Many machine manufacturers have been selling according to the same pattern for decades: machine out, service in, spare parts later. This works – until customers start to pay more attention to capacity utilisation, skills shortages and predictable costs. Then it’s no longer “the machine is good” that counts, but: How fast is production? How stable? With how little effort?”
TRUMPF: Pay-per-part as the answer to bottlenecks
TRUMPF describes its “pay per part” model as an approach in which TRUMPF helps to control the operation, programming and maintenance of the production cell and detects malfunctions at an early stage because the system is monitored remotely.
The logic behind this is crucial: the provider does not earn money from the sale – but from the customer’s results.
This is more than just a pricing model. It is knowledge as a competitive advantage:
-
Knowledge about failure patterns.
-
Knowledge about real utilisation.
-
Knowledge about which parameters drive part costs.
Why this is so valuable for SMEs
Many medium-sized customers have similar problems:
-
Machines do not run continuously.
-
Set-up times are too long.
-
There is a lack of personnel for programming and optimisation.
-
Downtime is expensive but poorly documented.
If a provider makes these points visible and solves them based on data, the benefits are very clear:
More output per hour, less stress, better calculations.
What “in-depth knowledge” means in concrete terms
For such models, you need three layers of knowledge:
-
Status knowledge: What is the machine actually doing right now?
-
Cause knowledge: Why is it standing still? Material, programme, operation, maintenance?
-
Economic knowledge: What does an hour of downtime cost – and what are the benefits of a measure?
AI is particularly helpful for layers 1 and 2: patterns in sensor data, anomalies, forecasts.
People are crucial for layer 3: prioritisation, economic efficiency, responsibility.
The transfer for medium-sized mechanical engineering companies
You don’t have to go straight to “pay per part”. But you can use the principle:
-
Link service packages to availability.
-
Plan maintenance proactively.
-
Provide customers with clear OEE/downtime logic.
-
Define pilots: 1 line, 1 customer, 90 days.
What SMEs can learn from this
-
Good products are no longer enough. Win good results.
-
Data is not an IT project, but a building block of the business model.
-
AI is most effective when it targets real bottlenecks (downtime, set-up time, waste).
-
Start small but measurable: 90 days, clear target values.
