MeridianLink is a leading provider of cloud-based software solutions that empower banks, credit unions, mortgage lenders, and consumer reporting agencies to accelerate digital transformation. The company serves over 2,000 financial institutions across the U.S., helping them streamline workflows, improve lending efficiency, and make smarter, data-driven decisions.

In the first part of this interview, MeridianLink’s VP of Data, Chris Eldredge, led by Afidence’s Business Development Manager, Spencer Hogan, explored how financial institutions can overcome data silos, foster thriving data communities, and standardize definitions.

In this part, Eldredge dives deeper into sustaining data communities, identifying critical KPIs, and tackling the challenges of siloed systems.

Why siloed thinking creates exceptions everywhere

Eldredge states that the rapid evolution of businesses demands having a team tasked to keep everyone aligned with the updates. He says, “Business moves so quickly today that you might have a meeting this month and get everybody aligned, but if you don’t have a team to keep those meetings going, everybody forgets.”

Further, Eldredge states that sustaining a data community requires dedicated resources — not just data engineers. “Your data engineering team is there to provide content, but it’s not their job to go and convince you that what they’ve just built is the one that everybody in the business should use. You need people whose job it is to focus on the context.”

On top of it, Eldredge reiterates that people management is as important as data management. He explains that anyone in the business who engages with data has a role in keeping it aligned across the organization.

To achieve this, leaders must provide practical ways for employees to stay connected. He suggests building ways so that anyone with an interest in data can ensure that “data is aligned across the business and make sure that they have ways that they can connect.”

Listing some examples, Eldredge states that it could be a regular cadence of meetings wherein anybody wanting to be in that community can get involved. It can also be a repository that the teams can access for content and best practices, or a forum or chat room to get people access to information.

Identifying what matters the most

When asked how organizations can identify the most critical data and KPIs, Eldredge points to a finance-first approach. “What you report to the street is what most people have to work with. Whether that’s revenue, products, or counting customers. Those are the things you can’t go and have a different story about another day.”

He suggests starting here and then understanding the differences that exist across businesses in terms of customers, products, or employees. The next step is to get people together and discuss what needs to be differentiated in the data model.

Eldredge then highlights the role of certified data models in cutting inefficiencies: “Business analysts and data scientists spend up to 90% of their time on prepping data. If you can create a certified data model that gives them a predefined set of data, you can speed up the pace at which your data professionals can work.”

Lessons on misaligned KPIs

Moving forward, Eldredge shares a cautionary note from his career where an organization struggled with multiple definitions of recurring revenue. “The multiple definitions were not only across functions like product, sales, and finance, but also within those departments. Why? Because there was no established definition.”

When the team delved deep to understand the actual difference, it found inconsistencies, gray areas, and cross-functional teams naturally favoring definitions that best suited their goals.

Resolving this required aligning product and customer hierarchies, establishing clear start-and-stop rules for customer counting, and enforcing consistent definitions companywide.

Another challenge Eldredge often sees is organizations discarding data when consensus breaks down. “You might get into an executive meeting where one camp has one set of numbers and the other camp has another, and then those numbers become discounted. People say, ‘If we can’t agree, maybe we should make this decision another way.’ That’s how organizations become not data-driven and go with opinions or gut feeling.”

He warns that without alignment, data becomes an excuse for avoiding accountability rather than a tool for decision-making.

Breaking down siloed systems

One of Eldredge’s key mandates at MeridianLink is addressing siloed IT systems: “How do we make data work across our systems as opposed to making the data in our systems work?”

Discussing further, Eldredge draws a sharp line between focusing on individual platforms versus creating business-first data models. “If you start with individual systems, and most companies, MeridianLink included, have grown up with a variety of systems across a variety of time periods that kind of have persisted… And what has happened is you didn’t have the full picture when you were working on those individual systems, especially if you have mergers and acquisitions.”

Because organizations rarely control how legacy systems evolve, those gaps in visibility often lead to decisions that no longer serve future business goals, he adds.

To overcome this, Eldredge recommends flipping the usual approach. “You can change that dynamic to say ‘we don’t want data models that are specific to any given system; we want data models that work for our business that then we can apply to the systems.’”

He cautions against relying solely on vendor “best practices.” Eldredge notes, “If you don’t have an opinion of how your business needs to operate and the way that it should always work for your business, you’re going to get locked into something that might not be future-proof.”

To avoid that trap, organizations must go beyond data pipelines and enforce rigor through clear documentation and governance. “It’s making sure that you have process maps, and understand what your business processes are and what adds value in your business. It’s making sure that you have data dictionaries and glossaries so that you know what your data looks like and what it means, and you have defined metrics so you know how you’re measuring these things.”

Looking back, Eldredge contrasts the old system-first mindset with today’s opportunities. Unlike the past, today, there are numerous data platforms and data tools that can be leveraged to define data models.

In conclusion, Eldredge underscores that this isn’t a skill just any technologist can pick up. “Understanding how to architect and map out a data model that’s relevant for a company is a very specialized and important skill, and it’s something that companies should seek out.”

CDO Magazine appreciates Chris Eldredge for sharing his insights with our global community.