools.

Measuring the present is often seen in the way credit unions value their collateral in a car loan portfolio. Instead of assuming a consistent depreciation rate across the portfolio, credit unions can tap into data on the resale value of each car by make, model and year. The information can be updated as often as the credit union wants.

Whether car collateral values are estimated by a formula applied to the entire portfolio or based on actual one-to-one book values often depends on the extent of risk of falling values.

“If your portfolio is weighted heavily in trucks and SUVs because of the demographics of your membership, it's probably not as important because they don't depreciate as fast. If they're weighted more in sedans, you're at greater risk because those vehicles tend to depreciate faster.”

Predicting good customer targets and matching them to the right loan is trickier.

CU Direct has been working with credit unions to set up dynamic pricing. Instead of basing a rate on credit score and the length of the loan, they can help credit unions price based on the predicted profitability of the loan, taking into account loan-to-value, and ability to repay based on their current financial situation.

“Instead of offering every product to every member, they're offering the right product to the right member at the right cost,” Cochrum said.

Credit scores are not enough because credit score is a predictor of default at a point in time.

“It's almost exactly like a weather forecast: It's going to be more accurate the closer you are to the storm actually occurring,” Cochrum said. “Credit unions are beginning to realize that credit score at origination changes dramatically throughout the life of the loan.”

The next step is for credit unions to use big data.

Big data doesn't mean a lot of data. It means tapping sources outside the credit union and linking it to members, Cochrum said. For example, it could mean studying Facebook “Likes” or other social media trends to see if there are ways to better identify members who might prefer certain products.

Is it creepy?

“I don't find it creepy at all,” Cochrum said. “My wife thought it was creepy that she could see on Facebook which of her friends were close by. I told her all she had to do was turn off her location services.”

Cochrum, who described himself as a “fairly prolific social media user,” said he would rather have someone from his credit union call him based on social media data that indicated he had recently been promoted, than to have the credit union send him a flyer in the mail advertising rates on CDs that he doesn't want.

“I'm kind of appreciative of people who use data and analyze my behavior so they're serving up offers that are important to me, and I'm not seeing offers that don't apply to me,” he said. “If you approach it correctly and not overdo it, there's a real benefit to your membership when you're targeting them better.”

CUNA Mutual AdvantEdge Analytics sees its mission as helping credit unions organize, translate and act on information.

Before announcing its launch of AdvantEdge Analytics in May, CUNA Mutual enlisted the help of the McKinsey & Company management consulting firm “to uncover the next-generation analytics,” Peterson said.

The CUSO met with more than 400 credit union executives and surveyed another 400 credit unions. It hosted four-hour “deep dive” strategy workshops at credit unions about how to launch this new business.

“We've been heads-down researching this with our customers,” Peterson said. “It helped us hit the ground running.”

AdvantEdge Analytics can leverage talents from within the 3,500 employees of its CUNA Mutual parent. But among them, 500 people are dedicated to the subsidiary as data developers, data scientists, data translators, implementation specialists and other roles.

CUNA Mutual had the foundation of resources it needed, but it was also looking outside for innovators. In February, it spent an undisclosed amount to buy SavvyIntel, a Chicago-based data analytics company. Blesson Abraham, its CEO, had worked for Baxter Credit Union, based in Vernon Hills, Ill., 35 miles northwest of Chicago, for five years before co-founding SavvyIntel in 2014. He is now AdvantEdge Analytics' director of analytics.

SavvyIntel enhanced the reporting capabilities of the AdvantEdge Analytics product suite that CUNA Mutual developed in 2016. The software is designed to allow credit unions to draw from a deeper pool of data from a wider array of sources than an individual credit union's resources would normally allow.

The CUSO can manage call center and email marketing. It can provide back-office fulfillment and build-to-order services.

“Rather than casting a lead for a loan off to a credit union, we work further in that process, proving the service to actually close the loan on their behalf,” Peterson said.

One need AdvantEdge Analytics fills is helping credit unions find new members and identifying which members might be at risk of leaving.

“Every credit union will tell you they can bring in members; their challenge is identifying and keeping high-value members,” Abraham said.

Another need is housekeeping. AdvantEdge Analytics helps set up systems for credit unions to organize data and import data into a unified environment, where it can be compared with third-party data.

The CUSO also helps set up systems to maintain data integrity and the quality of reports, visuals, dashboards and performance metrics.

“Once they're ready, we help them with that predictive piece as well,” Abraham said.

AdvantEdge Analytics buys big data from Acxiom, Experian, Epsilon and other venders to help credit unions discern demographic patterns of members, how many members in the household, the types of goods they purchase and where they buy them.

“External data can get pretty expensive,” Peterson said. “Collaboration is really key across credit unions. One role we can play is to be a facilitator in that collaboration by leveraging our scale to bring value.”