The Key to a Successful Data Products Initiative? Effortless Maintenance

By Donna Bridge

Published on May 15, 2025

You've invested time, money, and people into launching a new data products initiative. You've picked the right platform, built out your first few products, and maybe even rolled them out. But what happens after the initial shine wears off?

The key to long-term success isn't just launching data products — it's maintaining them.

Why data maintenance breaks down

As the Head of Product Research at Alation for the past four years, I've talked to hundreds of data producers, consumers, and governance leads across industries. I've seen a lot of successful data governance programs, and I've seen a lot of failures.

One thing is consistent across all data programs — maintaining relevant, fresh knowledge content is hard. It is so hard that I've seen countless organizations stumble after the initial excitement fades. Not because of a lack of effort — governance leads and data stewards work incredibly hard — but because they lack centralized power, business buy-in, and sustainable maintenance processes. The result? Their governance programs get deprioritized and defunded.

At Alation, we've developed strong onboarding programs to help customers launch successfully. Our 3-month expert services Right Start program trains admins and stewards on how to set up and manage a catalog. We host "Docujams" to bring business users together to curate their assets in a group session filled with drinks, snacks, and prizes. These initiatives work — they get customers off the ground. But they're not sustainable.

Things fall apart later in the day-to-day: keeping data fresh, relevant, and usable. That's the real challenge.

Starting is easy, sustaining is hard

Maintaining knowledge requires ongoing, often thankless effort. You can't just create documentation or define a data product and be done with it. The second you publish, the clock starts ticking. Things change. Systems update. Business priorities shift.

And yet, most data management tools aren't built to support the daily work of keeping knowledge fresh and accessible. Traditional platforms are powerful and flexible, but they often prioritize technical capabilities over usability, creating a disconnect between the people who need to use the tools and the tools themselves. What's missing is a user experience designed for the people who hold critical institutional knowledge — the ones who need simple, reliable ways to share that knowledge with others.

It's like buying a high-performance sports car without understanding maintenance requirements. Sure, it's powerful — but without regular oil changes, tire rotations, and system checks, that expensive vehicle will quickly lose value and functionality. Similarly, without proper guidance and support, even the most sophisticated data environments deteriorate rapidly.

As a result, the burden of keeping data ecosystems clean, organized, and up-to-date falls heavily on data stewards and especially on the data program admins. In organizations with strong data governance programs, this burden is shouldered by large, dedicated teams working day in and day out to maintain the system. But not every organization has that luxury — nor should they need it.

Usable, trustworthy data shouldn't require a small army. The tools themselves should carry more of the weight. That's the future we're aiming for — where AI helps do the heavy lifting and makes sustainable data management possible for every organization — big or small, with mature or immature data programs.

The myth of non-invasive governance

One of the biggest pitfalls in data management I've seen is the idea of "non-invasive data governance." The concept sounds great on paper: it's "non-invasive" to the business because it doesn't require creating new roles or dedicating full-time resources to data stewardship. Instead, the idea is that everyone who touches data simply does their part. If every data user takes responsibility for curating and maintaining documentation for their own assets, the collective effort keeps everything up-to-date.

The theory is appealing. It suggests that governance can be lightweight, embedded, and shared — avoiding bureaucracy and spreading the responsibility across the organization. In reality, though, it almost always fails.

Why? Because documentation and stewardship are not part of most people's core job responsibilities. Business users aren't incentivized to do it. It's not in their OKRs, they don't have time, and — most importantly — it's a thankless and never-ending task. Adding documentation duties to already overburdened stakeholders doesn't make the work more efficient — it just shifts the load onto people who are being measured against completely different outcomes. Instead of doing the work, they avoid it, because doing it invites questions, follow-ups, and distractions from their real priorities.

There are three fundamental issues. One is accountability. When "everyone is responsible," no one is truly accountable. Without clear ownership and specific metrics, these distributed efforts inevitably collapse under the weight of competing priorities. 

The second issue is that curation efforts are not easily tied to business value. Sure, having usable data that everyone in the organization can discover is nice, but how does it directly impact the bottom line? Measuring the value of data governance programs is notoriously difficult, which makes it hard to get business stakeholders on board for curation duties. 

The third barrier to success is the ongoing effort required for a largely thankless job. We ask a lot of our stewards—to keep institutional knowledge fresh and accurate—but maintaining the standards needed to make data usable and valuable is an unsustainable burden.

Non-invasive data governance might work in very different conditions. If stewardship were easy, clearly embedded in people’s core responsibilities, intuitive, and genuinely rewarding, more people might be willing to take it on. But that’s not the reality today.

The case for data products

How can organizations address the accountability issue? It may be time to formally build data stewardship responsibilities into people's roles. Data products offer a promising path for this approach.

Unlike non-invasive governance, the data product model is explicit about ownership. It assigns responsibility for a defined set of high-value assets to a specific individual — the Data Product Owner. Maintenance isn't a side task; it's core to the role. It's right there in the title.

So, step one in making this model work: establish accountability by adopting a true data products structure.

But that's a major commitment—and not every organization is ready for it. Many I've spoken with are excited about launching data products but are hesitant to do a full reorg or create entirely new roles. It's a big investment. And yes, there's risk: What if data products don't stick?

Still, most of these organizations are moving forward with data products. So the question becomes: how will they make them sustainable?

The value-driven opportunity

This is where data products have a real advantage, because they promise clear, tangible business value. They're built for specific use cases, with defined consumers and measurable outcomes. This focus gives them a better shot at long-term stickiness.

That business value is what makes data products so compelling to the broader organization. When the benefit is obvious and immediate, the incentive to adopt — and maintain — becomes much stronger.

But here's the catch: you need both strong business value and an experience that is relatively effortless and inherently rewarding to the individual contributors. If maintenance is too hard, time-consuming, or disconnected from people's daily workflows, adoption will stall and products will decay. 

Moving from theory to practice

The challenges of data maintenance are real, but they're not insurmountable. In our next blog post, "4 Ways AI Can Transform Data Product Maintenance," we'll dive into practical AI applications that will make ongoing maintenance not just possible, but efficient. We'll explore how to incentivize with business value, leverage social proof, reduce friction with great UX, and let AI handle complexity. Because the future of successful data products isn't about working harder—it's about working smarter with the right tools and approaches.

Stay tuned to learn how you can transform data product maintenance from a burden into a sustainable, value-generating practice.

Learn more about data products:

    Contents
  • Why data maintenance breaks down
  • Starting is easy, sustaining is hard
  • The myth of non-invasive governance
  • The case for data products
  • The value-driven opportunity
  • Moving from theory to practice
Tagged with

Loading...