You may have heard the term “big data” and wondered what it means — and how it diff ers from the regular, plain old data you’re used to seeing. It’s really a matter of scale: Big data refers to huge, complex data sets from diff erent sources that can be analyzed for patterns and trends. Credit unions use the results to:
- market the right services to the right members at the right time;
- develop the products their members need most;
- know when members are about to jump ship so they can try to retain them;
- pinpoint when members change jobs, move or are about to make large purchases;
- identify which members are at risk of defaulting on a loan;
- know where to build new branches; and
- drive more engagement with members.
We interviewed several data experts and credit unions that are collecting and analyzing massive amounts of data and asked them to share their best practices.
Delineating the Data
Before you can start collecting data, you need to know exactly what you want to collect and why. Th e following categories can be mixed and matched to meet your credit union’s particular needs.
Demographic and financial data
Demographics are what we often think of when we hear the word “data”; that is, where members fall in terms of age, income, education, homeownership, job status, marital status and so on. Combining this information with the other categories can turn up valuable insights.
Knowing members’ demographics can also help credit unions send the most relevant promotions and messaging. InFirst Federal Credit Union in Alexandria, Va., for example, delights members by sending them a virtual birthday cake via their mobile app on their special day.
Then, there’s the financial data that credit unions collect, which includes credit scores, the value of members’ homes and automobiles. InFirst uses Visible Equity data analytics software to glean this information. “It has been really one of our more successful products as far as being able to get people’s credit scores, the value of their home and even information about their automobiles,” says Marty Wye, InFirst’s president and CEO. Data such as this allows financial institutions to determine whether a home equity line of credit (HELOC) should be revoked because the member’s home value or their credit score has decreased. Monthly, InFirst’s Member Solutions department (collections team) uses Visible Equity to actively manage risk by identifying and flagging real estate loans that should be closely monitored based on the changes in home values and credit scores.
Payment data
Knowing what your members are paying for has clear benefits. For example, FedChoice Federal Credit Union in Lanham, Md., runs a report out of its Core that identifies and categorizes outgoing ACH transactions. “At this time, we’re looking for members who are making payments on student loans, so we can target them with student loan refinancing,” says Mark Shell, FedChoice’s information and creative services director. This same strategy could also be used to identify car payments and mortgages outside of the credit union.
Knowing where your members hold cards and accounts at other financial institutions also provides benefits. First, it helps credit unions figure out where they need to change or improve their own offerings. Take Michigan State University Federal Credit Union (MSUFCU) in East Lansing, Mich. “We included payments data in our analysis to determine where members are sending payments to other cards that are similar to ones that we would like to provide as a product choice,” says April Clobes, MSUFCU’s president and CEO. “It’s possible if they’re not using our cards it’s because we don’t have a level of rewards or high-end services that you receive from, for example, American Express or certain Capital One cards. We could use their payment data to determine if they have those products elsewhere, to get a sense of what the incremental revenue value could be for us.”
Second, this data gives you a peek into your members’ needs so you can create relevant products for them. “We’re evaluating direct deposit, account balances, transactional data and demographic data to determine how people are using the product,” says Clobes. “Then, through the transaction data, we’re determining when members are making transfers through other financial institutions and trying to evaluate what they might be utilizing elsewhere as a checking product.”
MSUFCU will be combining this information with data from focus groups to determine whether they should offer rewards checking, identity theft protection plans or a high-rate checking account, and what their members’ fee tolerance is for such checking accounts.
Customer behavior data
Understanding member behaviors such as saving, spending, insuring and investing helps credit unions figure out when members are experiencing major life events — so they can promote the most suitable products to them. “We often see people saving more when they’re looking for a new home because they’re looking for that down payment,” says Aaron Junod, vice president of product management at Geezeo, a fintech partner for financial institutions, core processors and complementary digital banking solutions. “When people are about to look for a HELOC, you’ll see a lot of shopping at Home Depot or Lowe’s because they want to expand their home.”
Where members are right now
Geofencing is using GPS or radiofrequency identification to draw a geographic boundary that can trigger a response when a mobile device enters or leaves it. MSUFCU, for example, uses geofencing to pop up a home loan ad on the member app whenever a member is near a Parade of Homes event. Credit unions could also use location data to target members when they’re inactive and using their mobile devices, meaning they have the time and opportunity to engage with you — such as waiting in the carpool line at an elementary school.
Email and ad click-through rates
Gathering statistics on how many members are opening your emails, clicking on links in emails and clicking on online ads helps credit unions know what kinds of offers and messaging work for their members. Faye Fardshisheh, first vice president of marketing at United Nations Federal Credit Union (UNFCU) in Long Island City, N.Y., uses data such as clickthrough rates to see not only which messages are resonating, but also how the credit union should further refine its communication to improve on those results.
FedChoice monitors web traffic, including where the traffic came from and what website pages people visited. “We then do a correlation with membership applications to determine what percentage of visitors become members,” says Shell. Reviewing the entire data chain, then, you might discover that a particular email campaign drove X number of members to the website, whereas Y percent of them responded to the offer.
It’s not all about the clicks, though; even knowing how long viewers watch your videos can be helpful data. By analyzing drop-off reports on FedChoice’s video commercials, for example, Shell discovered that shorter is better. “We used to do a lot of 30-second commercials,” he says. “But the data showed a high dropoff after 15 seconds, so we moved to 15-second commercials and increased click-through rates from around 0.75 percent to more than 1.5 percent.”
Putting the Data to Use
Once you know what types of data are most relevant for your credit union’s needs, what’s the best way to collect it? And what do you do with it once you have it? We asked our experts for their best practices in data collection and analysis.
Use third-party platforms to get the whole picture
Unless your credit union has the resources for customized software and a complete data analytics team, it can help to bring in third-party partners that integrate with the systems you use and offer different types of data or analytics services.
The quarterly surveys UNFCU conducts, for example, are administered by a third party. FedChoice uses a service from Ser Technology Corp. (Ser Tech) that identifies members who have applied for a loan outside of the credit union and automatically targets them with marketing materials in an effort to recapture their business.
Mix it up
Credit unions have access to demographic, transactional, payment, behavioral and other types of data from member records, third-party platforms, marketing campaign reports and member surveys. Unfortunately, though, while many credit unions are on top of collecting data, they fall short on the analysis part of the process. “In a lot of cases, data is stored in separate systems, and it’s often difficult to query, difficult to link up and, in general, difficult to use,” says Junod. “They’re collecting it for different purposes and then not necessarily using it to its greatest potential.”
Instead of sequestering each type of data in its own silo, combine this massive amount of information to tease out the insights you need. “We conduct surveys and roundtables, and then combine this qualitative information with quantitative data, covering things such as transaction history, adoption rates and satisfaction rates,” says Fardshisheh. “Then we take a holistic view and analyze all of this to help us determine how to best meet our members’ needs.”
Don’t analyze in a vacuum
Data analysis can reveal basic information on how your credit card is doing in various categories. But what does that mean for your credit union? “Doing data analysis in a vacuum can yield your own results but may not yield the full picture, depending on what you’re looking for,” says Junod.
That’s where peer comparison data comes in. “It can really help you understand where you might be pulling away from the pack or where you may be behind,” says Nicole Jass, senior vice president of product at payment service provider Worldpay. “Those could be interesting places to take action on.” Third-party systems such as Worldpay and Geezeo can help bring in this comparison data, since their data is fed by multiple institutions.
Keep it clean — but not too clean
Keeping data clean is an enormous task, but it does improve the results. If you can’t parse out the five different codes under which Starbucks may appear in your database, for example, you won’t get a full picture of your members’ spending habits. A third-party system or in-house analyst can often make sure your data is clean.
Keep in mind, though, that you’re not looking for 100 percent accuracy. “You’re really looking for general trends, and small anomalies shouldn’t throw you off too much,” says Jass. “It’s just important to know that these issues exist, because if you think you’re looking at 100 percent absolute truth, you might guide yourself in the wrong way.”
Don’t let perfect be the enemy of good, or you’ll never get to the place where you’re deriving actionable insights from your data.
Take small steps now
The most important piece of advice experts have for using data to drive business results is to just get started. Says Jass, “Anyone who is early on in the data journey can’t get too frustrated with that process. Making data-driven decisions is so valuable and important that it’s worth it.”
From the May-June 2019 edition of The NAFCU Journal magazine.