Consider the simple goal of a credit union executive: She wants to reward her institution’s most valuable members by paying a special dividend to them each year. But who are they?
They’re the members who help the credit union stay capitalized by maintaining an average daily balance of at least $1,000, or by having an outstanding loan balance of more than $5,000. They also save the institution money by using digital services; they haven’t visited a branch or called the call center in at least a quarter.
So the task seems simple: Just identify those members, and pay out the dividend.
Now consider the plight of the credit union data analyst charged with producing the list. It’s not so easy, because the relevant data is maintained in different systems. Combining those data sources to generate an answer to a seemingly simple business question is a challenge.
“For the business, this seems like a simple problem,” said Raj Rathi, director of data management with AdvantEdge Analytics. “‘Why can’t I get this list of people who fit this criteria? But it’s a technical, difficult problem for the data team.”
Credit Unions Need Advanced Data Analytics to Compete
Speaking at the annual meeting of the Credit Union National Association (CUNA) Technology Council in Chicago, Rathi said credit unions must make data analytics central to their growth plans because their traditional selling points are being challenged by emerging competitors and the advance of time.
Online-only banks and other fintechs are able to offer the kind of attractive interest rates that credit unions have long been known for. And while their reputation for personal service remains good, customers in the digital era are used to serving themselves online — and want to do so on their own time, not during bankers’ hours.
To compete, credit unions must use data to understand their customers on a personal level so they can develop digital products that meet their needs.
“What we used to do in the branches — we knew Joe, we’d talk to him and give him personal service — we now want to give people the same the same experience in digital channels,” said Rathi.
But that requires a sophisticated approach to data analytics and governance. Rathi argued that credit unions should begin with a data strategy that answers basic business questions (like how to reward the most valuable members) and then use those questions to build a multiyear data roadmap.
After that, he said, an organization’s “data journey” moves on to answering more difficult questions about data governance and how to manage both raw and structured data, before eventually achieving the high-level benefits of guided and advanced analytics.
Those benefits include the ability to identify trends, forecast outcomes if those trends continue, and prescribe solutions. But it all begins at the most basic level, Rathi said: “To be successful, your data journey has to start with strategy.”
Why Credit Unions Should Consider Data Lakes
Many credit unions, if they maintain data in traditional data warehouses, will find themselves stymied by an inability to answer new questions and solve emerging problems. That’s because the structured data model they’re using lacks the flexibility to adapt to new business challenges.
A key solution is the implementation of a data lake. Unlike a data warehouse, which maintains data process for a specific purpose, a data lake is a vast pool of raw unstructured data. For credit unions, data lakes don’t replace data warehouses — they complement them.
“The data warehouse isn’t going away,” he said. “But a modern data management solution typically includes a data lake, which allows you to take in data from multiple sources, both structured and unstructured. The idea is to free the data from the rigid data model that exists inside the data warehouse.”
As Amazon explained on its website, “different types of analytics on your data — like SQL queries, Big Data analytics, full-text search, real-time analytics and machine learning — can be used to uncover insights” with data lakes.
Using a data lake in conjunction with a data warehouse allows credit unions to use their data to identify and solve any business problem or question, said Rathi, because they’re flexible and extendable.
“In data management, we need to listen carefully to what the data problem really is, and focus on solving that,” he said. “Make it simple for end users: There will no doubt be complexity behind the solution, but for the end users — our executives customers and business users — the solution needs to be intuitive.”