13 minutes
Three credit unions share their analytics successes, efforts and plans for a data-driven future.
Credit unions are using analytics to do everything from providing personalized service to members they rarely see in a branch to identifying and contacting members who are in danger of running into financial trouble, even though the member may not realize it yet.
These efforts can be complex and expensive, especially for smaller credit unions that need to find a way to break down information silos and consolidate their data before moving on to analysis and action.
Paolo Teotino, chief product officer at CUES Supplier member Trellance, Tampa, Florida, says predictive analytics really is the new frontier for credit unions, but the cost of acting on their own is prohibitive for all but the largest. The good thing is, they don’t have to.
“We essentially offer AI (artificial intelligence) as a service to the credit union and spread the cost of having a team of data scientists and the costs of running a machine-learning platform across a variety of credit unions,” he says.
All credit unions need basic data management and basic descriptive analytics—“the process of using current and historical data to identify trends and relationships,” according to Harvard Business School—so they can have a sense of where their business is going, Teotino says. “What we say is that any decision without data is just a guess.”
Teotino notes that consumers these days expect companies to use the data they have to provide useful suggestions for everything from what show to watch to what products to buy. If your credit union isn’t making useful suggestions about products and services online, via text or in person, many members will conclude you don’t really know them or their needs, so it’s “actually a negative message that the member is getting.”
Some CUs are realizing how much valuable information they have about their members that other organizations and fintechs don’t have access to, and they’re learning how to use it, says Barry B. Kirby II, SVP of CuneXus Solutions, Santa Rosa, California.
“We’re seeing our clients start to leverage existing data they have on a member to be able to extend credit above and beyond just credit bureau data,” he says. For example, if a member has direct deposit and other products but is transferring money regularly to another financial institution, perhaps for an investment, the credit union can suggest an alternative that would keep the money in-house.
Kirby says a frequent question he gets from smaller credit unions is “How do we start?” He recommends selecting a couple of simple characteristics, such as searching for members who get direct deposits and then transfer the total off to another FI. Contacting those members to offer your services is a natural place to start.
To get a sense of the current landscape, we talked to three credit unions about their current data analytics efforts and where they are headed.
Pandemic spurs use of analytics to contact members in trouble Teachers Credit Union, based in South Bend, Indiana, with more than 300,000 members and $4.6 billion in assets, has been developing its analytics capabilities for many years, so it’s not surprising that it turned to data to help it serve members during the pandemic.
“We wanted to identify members who possibly had been impacted by sudden job loss at the beginning of the COVID-19 pandemic,” says CUES member Mitch Speer, manager of business intelligence and business transformation at Teachers CU. “Our goal was to reach out to those that may need help and educate them about the services and assistance that TCU offered during this unprecedented time.”
The challenge was that the credit union couldn’t contact each member and had no easy way to determine who might be at risk of losing their job; it had to look for signs—missed direct deposits, for example, or other atypical transaction behavior—that members were impacted without having direct information.
“We needed a way to identify which members were potentially impacted to build our database and divide the call lists between our branches based on where the members resided,” Speer says. “By determining the correct data points to use, we were able to effectively narrow down our list and make it a priority to contact those that could likely be affected.
“Armed with daily transaction data, types of accounts and basic member demographics, we set out to understand how to identify income/job loss by proxy of other data trends. However, running those queries on a daily basis was also not an easy feat. Member transactions were quickly shifting from inside our branches onto our digital and mobile platforms.”
Teachers CU developed an efficient process and produced a list of members who showed potential signs of financial distress. “As a result, we were able to make thousands of outreach calls to our membership to offer services to assist them during the pandemic,” Speer says.
Employees who worked in Teachers CU branches that were closed due to the pandemic joined in the outreach and thus were able to continue to serve local members despite the lack of in-person interactions, he adds.
He says members appreciated the effort the credit union made to stay connected and offer help, and the project gave its “front-line team more opportunities for relationship-building conversations.
“From an analytics perspective, we learned that communication with multiple parts of the organization in times of crisis is what will separate success from failure,” Speer says. “Knowing what to look for and how to identify those members was not just imagined in an analytics back-office meeting room. It came from talking with people who understand the intricacies of the issue more fully. This was a total team effort.”
But the credit union’s analytics initiatives didn’t start and end with the pandemic. Each of the past four years, Teachers CU has had a group of students from the University of Notre Dame work on a set of analytics problems. “Collaborating with a team of four to five students each year, we tackle a different topic using masked data from our credit union,” Speer says. “The problems presented to the teams have ranged from how to target high-propensity indirect memberships for cross-sales, developing a next-best-product recommender, branch transaction analysis and member retention.”
The credit union has also created three separate predictive models that are currently running in tandem.
“As part of an ongoing call campaign for our retail team, we have built in our indirect member propensity model to identify the indirect members most likely to grow their product relationship with TCU,” Speer says. “Also, we’ve layered our next-best-product recommender on top of that to provide specific, tangible solutions for these members to enhance their relationship with TCU. In addition, this year, we have developed a member retention model to help us identify the trends and characteristics of members that decide to leave TCU. As we continue to fine-tune this model, we are hopeful to capitalize on the opportunity to identify those members before they close their accounts.”
Boosting Revenue Through Better Analysis
Magnifi Financial, a $1.7 billion credit union with 74,000 members, based in Melrose, Minnesota, has begun using analytics to help grow its revenue, says CUES member Neal Kaderabek, chief digital/information officer.
The data is allowing Magnifi to study how members engage with the credit union now and use predictive analytics to recommend the next best product for them, he explains. In addition, the CU is using analytics to identify members who are in danger of leaving, so that it can find better ways to meet their needs.
Kaderabek says Magnifi is also planning to use analytics to pre-approve member loans and to approve loans that in the past it would have rejected. “Now, we’re bringing other criteria to the table and saying ‘OK, normally this loan would be denied, but we’re going to approve these loans based on their utilization of financial services that we offer.’ We’re using data to be more critical of how we are engaging with our members and trying to create a more personalized engagement.”
The insight that data creates for the credit union is appreciated, he notes. “It’s allowing us to be more proactive with our members, which is good because members want that assistance. They want us to know what they need to know at the right time, so they can act accordingly.”
Kaderabek says he expects it will take six to nine months for the credit union to see the full benefits of its use of analytics. “It’s a matter of execution, but certainly it’s pointing in the right direction, and it should take the engagement with members to a whole new level.”
Magnifi created an analytics team by identifying the existing data experts in each area of the organization, not hiring new staff.
“It seems like every department had somebody on their staff that, … amongst many of their duties, they were the data person,” Kaderabek says. “We cherry-picked some of our best data people and said, ‘You know, it looks like this is what really gets you excited about your work, so how about we have you do it full time?’”
This isn’t a journey most credit unions can take on their own. Kaderabek says Magnifi has obtained a lot of valuable information from other credit unions through Trellance’s user-group forums.
The next step for Magnifi is to embed the analytics information in its CRM “so that anybody who’s engaging in in the conversations with our members will have this 360-degree view,” Kaderabek says.
The information will also change the approach to online service. Magnifi is developing ways to present such suggestions as next best product without needing to interact with a credit union employee.
“That’ll be something new for us, to communicate that to our members,” he says. “It’s going to be timely and it’s going to be specific to them. Our members are not going to see it as carpet-bombing marketing that means nothing to them. Most of this information should be relevant.”
Strong Analytics Foundation Allows Action
Phillip Swift, VP/data analytics and business intelligence at $1.1 billion Centris Federal Credit Union, based in Omaha, Nebraska, with 123,000 members, says the key to a successful analytics program is the foundation: a single data warehouse that brings together all of the CU’s information.
“One of the major issues for credit unions is the need to get a single repository where you can pull data across your spectrum, to be able to blend loan origination and servicing and everything you need into one place,” Swift says. “A lot of folks still cobble together information from different places and use a spreadsheet. First thing you know, you go to a meeting, and finance has one report and retail has another, and they’re supposed to be the same data, but they don’t tie together.”
He says credit unions need a complete data strategy to ensure they can merge their data together and get a holistic view of members across their loan portfolio and their deposits.
Swift reports that Centris FCU had a data warehouse in place but decided to upgrade when it replaced its 30-year-old core banking system. He went looking for a system that provided analytics capability designed for financial institutions.
He settled on Trellance and its data analytics platform because of its focus on credit unions. “A member can be seen across all of their different pieces and parts they have with a credit union,” he says. “It’s really important that you find the right vendor that that will meet that single repository need, and Trellance does that for us.”
To set up a system, a credit union needs someone with an understanding of database administration and someone with business intelligence experience who can turn that data into actionable reports, Swift says. Depending on a credit union’s size and the number of reports it wants to generate, it may need several business analysts. Centris FCU has experienced staff in these roles who have been at the credit union for several years.
“The ideal staff to have would be a business analyst that can bridge the gap between the end user and the technology side,” he says.
Swift says Centris FCU uses analytics that identify members’ needs in tandem with incentive payments to retail staff so it can focus on selling products and services that generate profits. “We drive the ship to meet our members’ needs as well as make money for the credit union.”
The system helps retail employees increase their incentive pay while focusing on selling CDs one month or loan origination the next, depending on the credit union’s needs.
“It really helps to have that member analytics when they come in and have a meaningful discussion,” Swift says. “It really helps us to meet that member’s needs in a much more efficient manner.”
Centris FCU is currently using its analytics capability to watch for members who may be getting into financial trouble as inflation rises and there is a threat of a recession.
“We are proactive and understand our portfolio,” says Swift. “If things are going well right now, we want to make sure that we keep our eyes open in case they start going south. We’re starting to do a lot of credit analysis, looking at what the credit scores were at the time of origination and now.”
Also, Centris FCU is looking at early-stage delinquencies, with payments less than 30 days late.
“We’re not really sure how people are doing out there and how things are reacting. Analytics is going to be critical to get a proactive look. ... If things start going south, we definitely want to start contacting our members to help them before they get themselves in trouble.”
Swift says that in the past it was virtually impossible for credit unions to carry out this sort of analysis because they didn’t have detailed information in a format that could be used.
“I think that we’re serving our members well because we want to be proactive with them if they’re getting themselves into trouble. We certainly don’t want to be charging anybody off. We want to help them ahead of time.”
This approach helps differentiate credit unions from banks, Swift adds, which are much more likely to follow their strict protocols and not focus on helping their customers.
Analytics can also help credit unions serve their time-starved members who are often in a hurry, online or in-branch, he says. When front-line staff have the data about a member, they can make the most of an interaction by knowing what they need.
“You don’t get to see him very often, so this data is critical to help that member within that 30 seconds you might have,” Swift suggests.
Centris FCU also plans to use predictive analytics to identify members who look like they might be leaving. “We will build a list and reach out to say, ‘Hey, it looks like there is something going on here. Is there something we can do to help?’” cues icon
Art Chamberlain is a reporter who focuses on the credit union system.
