The Care Stack

Why most health system data platforms stall after year one

Victor Wilson · · 4 min read

Year one of a new data platform usually goes well. There is a budget, a vendor, a project plan, and a clear goal: get the data into one place. The team moves the EHR data and the finance data. Old reports get rebuilt. There is a launch.

Then year two arrives, and things get quiet.

The platform still runs and the bills still come. But fewer people log in than you hoped, the request backlog keeps growing, and someone in finance asks what the organization got for the money. Nobody has a crisp answer.

That question is harder to dodge this year. Kaufman Hall reported that the median hospital operating margin closed 2025 at 1.3%. At a margin that thin, every line in the budget has to show what it returns, and the data platform is one of the bigger lines.

I’ve seen this pattern in many health systems. It is rarely a technology problem. The tools work. The problem is that the plan ended at “the data is in the platform,” and that was never the real goal.

Building is the easy part

Moving data is a project. It has a start and a finish, and it fits neatly into a capital request.

Using data is different. It is ongoing work that changes how people make decisions every week. That work does not fit in a project plan, so it often doesn’t get planned at all.

Here is what I usually find at the end of year one:

  • The data is loaded, but no one owns whether it is right. When a number looks off, it bounces between IT and the department for weeks.
  • The new dashboards copy the old reports. Same measures, new colors. Nobody’s job changed.
  • The team that built the platform now also runs it day to day, from tickets to broken loads. They have no time left to build anything new.
  • Leaders were told the platform would pay off once the data was in. It’s in, and the payoff hasn’t shown up.

None of these is fatal on its own. Together they make the platform look like a cost center. Cost centers get cut.

Plan for use from the start

The health systems that keep their momentum treat year one as setup for year two. They decide early what the platform is for, in terms a CFO would recognize.

That starts with a short list of decisions the platform should change, such as which contracts to renegotiate or which denials to fight first. If you can name the decision, you can name the person who makes it, and that person becomes your customer.

Give each decision an owner on the business side. I don’t mean a sponsor who shows up at the steering committee. I mean someone who uses the data every week and will tell you, loudly, when it’s wrong. Data gets better when someone depends on it.

Next, protect time for building. If the same handful of people run operations and also build new things, operations will win every time. Split the work, even if the build side is only one or two people at first. Otherwise year two turns into a help desk.

Finally, measure value in dollars and hours, and start early. A platform that saved a known amount in its first year is easy to fund in its second. A platform that “improved visibility” is not.

There’s a new reason to get this right. AI agents are coming to health systems, and agents need data to read. In March, Epic previewed Agent Factory, a no-code tool for health systems to build and monitor their own agents. Other vendors are moving the same way.

An agent built on your platform is one more kind of use. It can draft a denial appeal or flag an invoice that doesn’t match the contract. But it inherits every problem on the list above. If no one owns whether the data is right, the agent will be wrong with confidence, and much faster than a person would be.

So the same rules apply. Tie each agent to a decision. Give that decision an owner who checks the agent’s work. The agent prepares, and a person decides.

What to do this quarter

If you are near the end of year one, or already past it, you don’t need a new strategy. You need a short, honest review.

Get the people who use the platform and the people who run it in the same room. Ask a few plain questions:

  1. Which decisions are made differently today because of this platform?
  2. Which numbers do people still pull by hand, and why?
  3. Who gets the call when a number is wrong, and how long does a fix take?
  4. What has the platform saved or earned so far that we can actually point to?

The answers will be uncomfortable in places. That’s fine. They will also show you where to focus.

Pick one decision from that conversation and make it the only focus for the next 90 days. Give it a business owner and a dollar target. Put a small team on it that does nothing else. When it lands, tell the story in plain numbers to the people who approve the budget, then pick the next one.

This is slower than a big relaunch, and it is less exciting. It also works. Each win earns trust, and trust is what keeps a platform funded long enough to matter.

Before you approve next year’s platform budget, ask for one sentence in writing: which decision will this platform change next year, and who makes it? If nobody can answer, fix that before you spend another dollar on the technology.

Discussion

    Join the discussion