Oregon Community CU's Business Intelligence Unit Drives Growth
EUGENE, Ore. – Oregon Community Credit Union has spent the past five years grooming its internal business intelligence department, and management says the unit has helped every area of the $1.3 billion CU improve and grow.
While many CU executives are generally familiar with the terms "data mining" and "predictive analytics," most are unfamiliar with how using data can reshape how a credit union does business. Oregon Community said its business intelligence department has allowed it to make "impressive gains" in improved decision-making, segmentation, pricing, financial forecasting, marketing and more.
OCCU boasts that its business intelligence department grew from a single person running reports to the "critical" area it is today with five full-time professionals and three interns managing 170-plus projects per quarter. The team's representatives hold degrees in economics, applied mathematics, computer programming, business administration, architecture and applied information management.
A few examples of the department's recent successes:
- By using “Next Best Product” segmentation modeling, OCCU’s business intelligence was able to predict which households would best benefit from its credit card ITAs. One recent mailing resulted in a cost-per-acquisition of just $66 – which the CU said is far below industry standards.
- When preparing to roll out a significant deposit product enhancement in the first quarter of 2016, business intelligence played a critical role in helping OCCU discern the profitability impact of various scenarios.
- Marketing asked business intelligence to estimate the monetary value of new memberships using account-level profitability figures over the course of three years to discern break-even points.
Ron Neumann, OCCU's chief financial officer, is given credit for supplying the "vision" for expanding the business intelligence department. Dave Schiffer, director of finance, has day-to-day responsibility for the group and is actively involved in much of the financial forecasting/modeling efforts. Casey Foltz, business intelligence manager, has extensive experience in data mining, information design, multivariate analysis, data architecture and database programming.
In the Beginning
When the CU completed a core conversion in early 2009, the business intelligence department was born. Neumann joined OCCU in September 2010. He told Credit Union Journal the original vision for the unit concerned understanding data and managing data better.
"When I started there were three people in the business intelligence group," he recalled. "We had 800 reports, and the first thing I said was to cut that down to about 100."
From data centralization, the BI group began building trust in the organization. The CU's loan portfolio was growing, and analytics offers "many different ways to slice and dice the portfolio," Neumann noted.
"We began to do more and more sophisticated work," he said. "Centralization allowed the organization to put demands on the group. As we understood the demands we were able to bring better tools.
"We were very deliberate in building the business intelligence area so we could leverage the data to better serve our members and be operationally efficient," Neumann added.
Schiffer said: "I joined early on in the journey. It was recognized early on there was real value coming out of that group. At that point the organization realized there could be a real competitive advantage in leveraging the reports. There was real opportunity to have information that could be leveraged."
Foltz noted he also became involved early on, and said the BI department has evolved from just cranking through reports – which he said many credit unions believe is business intelligence – to operating at a "higher level."
"We use a high degree of communication," he declared.
Foltz said the group has grown significantly from 2010, which saw elementary data usage, to predicting what type of branch a given member would use. He said this allowed OCCU to divide its branch network into "anchor" versus "satellite" branches.
"The projects are much more complex today than they were four or five years ago," Foltz said. "The first couple years we were just trying to get reports out but didn't understand the full potential. In 2010 we really looked at the way we were doing things and adopted an agile philosophy. Our data warehouse now has grown into an enterprise-level operation in very short order."
Information Leads to Opportunities
Once the business intelligence department became a full-fledged segment of the credit union, it really took off, according to Neumann.
"We took a look at our member data and classified the membership into different categories – by demographics, by products/services used and others," he said. "Once we had that information, we saw opportunities. We saw areas to increase usage – by reaching out to members who did not use their credit card as much as others."
Eventually the unit was able to do full predictive analytics. "We could say if they used one service, there was a good probability they would want A, B or C next," Neumann said. "That allowed us to market with a higher success rate. Instead of marketing with a shotgun approach with a mass mailer, we were able to approach it in a deterministic fashion. We have become very focused and structured. It drives an efficient, lean marketing campaign."
Deborah Mersino, OCCU's chief marketing officer, said at any given moment, she can read a report on one of the credit union's flagship projects and can see, for example, how many checking accounts were opened in the last week as well as drilling down to note how many were opened by millennials or by baby boomers, and/or by residents of a certain county.
"I can better understand the success of our marketing campaigns. By working in concert with business intelligence, it has monetary impact," Mersino said.
Foltz reported the BI group has developed its own profit models it created in house. "We can provide more or less granular data, depending on which part of the credit union is reading the report. Automation has helped make the process better. We have predictive models on product acquisition showing the likelihood someone will want a new product, which leads to more relevant messaging.
The messaging can be done in real time, Mersino added, which allows for many possibilities.
"In terms of membership growth and targeting our segmentation models, we have seen terrific membership growth in all segments," Mersino said. "Our overall growth is strong, but our millennials in particular have shown exciting growth because we have targeted them specifically. All of this is because we are marketing in alliance with business intelligence."
Mersino said another "exciting" use of modeling is the manner in which OCCU is able to market new credit cards. She said cost per new account has dipped dramatically from $300 or even $400 per card to $100 per card.
"In a $10,000 mailing, that is pretty significant savings," she said.
Foltz said internally the key term is "Next Best Product Available." He said the model reviews not only the products people have, but the sequence in which they acquired products and the time in between products.
"It finds what the most probable next product will be. Every night the model retrains itself with fresh data, then makes new predictions," he explained.
There are five models that "compete" with each other, and the BI group uses the best performer each day, Foltz said.
"We know models have to change some times because people change. Each model is based on different algorithms. Most have a basis in mathematics and have different assumptions. We stack them up against each other to determine how often a model was right and how often it was wrong."
'Just Scratching the Surface'
The BI group has six active projects for 2016. The team recently finished mapping out the projects, which will be followed by a refresh of the data warehouse to allow other people to get the data they need.
"We want to make data ubiquitous throughout the organization," said Foltz. "We are training the staff how to use analytics. We have an active, collaborative road map."
Neumann's vision is coming to fruition in how well the business intelligence group works with all departments of the credit union, Mersino said.
"Having great data impacts every area," she added. "It is a gold mine to marketing, and is an incredible partner to us that allows us to do our job so much better."
Mersino said quality business intelligence also works hand-in-hand with the finance department on liquidity models, and has many uses within lending. "We feel very fortunate to have this level of sophistication woven into the organization."
Schiffer said it all starts with taking serving the member "very seriously."
"We realize everything we do supports the organization in supporting the members. That is why we make the effort to reach out to everybody," he said.
According to Schiffer, there is a "strong appetite" in the organization to use business intelligence. "It is our hope we can serve people by getting them to self-serve. It has been very strategic to set up business intelligence and making it better. We want to continue to help everyone move forward."
Foltz noted many of the BI team members have advanced degrees, and the internship program is helping it continue to build. "The interns are not from outside, they come from other areas of the credit union. They bring a new perspective, and then they take expertise back to their areas. Right now one intern is from marketing, one is from lending and one is from operations.
"We want to make the credit union movement faster and more agile by sharing this," Foltz added.
Neumann insists OCCU is "just scratching the surface" on the use of predictive analytics. He said the credit union will continue to make sure it is getting products in front of members who want them, and in a timely fashion.
"We want to be operationally efficient, make our balance sheet stronger, improve earnings and net worth," he said. "There are other, sophisticated tools we can use to even better centralize our data, so we will continue to invest in the technology that enables us to do this good work."
Neumann said business intelligence should be seen as an important aspect of running a CU. "All credit unions have so much data available to them, and it is fascinating to organize that data to help make our membership stronger and more satisfied. We look forward to having the department support the organization."