The budget for production automation: how to set it without knowing the price of the solution

How to set an automation budget – before you start collecting quotes

Automating a production process should rest on numbers, and the budget should follow from the cost you already carry or will carry in the future (in the case of a new project or hiring another person). In practice, when investment comes up, what I usually meet is a situation where an idea has appeared, or at least a feeling that something needs to change and improve, but there is no information on which to base that decision consciously — which usually reduces the whole discussion to one question: how much will such a station cost? Only once some quote appears do the first conversations and analyses begin. The moment for a decision arrives, questions are asked whether we can spend that much and whether it will pay off — and I often wonder what goes into the calculation at that stage and on what data the decision is based. This approach creates a series of misunderstandings between supplier and client, stretches out the time needed to prepare a quote, makes it harder to match the optimal solution to the specific situation of a given plant, and often takes up the time of both sides unnecessarily. That is why in this article I go through the whole process comprehensively and try to justify why the approach needs to change.

In what follows I will try to explain why I believe the most commonly used return on investment indicator (ROI) is used wrongly — too early in the buying process — and what happened in Poland between 2020 and 2025 that so drastically changed the situation of manufacturing companies.

The pay calculations you will find in this article are based on reports and official publications from GUS; the sick leave data comes from a ZUS report. The wage trend was derived from data on average pay in Poland between 2005 and 2025.

Before I start, a few terms need explaining, because most of the differences come precisely from both sides counting different things. These are indicators I have selected; they do not exhaust the subject – many more are in use – but I take the ones that have the greatest influence on the calculations.

To answer the question of what budget we have, we will need:

1. The definitions from the legend above, which form the basis for the calculations further on
2. A time horizon
3. Data on production in your plant:
   • 1/2/3-shift operation
   • current throughput, how many people work and for how long to cover current and future demand

Why ROI cannot be the starting point of a conversation about automation and what the COI indicator is that defines your situation today

The most common mistake I see in conversations about investing in automation is using the return on investment indicator, ROI, which requires knowing the price of the solution — and that is the last thing you learn in the buying process.

The formula used here looks as follows:

Example:

As you can see, calculating ROI requires the cost of the investment to be known – you can only work this indicator out once you have a quote, which is why most companies push for an offer from the start: how much will it cost? That reduces the entire conversation to price. The lower the price, the higher the ROI – is that really so? I check this on concrete examples in section 5.

Why does this matter so much? Production operators are paid differently, companies are at different stages of development and in different financial situations, and boards disagree on what the payback period should be. Let us look at an example:

Example 1: A station palletising cartons of finished product

Automating exactly the same process, two different companies, and a budget difference of 150 thousand against 1.7 million. What is a supplier supposed to quote without knowing these facts? The priorities will be completely different, and so will the expectations.

I will come to the figures in the table above in the next section, where I calculate them on a concrete model.

In the first case automation does not make sense yet; some simple, inexpensive improvements to packing are probably worth introducing. It is worth returning to the analysis once a full second shift has to be started and a good financial position encourages the investor to extend the required payback period to 2 years.

In the second case the cheapest offer, based on lower-quality components, a shorter warranty, lower energy efficiency or lower throughput, will ultimately deliver a lower ROI. The question is whether the investor took that into account at the moment of deciding.

Compare it with the decision to buy a car – what do we usually take into account:

• the purchase price (CAPEX)
• average fuel consumption (OPEX)
• reliability, the cost of spare and wear parts (OPEX)

and much more.

If I have a budget of 50 thousand PLN and cover an average of 5 thousand km a year, I am looking for a different car than if I had 150 thousand PLN and covered 25 thousand km a year.

For that reason I believe a conversation about investment should start with an analysis of the situation in your plant and be based on concrete figures rather than guesses – before the first contact with a supplier. If you do not know how to go about it but would like to check whether such an investment could make sense for you, get in touch with us. In a short, free phone call we will establish whether there is real potential for automation, or whether it is better to wait or look at a different process.

Coming back to setting the budget – the COI indicator comes to our aid, and the conversation about automation should start with it. COI answers the question of how much this process costs me today with no investment at all, and how much leaving it unchanged will cost me over the coming years. For any investor, business owner or decision-maker on a corporate board that information is crucial, and it allows a prospective supplier to select a solution matched to the expectations, the time horizon and the financial means.

What a production operator really costs in 2026

One production worker, assuming 1-shift operation, costs an employer in Poland 121 986 PLN a year, that is 10 165 PLN a month — calculated on the basis of median gross pay of 7 690.82 PLN GUS, Feb 2026.

Here, however, the first discrepancy appears – the cost of an operator is not the same as the cost of a staffed station – and at this point I return to calculating the first indicator from the table of terms:

Station staffing factor (SSF)

One full-time post does not fill the nominal working time of one station, and this is an aspect that is often overlooked.

I derive the SSF from actual data. Nominal working time in 2026 is 251 working days (excluding public holidays and weekends) of 8 hours. I apply the following deductions:

• holiday 26 days
• sick leave: 12 days, reduced against the average in ZUS data (14.65 days)
• training/medical check-ups 2 days
• other absences (compassionate and carer's leave) 2 days

In the calculation for 1-shift operation I assume sick leave of 12 days on average. In companies of more than 1000 people and on 3-shift operation these values are correspondingly higher, and that is accounted for by applying the additional SCF factor.

Shift cost factor SCF (Shift Cost Factor)

This is my own multiplier, allowing a more precise calculation that accounts for the additional costs of running a second and third shift. The reference point is concrete data:

Looking at the size of the entity, the highest median pay was recorded, as in previous periods, in entities employing 1000 people and more — 9 146.18 PLN (118.9% of the overall figure), and the lowest in entities employing 9 people or fewer — at the minimum wage level, that is 4 806.00 PLN (62.5% of the overall figure)

It is in companies of more than 1000 people that we most often find 3-shift operation and higher total pay, which follows from statutory night-work and shift allowances. On top of that, higher turnover and the associated recruitment costs, as well as longer sickness absence, have to be taken into account.

So I assume the following:

Total cost to the employer for one operator

The cost of staffing a station

COI — what it costs to keep the process unchanged

For a station running three shifts the cost of inaction is 1.07 million PLN over 24 months, 1.65 million over 36 and 2.89 million over 60 months — with a conservative indexation of 5% a year. You carry this cost whether or not you decide to automate.

Why 5% is conservative I explain in section 4. Average pay in the national economy grew by 11.5% a year between 2020 and 2025 GUS.

How to calculate the total cost of the machine that replaces an operator

The subject has to be approached honestly and stated plainly: an automated station also generates additional costs. OPEX has to be subtracted from COI to arrive at the budget — in practice, working out the fixed costs after automation means accounting for:

• Energy: for example 3 kW of average draw (compressed air compressor supply + robot supply) × 2 008 h × the number of shifts × 1.2 PLN/kWh ASSUMPTION — from 7.2 thousand PLN on one shift to 21.7 thousand on three.
• Service, spare parts and wear parts: 3% of the outlay per year ASSUMPTION.
• Residual operator attendance: it depends on the autonomy achieved and on how much involvement is needed; for example, in one of our projects the operator comes to the station after automation usually for 10 minutes once every 4 hours — they top up components and collect the finished product. That is time which has to be subtracted in the calculations.

CAPEX + OPEX over the entire period of use is TCO — and it is TCO, not the price in the quote, that should be compared with COI. Breakdowns and downtime are the hardest item to estimate in this account. Some machines, particularly the most complex ones, do not reach full throughput in the first weeks of operation, often because setting them up optimally requires steady series production, properly trained operators and dimensional stability of the components. That is why in practice, where the technology and the budget allow, stations are designed with a throughput margin that covers those differences.

BUDGET – the result of the whole calculation

An example for a robotic station fitted with 10 additional pneumatic actuators.

The budget is COI over a given time horizon less the station's OPEX over the same period.

Why is 2020–2025 a turning point for industrial production in Poland?

It is hard not to notice how much the cost of living in Poland has risen in recent years – record inflation and rising loan instalments, fuel and food have affected all of us. That change also had consequences for employers, because average pay grew by 11.5% a year over that period — twice as fast as in the previous fifteen years. As a result, the budget that can now be put towards automation, assuming the same payback period, has risen by 70% in five years. COI – the cost of inaction – has risen by the same amount, and in the coming years it will be felt more and more, bearing directly on price pressure and competitiveness in the market – now also for the smaller companies which until now have approached production automation with a degree of reserve.

Between 2005 and 2025 average pay rose by 274% — from 2 380.29 PLN to 8 903.56 PLN, an average of 6.8% a year GUS. That growth was, however, unevenly spread: 2005–2010 was 6.3% a year, 2010–2015 only 3.9%, 2015–2020 — 5.8%, and 2020–2025 as much as 11.5%. Over those five years pay rose by 3736 PLN gross, that is more than in the previous fifteen years put together.

That change overturns investment calculations made only a few years ago and points clearly to the only possible direction of development, leaving no room for debate about whether lower pay in Poland can go on driving economic growth. Only a higher degree of automation, better quality and higher throughput can do that.

When automation makes sense and when it is better to wait

Below are three real examples, calculated according to the model described in section 3.

Three people picked six different parts into a pouch, which was then heat-sealed and packed into a master carton by hand. The cycle time of one operator: 50 s per pouch, that is roughly one pouch every 16.7 s from three stations. After building an automatic component feeding station the cycle time is 10 s for a finished, sealed and packed pouch — a 67% increase in throughput.

Alternatively – a cheaper station with a 16 s cycle. Let us work out which pays off better if we are able to sell the surplus output.

Operator involvement fell to 10 minutes once every 4 hours.

1. Station 1 – 16 s cycle, investment cost 340 thousand
2. Station 2 – 10 s cycle, investment cost 420 thousand

Case study 1 — two station variants with the same staffing

Station 1 pays back faster, because for the same freed-up staffing it costs 80 thousand PLN less. The difference, however, lies in throughput – Station 2 is able to produce 271 080 pcs more per year. For the comparison to account for different throughputs enabling additional sales, both variants have to be brought to the same throughput, and that means adding operators working by hand on a 50 s cycle to Station 1.

At the same output of 722 880 pcs a year the account therefore looks like this:

The cheaper station only wins in the first year, while sales stay at the same level. From the moment additional manual staffing appears, the advantage reverses for good: after 24 months Station 2's ROI is higher by 37 percentage points, and after 36 months by more than 100 points. Clearly – if there are no additional sales, Station 1 will ultimately generate the higher ROI thanks to its lower purchase price. It is worth being aware at the point of choosing a solution what a particular change can bring, so that you can ask yourself whether it is worth taking the risk today and paying more for a station with higher throughput. In the example given, increasing the investment by 80 thousand PLN could, over 3 years, bring an additional 583 593 PLN of savings compared with the cheaper solution – assuming that sales only grow in the second year after the solution goes live.

Two operators were placing polystyrene protectors onto finished product leaving on a pallet. The size of the spacer forced two-person handling. The robot took over the placing operation, one person was moved to another station, and the second performs the remaining operations to prepare the despatch.

One station running three shifts is freed up, that is 522 588 PLN a year. Taking, by way of example, an outlay of 500 000 PLN, OPEX comes to 58 461 PLN a year: energy 21 686 PLN, service 15 000 PLN (3% of the outlay) and residual attendance 21 775 PLN. The net saving in the first year is 464 127 PLN, so the outlay pays back after 13 months.

The operator loaded a component into the test machine, verifying 100% of production before despatch to the client. After the test they placed the part in a finished-product container. A second operator moved the containers to the warehouse on a trolley — that operation took 10 minutes once an hour. After automation a robot serves the test machine and places the parts into an automatic vertical store, which once full is collected by an AMR mobile robot and taken to the despatch warehouse.

One EOL station on three shifts is eliminated, plus 16% of the involvement in internal transport — together 606 thousand PLN a year. As you can see, without knowing the value of the investment you are already able to set your budget.

When it is better to wait or look at a different process

1. The saving exists only on paper. The model works only when the staffing cost actually disappears: the person moves to another station, you do not start another shift, or you do not hire additional people as orders grow.
2. The process is not repeatable. If the dimensional tolerances of the components are wide and success depends on what the operator compensates for with experience and a practised hand, automation can be expensive and hard to implement.
3. One workstation, one shift. At a median of 7 690.82 PLN the 24-month budget for one staffed station is 258 thousand PLN. Most stations that deliver the autonomy which generates the largest savings cost more, so such a project has to be calculated over 36–60 months.
4. A short time horizon with no possibility of reuse. If production of a particular product is ending soon and the station is a dedicated one, continuing to use it may prove impossible and modifying it uneconomic. In that case robotic stations have the advantage that they are easier to adapt to a new product or an entirely different process.
5. Small batches, high production variability. If production consists of small batches forcing frequent changeovers, changes to tool geometry or the swapping of assemblies, operator involvement may stay high, which will significantly reduce the budget and lengthen the payback period.