August 24, 2020

7 Lessons Learned From a Credit Union That Took Their Data Analytics to the Next Level

Robert Burger

Topic:   Data

Guest Blogger: Robert Burger, Chief Data Officer, PSECU in Harrisburg.


How PSECU Turned Spreadsheets Into Data Stories

I thought our data and analytics program was pretty sophisticated for a credit union. After all, we’d had all our information in a database for 25 years. But when I saw the state of the art in data and analytics, I realized that we weren’t really using true analytics. All our database did was spit out endless grids of numbers. We had nothing to highlight what was important or to identify anomalies and trends. And yet we spent much more time producing these reports than we needed to.

So, with management’s backing, we spent the past year reworking our approach to gathering, analyzing, and presenting data. It’s already making a big difference. We’ve been able to navigate the stormy waters of the COVID-19 pandemic with a far better view of our financial condition than had we been using the old reports.

Here are seven lessons we’ve learned so far.

1. Paint a picture for management

The No. 1 reason for our investment in analytics was to give the credit union’s leaders a faster and more accurate way to assess the institution’s health and growth. But to be worthwhile, we needed their enthusiastic support. So we presented a lot of information about the benefits of the program, advantages ranging from productivity and efficiency to business planning. What clinched the deal, though, were the pictures—examples of the sorts of visualizations and dashboards we would be able to provide.

2. Get on the same page

We have several business units, each with its own analysts and metrics. For example, if we were working on re-engaging inactive members, everybody had a different definition of a “dormant account” based on the business units’ needs. As we built our new analytics system, we created a single set of numbers and definitions on which we all agreed. Now we can focus on finding solutions rather than debating what the problem is. And we kept everyone involved throughout the project, getting regular feedback on how to make the reports more useful.

3. Pick the right tool

Once we decided that our reporting software wasn’t enough, we had many business intelligence packages to pick from.  Most didn’t handle the whole problem. We wanted software that would build a solid data foundation, with the right governance to create a single source of truth. But we also needed a lot of power to present information with compelling and modern visualizations, and interactive tools for users.

4. Tell a story

Our finance department used to email a spreadsheet each day to our executives, with the latest share balance (deposits) numbers, loans, and investments. If you wanted to see changes over time, you’d root around in email for old reports, and compare. Now we have a daily dashboard with a large chart that visualizes balance trends over time.  You can quickly see if there’s a change we need to address.  The old spreadsheets were mainly used by finance people. The new dashboard is checked regularly by the entire management team.

5. Get your priorities straight

Our monthly scorecard report—key metrics monitored by management—took two weeks to create, every month, and was hard to follow. The ibi software we selected now automates a process that had previously required someone to manually merge spreadsheets from multiple departments. Instead of a monotonous grid of numbers, executives see a visual dashboard designed in cooperation with the marketing department. We selected 12 indicators and arranged them in five pillars—basically columns reflecting overall objectives, such as growth, efficiency, and employee engagementThe latest measures are in large type and color-coded to warn if we are lagging behind plan. The budgeted amount for the month and last year’s result are in smaller type.  By prioritizing and highlighting what we want to communicate, the dashboard makes it easier to absorb a lot of information quickly. 

6. Drill, baby, drill

There’s more to our new analytics than pretty pictures. If an executive sees an anomaly or issue on one of the dashboards, it’s easy to bring up the data behind the number, along with the historical trends. It’s fast, and you get only as much detail as you need.  We’ve also brought in benchmark data from a group of similar credit unions so we can compare our results to the industry.

7. Don’t stop thinking about tomorrow

We’re starting to use our new analytics infrastructure to build predictive models that will help improve our performance. Our first project helps us market to people who join the credit union when they get an auto loan from us through a dealer. The model identifies which of these members will be most receptive to other products—credit cards, checking accounts, etc.— so we can be more efficient in our marketing. We see many other opportunities throughout the business.

We’ve certainly got a lot more to do. My goal is to make sure our executives have the up-to-date information they need to make the best decisions, any time, any place, on any device. Still, we’ve done a lot in just one year with ibi.


Robert Burger earned a bachelors degree in Computer Science from Shippensburg University and has worked at PSECU for 28 years. Robert served briefly as a computer operator, before being promoted to programmer. He wrote PSECU’s first home banking software (a desktop application) in 1995. Robert was programmer manager and chief technology officer, before his current position as chief data officer for the past six years. Throughout his tenure, Robert maintained that data was PSECU’s largest asset and worked to ensure its security and integrity. He looks to make sure PSECU’s data always tells a single truth.