Lead Big Data. Don’t Let Big Data Lead You.

Amy Sink, CFO at Teachers FCU, advises credit unions on how to best use (or not use) member data.

Yun Ma

By Yun Ma


Big data is a new name for an old concept. Just ask Amy Sink, CFO of Teachers Credit Union ($2.4B, South Bend, IN).

“Big data is the new buzzword,” Sink says. “Here at the credit union, we’ve had a data warehouse for about a dozen years, probably longer than that. Back in the day, it was called the data warehouse, and before that, the repository.”

Defined as high volume information assets that enhance business insights and lead to better-informed decision-making, at credit unions, big data is generally collected from two sources. It is either generated by in-house systems, such as member transaction statistics, or through third-party sources, such as marketing groups and GPS.

Whatever name it gathers data under, in Sink’s 25 years at the credit union, it has never been short of data. Teachers uses it to analyze member spending and find ways to improve pre-existing programs or devise new ones.

“My guys put together data on everything, from investments to process improvement to contracting,” Sink says. “We use member data extensively for all of that.”

Wading into big data can be overwhelming, so Sink offers tips on ways credit unions can make it work for them.

Data is endless; time is not.

Too often, Sink sees analysts generate and look at data without a clear idea of what they’re looking for. So begin with an end goal to avoid getting buried in mountains of facts and figures.

“You have to start with the thesis before you try to solve for X,” Sink says. “Some people just gather data and give you a bunch of reports. That’s nice, but what does it mean? What do you do with it?”

After gathering data, make sure it gets into the right hands — such as an analysts who understand how to interpret it — to avoid being misled by it.

“My finance people are great with big data,” Sink says. “They understand what our balance sheet is, where we need cash, and where we need liquidity. They understand what drives fee income. They can say, ‘Hey, if I look at this fee income and compare it to this, that might tell a story.’ That’s the kind of people who need to interpret data.”

Big Data Or Common Sense?

Take this typical scenario: A market research group advises a financial institution on what products it should sell to certain customers. The research group claims that if a member has a mortgage and a car loan, then based on big data studies, there’s a 20% chance that member will be willing to open an IRA. Therefore, the financial institution should pitch a retirement account as the next product.

Sink doesn’t believe using big data in this way is helpful.

“If I get a list of 1,000 people and offer them all an IRA, 200 of those people will take it,” she says. “Two hundred of them will take a car loan. Twenty percent of people are going to take just about anything you offer. If you ask enough people, you’re going to get them. Big data doesn’t have anything to do with that.”

Sometimes common sense offers just as much insight as big data. When analyzing data, then, credit unions must differentiate whether the conclusion is insightful or obvious.

“Big data groups are out there claiming they can tell you what the next big product is,” Sink says. “I think that is hocus pocus.”

Big data isn’t magic. It can’t predict the future. Credit unions will do well to avoid allowing the smoke and mirrors of pseudo-analysis to cloud common sense judgments.

Put Big Data Into Practice

Instead of using big data to predict member habits, study member habits first and then use big data to determine how to modify habits. For example, use big data to determine how to customize a member’s products and services to be more sticky and user-friendly.

Sink and the Teachers’ team used big data to establish a member rewards program with a local grocer. Teachers began by studying where members were using their debit cards. They pulled the 20 most-used merchants, weeded out big box retailers, and identified local vendors, such as the grocery store chain Martin’s.

Teachers analyzed its members’ debit card transactions to determine how many of its members shopped at Martin’s, the average amount spent on each visit, the frequency of the store visits, and the average age group of shoppers. Teachers then used the data to convince the grocer to collaborate with the credit union on a member rewards program.

“Once the data is in front of people, it starts to make sense,” Sink says.

At the end of 2012, Teachers Credit Union and Martin’s partnered on a member rewards program in which Members, Martin’s, and the credit union share one-third of interchange income generated by using the Teachers debit card at Martin’s.

Although the program is new and the credit union does not yet have solid performance figures, it has been successful enough for Sink to consider implementing a similar program with other vendors, especially ones popular with younger demographics.

“We want members using our debit cards,” Sink says. “How do you do that? You make offers at vendors they frequent. We need to be more relevant. If they use our card, they have to have their money here. If they have their money here, then we have opportunities to sign them up for other products.”

Big data helped Teachers figure out which merchant to work with and convince that merchant to partner with the credit union. After that, the credit union’s creative savvy and understanding of members’ needs helped make the program a success.




March 28, 2013


  • Excellent perspective! So many people get "big data" completely backwards and end up with a bunch of completely obvious conclusions. As Sink said, "You have to start with the thesis before you try to solve for X.” Read more: http://www.creditunions.com/articles/lead-big-data-dont-let-big-data-lead-you/#ixzz2OqoxhJGF
    Brian Wringer