The Card Services for Credit Unions paper, “Member Segmentation Improves Your Bottom Line,” said using data to segment member bases can help a credit union target members with the best products and services.

“There are plenty of ways to use segmentation analytics to find members that will respond to new opportunities and cross-selling efforts. For example, consider high-usage cardholders,” the association wrote. “Notice their spend patterns,” the paper quoted Dean Knudtson, senior consultant for CSCU. “Can you offer them a business card and realize a higher level of interchange? Or look at the spend patterns with your rewards cardholders. Can you shift their spend patterns to higher interchange merchants by segmenting them, and offering bonus points to shop at higher interchange merchants?” 

Knudtson also suggests breaking out lower FICO score members. “It might make sense to offer these members a secured card with limits equal to a deposit, thereby reducing your risk exposure while offering a great product to members who may be rebuilding credit or establishing credit for the first time,” he adds. “This is also a perfect way to stay true to the industry's mission of people helping people.” 

The sorts of member segmentation and planning that both papers describe fall loosely into what is commonly known as big data, a popular term for the collection and use of massive amounts of transaction and payment data and to help assess and predict consumer attitudes and behaviors. It is the information science behind such online giants as Google and Facebook, as well as all three national credit bureaus, and now, the consultants suggested, it could be put to use helping credit unions better serve their members.

“This technology is new enough that a lot of our work with an interested credit union revolves around helping them understand the questions that could be answered and then helping them formulate those questions,” said Todd Herren, chief technology officer for TMG. 

Most credit union executives, Herren explained, are used to thinking about data primarily in demographic terms and in terms defined by the national credit bureau's credit scores. Most have no idea of how powerful a tool in card management data generated every time a member swipes their credit card could be. For example, a credit union contemplating offering professional sports teams tickets in a sweepstakes to drive credit card use could mine its cardholder data to find out how many cardholding members use their cards to routinely purchase sporting event tickets, sports equipment or gym memberships. Presumably, those members might use their cards more often to try to win tickets to a football or basketball game, he noted.

“Segmentation allows you to spend valuable marketing dollars on campaigns that target those members and cardholders who are most likely to respond, which greatly improves your marketing ROI,” says Barney Moore, senior portfolio consultant with CSCU. “Improved revenues aside, segmentation is simply a better way to understand the needs of members in general and add value to their member experience.”

With a combination of the right analytics tools, thoughtful evaluation of the results and solid PAU strategies, credit unions can execute more effective targeting and overall growth. “It's a good exercise for better understanding the make-up of your portfolio and managing it more effectively. The key is not treating all accounts the same way,” added Moore. 

The technology is sophisticated and specialized enough as well, according to Karan Bhalla, consulting director for IQR, that it remains beyond the reach of all but the largest credit unions, which have the resources to seek out and keep employees with the specialized skills to do this sort of work.

In addition, the technology is not without its critics. Critics of the technology in other industries have noted that getting big data does not guarantee the end user knows what to do with it, and privacy activists have questioned the necessity of collecting such a large amount of data about what could be personal behavior.

But Bhalla dismissed the privacy concerns, noting that there are already opt out procedures in place in several areas that let consumers keep their data from being collected in big data searches and he welcomed still others.

“From our perspective, opt-out procedures are a good thing since we only want to work with consumers who want their information to be part of bringing them better products and services. Every consumer who doesn't want that and who opts out shrinks the overall pool that I have to work through to find the others.”

But Ondine Irving, a noted credit union card consultant who helps credit unions manage their card portfolios by managing their processing expenses, questioned the wisdom of credit unions adopting such an sophisticated and potentially expensive technique to drill down in member data when many still do not control their costs at the portfolio level.

“The overall portfolio analysis is the first step to understanding cardholder behavior–and unfortunately the status of quo of garbage in/garbage out is creating an industry of inaccurate data for [use in] program analytics,” Irving wrote in an email about the use of big data. “For example, when credit unions do not zero out outstanding credit lines on closed, lost, stolen and bankrupt accounts, the key metric of credit utilization ratios and average credit lines misleads the credit union into making wrong marketing decisions. Many credit unions are inaccurately reporting total credit line liability by as much as 20% due to lack of overall card program management techniques and best practices. The processing system does not realize this. Therefore, the data the processors use is inaccurate to start with.”

“The bottom line is there are two types of cardholder behaviors–transacters and revolvers. Considering up to 70% of a card portfolios revenue should be coming from finance charge income and card loan balances and less than 25% of revenue is derived from interchange income, it would remain prudent for credit unions to focus on cardholder behavior at the portfolio level, providing the input of data is correct in the first place.”