When mining data, most bankers spend lots of time (or too much) fretting about how to get pertinent information into one place and what technology to use—all to the end of getting a good overview of customer needs and wants.
And while Chris Nichols, chief strategy officer at Winter Haven, Florida-based CenterState Bank, hasn’t completely solved those problems, he has another big issue: Once the data is gathered, what should he do with it?
“It’s almost an embarrassment of riches,” Nichols says. “We have so many projects in which the data can be used that we have to prioritize where we can gain the biggest advantages.”
That applies, by the way, across a broad swath of areas: from fraud detection to marketing, engaging customers to improving customer services. Nor is CenterState’s dilemma unique.
“Just about every bank I talk to has or is working on a data center or data lake,” says Adam Blue, chief technology officer for Austin, Texas-based fintech Q2. “But most are focusing on the mechanics of gathering the data rather than looking at what to do with it once they gather it. Just assembling the data is not accomplishing the task.”
Determining how to use data before gathering it could affect the sources used and the ways to integrate it, Blue adds.
JP Nicols, managing director of FinTech Forge, agrees that a clearly defined strategy may outrank implementing advanced technology in importance. “There’s no end to the list of things a bank can do with data gleaned from mining,” Nicols says. “But in prioritizing the projects, banks have to look at what they really want to accomplish. If you don’t have a strong strategy to begin with, you will just be chasing technology.”
Institutions such as CenterState Bank prioritize potential data uses through a “20 percent gain rule.” Nichols explains what this means to his financial institution: “You almost always have some success with using information from data mining. But some projects will only get you a 5 or 6 percent gain. We like to see at least a 20 percent improvement before we dedicate our resources.”
But even that rule has its exceptions. As he puts it, “If we are dealing with projects where big dollars are involved, a 5 percent gain might be worth more than a 20 percent gain somewhere else.”
'Data all over the place'
In the data game, success often boils down to evaluating customer preferences: how they like to be approached, what pitches they respond to and what they may want regarding the next best product. Other plans look at branch performances, examining such factors as parking lot access and the position of the front entrance, Nichols says.
Given its Florida location, CenterState has been evaluating the special traits and needs of snowbirds with a second home and who often open a small account there. These customers have true potential to grow their account size over time if the bank understands and can anticipate their needs.
But just knowing a lot about customers may not cut it if you fail to act on the new knowledge. Blue puts it this way: “You have the data and have analyzed it, now what? How are you going to change your organization to act upon what you have learned?”
As many banks pile on assumptions about their customers — in some cases unfounded—data mining works to either validate or debunk them. Blue notes that banks often believe that digital offerings are mostly for millennials. But in examining customer data, CenterState found a strong need for and interest in these services across the board. That means changing the marketing plan in terms of who’s targeted and what they’re presented with.
Nicols of FinTech Forge agrees. “You need to define the problems you have before you gather the data,” he says. “Banks have a lot of data to sift through. If you don’t know what you are looking for, you are just pulling threads.”
While CenterState Bank’s biggest problem may lie in defining and prioritizing goals, it has also needed to improve how it gathers data from multiple sources. Nichols offers a frank assessment of where the bank stands to gain ground.
“We probably have thousands of different databases, and the technology we use is often different
Creating a clean data lake
For the past two years, CenterState has worked to develop a “big data lake” that pulls data from all these sources, available to all authorized bank employees.
Nichols notes that for his bank, “Getting data clean enough that it can be used is a big problem. And getting historical data is often difficult. Some databases were only started as recently as 2012. Those databases don’t have information from our last big economic downturn and may not have enough experience.”
Nor is it enough simply to integrate the bank’s data — not if it hopes to find new customers or increase the wallet share of existing ones. Financial institutions must explore outside sources such as credit bureaus, social media posts and health care sources.
Technology, including integration software and artificial intelligence, can help sort and analyze the right information. But it can’t yield lightning in a fintech bottle. “AI holds a lot of promise, but we’re still very early in the game,” consultant Nicols says.
“A lot of banks just put the data in one big pile and expect the answers to leap out,” Blue adds. “Technology is a double-edged sword. It can provide answers but it can also be a distraction if banks spend more time on deployment techniques than asking, ‘What am I really hoping to accomplish?’”
This much is certain: It will take more than dredging a data lake or digging in a data mine to find out.
Want more BAI Banking Strategies? Sign up for our free newsletter!
Lauri Giesen has spent more than 25 years writing about banking technology and payments for numerous business and financial publications. In the 1990s, she founded and edited Financial Service Online, a magazine covering Internet-based forays into banking and investment services.