Industrial robots: ROI depends on the process, not the robot
Why most mistakes in manufacturing automation happen before the robot is purchased - and how a manager can evaluate real returns.
Interest in industrial robots among manufacturing companies has grown noticeably over the past few years. Robot costs are falling, available solutions are multiplying, and success stories appear regularly in industry publications. Managers are asking "could we do this too?"
I have nothing against that question. Robotic automation works - this is not hype. But the ROI of a project is not determined by the robot's specifications. It is determined by what happens around it.
Why standard payback calculations do not work
The standard logic goes: take the cost of manual labour on the operation, divide by the cost of the robot, get the payback period. A simple, persuasive calculation.
The problem is that it only counts what is easy to count. Left outside the calculation:
- the cost of integrating into the existing production line,
- changes to the logic of adjacent operations,
- retraining staff and changing roles,
- maintenance and spare parts,
- productivity losses during the implementation period,
- what happens to the robot when the product line changes.
That last point is often the most important. Flexibility has a price. A robot optimised for one specific operation becomes a problem, not an asset, when the product changes.
Where problems actually arise
I have seen several failed robotics projects - not in the sense of technical breakdowns, but in the sense that the expected returns never materialised. The pattern is almost always the same.
They automated a process that was itself poorly designed. The robot faithfully reproduced the inefficient operation - only now it could not be quickly changed, because change required reprogramming and possibly retooling.
They underestimated the variability of the incoming flow. In a demo, the robot works with perfect parts under perfect conditions. In a real shop floor there are dimensional variations, contamination, non-standard feeds. For a person this is a minor adaptation. For a robot it is an error or a stoppage.
They did not think through the failure scenario. What happens to the production line when the robot stops? If it was a bottleneck - the line stops too. A plan is needed: a manual workaround, a backup operation, the maintenance response time.
How to evaluate a project properly
Before moving to equipment selection, I recommend answering a few questions.
How stable is the process to be automated? If the operation changes every six months, the adaptation cost may consume all the savings. Robotics works well on stable, repeatable operations.
What happens to the operation when there are deviations? How often do non-standard situations occur, and how are they handled today? You need to understand what will fall within the system's responsibility and what will remain with a person.
What is the true implementation cost? This should include integration, retraining, changes to adjacent operations, and support through the first year or two.
What happens when the product line changes in two or three years? Flexibility has a price - is it included in the calculation?
A practical conclusion
Industrial robotics is an investment with a long horizon and significant adjacent costs. Companies that approach it as equipment procurement are often disappointed. Companies that approach it as process reengineering with automation as a tool get durable results.
Before choosing a robot, describe the process to be automated as precisely as you would describe it to a new employee. If that description does not come out clear and unambiguous - sorting out the process is the first step.