The Benefits Of A Centralized Analytics Department

Credit union analytics teams touch nearly every part of the organization, from marketing and lending to finance and senior management. In their daily duties, they use specialized tools to generate hundreds of reports to track trends and uncover insights needed to make essential decisions for the cooperative.

Increasingly, they also are delivering predictive analytics for deeper insights and focusing on governance to help safeguard member data. And on the horizon: a host of artificial intelligence and machine learning tools expected to ripple throughout the organization.

With demand for data at an all-time high, credit union analytics teams generally are taking one of two shapes: employees that work together as a centralized, enterprise-level unit or embedded analysts that concentrate on specific areas of the business.

Cole Nuyen, director of analytics at Advia Credit Union ($3.0B, Kalamazoo, MI), leads a nine-member team consisting of data engineers, analysts, and data specialists. The data engineers focus on complex system integrations, data warehouses, and data marts. The analysts collaborate with business units on various projects, whereas the junior-level data specialists handle simpler, low-touch reports.

Nuyen, who joined Advia two years ago to lead the enterprise team, says he has worked under both organizational models — centralized and federated — and both have their advantages and disadvantages.

“It depends on what the rest of the organization looks like,” Nuyen says. “If you have an organization full of people who understand how to use data to make decisions, then a centralized model can work pretty well.”

Lee Brooks, senior vice president of data enterprise analytics at Virginia Credit Union ($5.3B, North Chesterfield, VA) joined the cooperative in 2017 to organize and build an enterprise data team. Today, he leads an 11-member team that handles data engineering, business intelligence, business analytics, and data science.

The team also supports analysts embedded in the marketing and credit risk teams, providing them with data to analyze and run reports on. The marketing analysts specialize in areas such as market segmentation, whereas the credit risk analysts focus on the portfolio. Both jobs require close coordination with their business units.

Brooks says a centralized data team offers many advantages. From a practical standpoint, it provides coordinated support of a large data ecosystem with a small team. Teamwork ensures continuity and also allows for coverage during employee vacation or in-between hires. It also helps ensure consistent data governance processes. Various parts of the organization might create their own spreadsheets, but the enterprise analytics team ensures there’s a single version of the truth.

“Group interactions also are important because we’re always focused on data and always pushing for continuous improvement,” Brooks says. “We as a team learn from one another. If somebody finds a better way to do things, we’ll share it with one another.”

Looking For A Better Way To Manage Your BI? Callahan’s Policy Exchange offers real, vetted org charts and job descriptions to inspire leaders to think differently about their credit unions. Whether you want to explore a centralized team or embed analysts across the business, save time, money, and energy — visit the Policy Exchange first.

The Demand For Analytics

The U.S. Bureau of Labor Statistics predicts the number of data science jobs will grow 36% by 2031, one of the highest areas of job growth. Big tech companies such as Google, Meta, Microsoft, and OpenAI are hiring thousands of data scientists and paying top wages — making it a competitive market.

One way Advia assures it has the right people in the right seats is to build talent from inside the organization. For example, it has data specialist roles designed to be entry-level jobs that can provide support for basic reports.

“If we lose an analyst, we typically have a specialist ready to step up into that role,”
Nuyen says. “And if we lose a specialist, we have someone internally ready to step up. We pull people that have business knowledge into those junior-level roles and then develop them technically.”

Another key element in meeting the demands of the business is automation, Brooks says. Every night, approximately 1,500 data load jobs pull in data from multiple systems and load it into data tables, data marts, and reports.

“For a team our size, the only way we could support 1,500 jobs is to automate it,” Brooks says.

At Virginia Credit Union, which just merged with Member One FCU ($1.6B, Roanoke, VA), all source data, plus nearly all business intelligence dashboards, is captured and sourced from the enterprise team.

“Other areas certainly build reports, but to be displayed on our enterprise portal, dashboards must meet governance standards and be published by our team,” the SVP says.

The credit union’s automation has saved time and created new opportunities. For example, its front-line organization used to create its own spreadsheet with monthly performance metrics on branches and individual associates that was used to support staff performance.

“We said it is probably a better model if you can automate the whole process and give the branches and the associates their data on a daily basis,” Brooks says. “So then they know how they’re performing throughout the month.”

Under the revised process, an automated overnight process populates the data warehouse, which then populates Tableau dashboards that can be viewed over the intranet.

“It’s automated daily, which is great source for coaching staff,” Brooks says.

The ‘Cool Stuff’

Prioritizing projects across various parts of the business is another major focus of enterprise analytics. The analytics team at Virginia Credit Union has its own prioritization model that allows it to immediately fix issues or problems with data. Large projects use a different model.

“For business requests, I host a quarterly data steering committee where I meet with all the executives and share with them all the bigger items that we have to work on,” Brooks says. “I’ll get their input, and we make changes as an organization.”

Although meeting the day-to-day analytics needs of an organization is crucial, teams also should allocate time exploring new possibilities.

“How do we do cool things?” is how Nuyen describes it. “How do we do important things for the business?”

For example, in recent months, Advia’s team has been focused on integrating the credit union with fintech marketplace firms to widen the reach of the credit union. Advia has been growing geographically, but as a low-income-designated financial institution, the analytics team must closely track the impact of growth on the financial demographics of the membership and borrowers.

Virginia Credit Union has been looking for ways to incorporate predictive analytics into its processes. For example, the team created a “next best product” field delivered through its CRM application that predicts the product a member is likely to open next. Another predictive tool for check holds helps tellers determine how much funds should be available to a member when cashing or depositing a check from another institution. A scoring system sets the amount and displays it in the core system.

“It creates a consistent and automated process across all the branches,” Brooks says.

The New Frontier: Data Governance

Although institutions have used frameworks to manage their data practices for decades, government regulations, privacy concerns, and growing cybersecurity threats have pushed data governance to the forefront.

“Every day there’s another breach here and a leak there,” Nuyen says. “Working in a financial institution where you have this level of transparency into people’s lives and their data, it’s critically important to safeguard it and make sure you’re using it responsibly.”

Advia recently stood up a three-member data governance team that is separate from Nuyen’s team. The analytics director says there’s a good reason for that.

“We want to make sure it’s an independent function that’s not inappropriately influenced by anyone,” Nuyen says. “It’s sort of our partner in arms on a lot of different projects.”

Data governance will also be a critical function as enterprises implement emerging generative AI tools and large language models into the workplace and eventually deploy them on public-facing applications. These tools need to ingest huge amounts of data for training, so data quality is crucial.

“With AI taking center stage, it is incredibly important to have good quality data to feed into those models,” Nuyen says. “As with any process, it’s garbage, in garbage out. There’s going to be a ripple effect as it starts proliferating, and now the whole industry starting to realize the inputs are not optimal. Everyone needs better data – and clean data.”

The promise of AI is that it will enable individual users to fulfill their own requests. That day might be coming sooner than later. Power BI already comes with Microsoft Copilot, and Tableau offers Einstein, which means users don’t need to understand how to write a query or use the drag-and-drop fields.

“Already this space is getting where you can take a proficient business user, and if they know the question they’re trying to answer, and they know how to articulate it, the answer is already baked into these platforms,” Nuyen says. “If I want to know the top-producing loan officer in this region last year, it spits it out in a report.”

That level of self-service could answer many basic requests, which would free up time for engineers, analysts, specialists, and more to work on things that push their organizations forward.