What your production data isn't telling you
There's a number that almost no manufacturing director can give you off the top of their head: what did unplanned downtime cost my company last year? Not a vague sense of "it was expensive." An actual figure.
Research from Aberdeen and ServiceMax shows that over 80% of manufacturers cannot accurately calculate their true downtime costs. More than 70% don't know when their equipment is due for maintenance. The average plant deals with 800 hours of unplanned downtime per year. That's over 15 hours every single week.
These aren't numbers from massive automotive plants. This pattern shows up in mid-sized companies just as much.
The data is there. The connections aren't.
Most production companies have more information available than they realize. The ERP registers orders, stock levels, and financials. Machine suppliers increasingly deliver data through portals or edge gateways. There's usually a planning system, sometimes a maintenance system, often a quality registration of some kind.
The problem isn't a lack of data. The problem is that all of it sits in separate systems that don't communicate.
A 2026 survey by ECI across 300+ European mid-sized manufacturers confirms this. 84% are still in an early phase of digitalization or connecting systems. 72% are positive about AI. But 47% report no measurable results yet.
There's a significant gap between having data and actually using it.
Industry experts from universities, IIoT platform providers, and system integrators discussed this at a recent industry session. Their conclusion: awareness is growing, but most companies are still in a transitional phase. They know it matters. They just haven't figured out how to make the business case work.
One issue that keeps coming up: most production floors run machines from dozens of different suppliers. Each supplier provides data about their own machine, but not about your production process as a whole. And not all machine data is equally useful. Companies collect and store everything, but use very little of it.
Three things that happen when information lives in islands
You probably recognize at least one of these.
The machine that breaks down every few weeks, and everyone on the floor knows which one it is, but there's no data trail because the failure log lives in someone's head or in a spreadsheet nobody analyzes.
The maintenance that comes too early or too late, because the schedule is based on fixed intervals rather than actual machine condition. An estimated 30% of parts get replaced before they need to be, while other failures get missed entirely because nobody was watching the signals.
The production planner who can't see when the next maintenance window is scheduled, because planning and maintenance live in different systems. So maintenance gets squeezed into a rush order, or gets postponed until the machine fails on its own.
These aren't technology problems. These are information problems.
The knowledge that's walking out the door
There's another factor that makes this urgent. Manufacturing across Europe is losing experienced people at a rapid pace. The generation that knows every sound a machine makes, every quirk of the production line, is retiring. And in many companies, that knowledge has never been captured in any system.
Industry data confirms it: workforce growth in mid-sized manufacturing has stalled. At the same time, willingness to invest in equipment is declining.
The question isn't whether you need to digitalize. The question is whether the knowledge will still be there when you start.
It doesn't have to be big
The examples in the media tend to feature large corporations. BMW running predictive maintenance across four factories. Siemens offering AI platforms that cut maintenance costs by 30%. That feels distant when you run a metalworking shop with 80 people.
But the underlying principle scales down. And it doesn't start with AI or sensors. It starts with connecting information that already exists.
National smart maintenance research labs have developed practical tools for this: simple inventories of what data is already present in your organization, before you invest in anything new. The consistent finding is that most companies are surprised by how much is already there.
Four questions worth asking yourself
These aren't theoretical. They're relevant for any production company. The more you answer with "no," the more likely there's an information gap that's costing you money without anyone noticing.
Can you put a number on what unplanned downtime cost you last year? Do you know which three machines cause the most unplanned stops? Is your machine supplier delivering data that nobody in your organization is currently looking at? Can your production planner see when the next maintenance is scheduled?
Where it starts
It starts with knowing what you have. Not with buying something new.
If you want a rough sense of what unplanned downtime costs in your situation, I've put together a simple calculator based on published research. No registration, no sales pitch. Just fill in your own numbers.
How many of those four questions can you answer with a clear "yes"?
In het kort: de meeste maakbedrijven hebben meer data dan ze gebruiken, maar die zit verspreid over systemen die niet met elkaar praten. Dit artikel beschrijft het patroon, de gevolgen, en vier vragen waarmee je kunt zien of het bij jou speelt. Op imperial-automation.eu/en/tools/downtime-calculator kun je uitrekenen wat ongeplande stilstand je per jaar kost.
Jan Keijzer is founder of Imperial Automation, an AI automation consultancy helping European businesses turn friction into flow. With a PhD in Nuclear Reactor Physics from TU Delft and 30+ years of software development experience, he helps organisations deploy AI effectively.