There is no general answer to whether factory automation is worth it, and any page that gives you one is selling something. What there is instead is a calculation, and the hard part is not the arithmetic. It is being honest about the inputs.
This is a method for building that case: how to establish what the current process really costs, how to price a project rather than a machine, the three traps that make payback figures wrong, and where the numbers stop being the deciding factor.
Ask a narrower question
“Is automation worth it” cannot be answered. “Is automating this operation, on this line, at this volume, worth it compared with continuing as we are” can be, and the second question is the one a board will actually approve or reject.
That reframing matters because it puts the comparison in the right place. The alternative to automating is not doing nothing at zero cost. It is continuing to run the current process, with everything that currently costs, for the next several years.
Baseline the cost of carrying on
What the task costs to run
Start with hours on the specific task rather than headcount. How many hours a week go into this operation, across all shifts, including cover for absence, overtime at peak and agency staff when volume spikes. Fully loaded, not basic rate. Most operations are surprised by this number because it is spread across people who each do several jobs.
What it costs when it goes wrong
Scrap, rework, mis-picks that reach a customer, credit notes, short deliveries, the time spent investigating. These are usually recorded somewhere already, in quality reports or customer complaints, and rarely totalled. A process with a small error rate and a high value per unit can cost more in errors than in labour.
What it costs in capacity you cannot sell
This is often the largest figure and the one most often left out. Orders declined or delayed, overtime run to catch up, weekend shifts, and growth that has not been pursued because the operation could not absorb it. If a constraint is holding back revenue, the value of removing it belongs in the case.
What it costs in risk
Repetitive lifting, manual handling at height and pace, people working close to moving equipment. The cost sits in absence, staff turnover, insurance and the disruption when something happens. It is harder to quantify than the others and it should still be written down rather than left as an unstated assumption.
Price the project, not the machine
The quoted price of the equipment is one line of several, and on most projects it is well short of the total. A complete figure includes integration and controls, safety assessment and guarding, any floor or services work, the conveyor and buffering that connects the new equipment to what already exists, spares, training, and commissioning.
Two costs are almost always missed. The first is disruption during installation: what does the operation do while the area is out of use, and what does that cost. The second is internal time, which is your own engineering and operations people spending months on the project rather than on their other work.
A case built on the machine price alone will look excellent and then fail to land, which damages the credibility of the next proposal as well as this one.
The three traps in a payback figure
Sizing on average volume
Equipment gets specified for peak but justified on average, and the two produce very different numbers. Use the volume the system will actually see, and be clear about which figure is doing the work in the calculation.
Counting labour that will not actually leave
A cell that removes four hours of work a day from an operator who is present for eight hours has not saved a person. It has freed four hours. That is real value, but only if there is something worthwhile for those hours to go into, and the case should say what. Savings claimed against people who remain in post, doing the same shifts, tend not to appear in the accounts.
Ignoring the ramp
Very little runs at specification on day one. There is a period of tuning, exception handling and learning, and during it output is lower than both the old process and the new target. Building in a realistic ramp makes the case more credible, not less, and it prevents the disappointment that follows a payback date calculated from commissioning day.
Where Irish tax treatment fits
Capital allowances change the shape of the numbers and are frequently left out of a first-pass case. In Ireland, plant and machinery is generally written off for tax at 12.5% per year on a straight-line basis, spreading relief across eight years.
Separately, the Accelerated Capital Allowance scheme allows a 100% first-year write-off on equipment listed on the SEAI Triple E Register, currently available until 31 December 2030. Equipment that operates directly on fossil fuels is excluded. Whether anything in a particular project qualifies depends on the specific products and the register, so it is a question for the register itself and for your accountant rather than for a supplier.
None of the above is tax advice, and rules change. The point is simply that the after-tax profile of a capital purchase can differ enough from the headline figure to be worth checking before a case is presented.
When the numbers say no and the answer is still yes
Some projects go ahead regardless of payback, and that is legitimate as long as it is stated openly. A safety issue that has to be resolved is a cost of operating, not an investment decision. A capacity constraint blocking a growth plan is a strategic decision. A customer requirement, a process nobody can be recruited for, or work that cannot be staffed reliably at any wage all fall outside a simple return calculation.
The reverse case is worth as much attention. Numbers can look excellent for a project that should not proceed: a product about to change, a line due to be reconfigured, a volume that exists this year because of one contract, or a project with no internal owner. A strong payback figure does not make any of those safe.
A quicker sanity check first
Before building a detailed case, three questions filter out most projects that will not survive one.
Is the task repetitive and well defined, or does it need judgement that has never been written down? Will the volume still be there in three years, or is it tied to something temporary? And is this operation actually the constraint, or is it simply the most visible?
A no to any of the three does not automatically end the discussion, but it does mean the detailed case is likely to expose a problem, and it is cheaper to find that out now.
Common questions
What payback period should we be aiming for?
Whatever your own capital policy requires, and it varies widely between businesses. The useful discipline is to agree the threshold before the case is built rather than after the number appears, because a target set afterwards tends to be set around the answer.
Should we automate the worst process first?
Usually not. The first project sets expectations for everything that follows, so the better candidate is the one that is best understood, most stable and easiest to prove. The worst process is often worst because it is chaotic, and chaos is the hardest thing to automate.
Does it only pay if we reduce headcount?
No, and framing it that way narrows the case unnecessarily. Most of the return usually sits in capacity, consistency and error reduction, with labour redeployed rather than removed. The case is stronger when it says specifically where those hours go.
How do we cost this without a full quotation?
Budget-level pricing is enough for a first pass, and a supplier can usually provide it from a description of the operation. Where the risk sits in whether something will work at all rather than what it costs, a proof of concept settles it on your own product, and simulation settles throughput and layout questions before capital is committed.
Where to start
Most manufacturers begin with one well-defined operation rather than a plant-wide programme. Palletising and depalletising are common first projects because the task is repetitive and the benefit is easy to measure. Machine tending suits shops where machine utilisation is the constraint. In distribution, picking, sortation and automated storage are the usual candidates.
Whichever it is, the sequencing matters as much as the choice, which is covered in our guides to factory automation and warehouse automation.
LVP Automation works with manufacturers and logistics operations across Ireland to scope projects and put realistic numbers against them, as part of our automation services. Tell us what the operation looks like now, and we will help you work out whether the case stands up.
