When the Machine Builder Says “It Can’t Be Done”
There’s a phrase we’ve heard more times than we can count in this industry:
“It can’t be done.”
Usually from a machine builder. Usually about data collection.
And usually… completely wrong.
The Setup
A global medical device manufacturer came to us through a partner. High-speed production lines making life-changing wearable devices—the kind of products that genuinely improve people’s lives every single day.
The problem? Their OEE was all over the map. Some days 30%. Some days 70%. Average somewhere in the 40-60% range. On high-speed lines running at parts-per-second rates, that unpredictability was costing them a fortune.
They’d been told by the machine builder that collecting meaningful data from these lines wasn’t possible. The equipment was too fast. The network was too full. The architecture couldn’t handle it.
So they accepted it. Because when the people who built your machines tell you something can’t be done, you tend to believe them.
The Conversation That Changed Everything
Our OT lead, Mike Manzi, walked into a workshop and looked at the lines. He asked a simple question:
“You’re running high-speed equipment with multiple operating cells, constant downtime, and you’re not collecting data off this? How do you expect to know why it’s breaking down?”
The answer was basically: we speculate.
Mike’s response: “I’ve deployed architectures collecting data at 20 milliseconds across hundreds of data points. This can absolutely be done. You just have to architect it correctly.”
What followed was a masterclass in pushing back—not against the customer, but against the limiting belief that had held them hostage. Every objection the machine builder had raised, Mike had an answer for. L80 processor limitations? Here’s the workaround. Network congestion? Here’s the architecture. Communication port bottlenecks? Pull from the backplane.
The customer had never had anyone in their corner who could go toe-to-toe with the OEM on technical depth. They’d been told “no” so many times they stopped asking.
What We Actually Built
The architecture isn’t magic. It’s just… done right:
- Dedicated ENT cards pulling high-speed data off the backplane instead of through communication ports
- Redundant interfaces handling approximately 40,000 transactions per minute each
- Local data archivers feeding a central enterprise historian
- Asset frameworks at both local and enterprise levels
The deployment happened in September 2024. By April 2025, the data analysts had reports ready for the floor.
The Results (In Three Months)
10% throughput improvement.
7% yield improvement.
On high-speed lines running parts per second, across multiple production cells.
Do the math on what that means in real dollars. Hint: it’s tens of millions annually.
And this was just from base reporting. What alarms are actually firing? What’s causing shutdowns? What’s our real yield right now? Nothing fancy—just giving operators the visibility to act in real time instead of chasing their tails.
What Made It Work
Strong executive sponsorship. The divisional VPs in both Americas and Europe were behind this. When things got stuck—and they always get stuck—we could pull that lever.
Proving it works, fast. Nothing kills skepticism like results. Once the numbers showed up, the conversation shifted from “can this be done?” to “how fast can we do this everywhere?”
Respecting the complexity without being paralyzed by it. This is a validated pharmaceutical environment. There’s paperwork. There are change control processes. Some sites have been operating for years and move fast. Others are newer and scared to tie their shoes. You have to meet people where they are.
The Uncomfortable Truth
The machine builder wasn’t lying when they said it couldn’t be done. They just didn’t know how to do it. And there’s a difference.
Most manufacturers accept limitations that aren’t actually limitations. They’re just the boundaries of what their current partners know how to solve.
Your controls engineers and automation folks? They probably already know where the problems are. They’re just too busy fighting today’s fires to win the war. And they’ve been told “no” enough times that they’ve stopped pushing.
Trust your team. Get them the support and air cover they need. And find partners who can push back on the “it can’t be done” crowd with actual technical depth—not just slideware.
Where It’s Going
What started as a proof of concept on a couple lines is now a multi-year, multi-site engagement. Phase two is digital twins, predictive models, and root cause analysis that goes beyond reactive firefighting.
The 10% improvement? That’s just the floor. We think 30% is achievable by the time this journey is done.
But more importantly—the people running these lines now have visibility they never had before. They can see what’s happening, understand why, and act on it. That changes everything.
Sometimes the biggest unlock isn’t new technology. It’s someone willing to say “yes, it can be done” and then prove it.
Axiom Manufacturing Systems helps manufacturers stop accepting limitations and start solving real problems. If your machine builder is telling you something can’t be done, let’s talk.