85% of industrial companies can't measure productivity. The other 15% grow 13.9% above their peers.
Productivity is the most spoken word in any operations meeting. And the least rigorously measured.
Most industrial companies track production volume, machine utilization and cost per unit. They call that productivity. It isn't.
Productivity is value generated per unit of resource consumed, time, energy, labor, installed capacity. When you measure the right things, decisions change. When you don't, you optimize how the operation looks, not what it delivers.
The metric everyone tracks, that measures nothing
OEE. Overall Equipment Effectiveness. It takes up half the wall of control rooms. It's on every executive dashboard. It's the favorite indicator of anyone who wants to look like they're measuring productivity.
The problem is what OEE hides.
A machine can run at 85% OEE and still produce items no one ordered, in batches that build excess inventory, at a pace that throws the line's flow off balance. The indicator stays green. The operation loses money.
In 2026, LNS Research published a figure that should be in every industrial board meeting: 85% of industrial companies have no real visibility into operational productivity. The 15% that do grow 13.9% above their peers over five years, per LNS Research's Industrial Productivity Index.
That isn't technology. It's measurement method.
What the companies that grow measure differently
The companies at the top of the distribution don't measure more. They measure differently.
They understand productivity as a system: the right output, at the right time, with the minimum necessary resource consumption. And they measure it across multiple layers at once:
- Flow productivity: order throughput time, from intake to delivery, not just time inside the machine.
- Labor productivity: value generated per hour worked, not operator presence on the line.
- Energy and input productivity: real transformation cost per unit, not total cost diluted across output.
- Decision productivity: the time between data becoming available and the corrective action being taken.
That last one tends to be the most revealing. In many industries, the data exists. The decision takes days.
Why the wrong measurement drives the right decisions for the wrong problem
When you measure machine utilization as productivity, your natural response to a drop in results is to raise utilization. You speed up the pace. You approve overtime. You buy more capacity.
And you worsen the bottleneck three stations downstream.
It's the basic principle of the Toyota Production System, still ignored at industrial scale. The factory isn't a collection of machines that need to be busy. It's a flow that needs to be balanced. Measuring each point in isolation produces local optima and terrible global outcomes.
The practical result: companies that invest in automation before fixing the measurement layer automate bad processes, faster and at lower variable cost, but with the same poor systemic result.
McKinsey estimates that up to 70% of digital transformation projects in industry fail to generate sustainable value. One of the most frequent causes: no real productivity baseline before implementation.
What Magellan finds in practice
In Magellan's operational diagnostic projects, the first structured activity is to rebuild the productivity visibility line, what the company can actually see of its own operation.
In more than half of engagements, we find three recurring patterns:
- Outcome indicators (cost, volume, lead time) with no breakdown by the causal vectors, flow, resource, decision.
- Data available in the system that never reaches the person making the operational decision in time.
- Productivity targets set as a percentage change over the prior year, with no reference to what's possible given the installed asset base.
Magellan uses AI to cross-reference production history, downtime, input variation and maintenance data, and pinpoint where the gap between measured productivity and potential productivity is widest. That map directs where the improvement effort will generate real return, not just the appearance of progress.
The most common result of this diagnostic: the company discovers it has between 20% and 35% of unused productive capacity, without buying a single piece of equipment.
The uncomfortable point
Most industrial leaders know their productivity metrics are incomplete. But they keep making investment, automation and headcount decisions based on them.
Not out of incompetence. Out of institutional inertia.
Changing the measurement system means admitting that prior decisions were made with the wrong information. It means confronting managers who have built their indicators for years. It means accepting that the executive dashboard everyone approves doesn't measure what everyone believes it measures.
That's the real cost of the problem. It isn't technological. It's political.
And while the discussion stays at that level, the 15% who measure correctly keep growing 13.9% above their peers, quietly.
Final question: If you removed every utilization and volume indicator from your executive dashboard and replaced them with value generated per resource consumed, what would you discover that you didn't know?