Controlling automation in a company with 150 employees

Data mess that cost 12,450 PLN per month

At this company, reporting looked like a battle with windmills. Every department – from production to logistics – kept its own Excel file. When it came to putting it all together at the end of the month, it turned out that inventory levels didn't match invoices by 4.3%. Chief accountants spent an average of 31 hours in each accounting cycle on this. This was time the company paid for, which brought no profit, only increasing frustration for the office team.

We calculated it precisely: overtime alone and order errors resulting from bad data cost the plant about 12,450 PLN per month. These are not small amounts, considering that this money could have gone toward machine maintenance or bonuses for the best operators. The owner saw only generic numbers at the end of the quarter, not knowing which specific machine on the floor generated the largest material waste losses. Honestly, they were operating a bit blindly, relying on the intuition of an old foreman.

It was a typical reactive approach. Financial problems only came to light when supplier invoices started exceeding revenue from orders by a few percent. No one was able to point a finger at the source of the cash leak. Working on this project required us to go all the way down to the level of machine sensors and sheets of paper that operators filled out with grease-stained hands at the end of each shift.

Results based on numbers are the only way to stop guessing how much the company actually earned last Tuesday.

First 14 days, or cleaning the foundations

We started by auditing 84 different data sources within the company. It turned out that 19 of them were duplicated, and 6 contained critical formula errors that no one had touched since September 2019. Our expert, Ewa Kaczmarek, spent the first 3 days talking to people on the shop floor, not just management. We wanted to know how information about each produced detail actually flows, because reports in the office rarely matched what the foreman saw on the morning shift.

Instead of building large and expensive systems, we focused on a simple connection of SQL databases with an interface that any employee understands without a PhD in computer science. We chose the 12 most important indicators (KPIs) that actually impacted the company's wallet. The rest of the unnecessary tables we simply deleted so they wouldn't clutter the view. This was the turning point – suddenly data started flowing into the director's panel every 14 minutes, rather than once a week in the form of a dusty binder.

Results based on numbers – specifics after implementation

34 days after the project started, the system was fully operational. The biggest surprise for management was the fact that one of the production lines consumed 18.4% more energy than assumed in the original budget. Thanks to automatic reporting, we caught this in 2 days, not after three months when paying the electricity provider's bill. Quick machine calibration saved 3,140 PLN in the first week after changing the settings. Hard data from the production line doesn't lie.

Administrative time spent on financial summaries decreased from 26 hours a month to just 47 minutes. Now the system itself generates a PDF with the most important results and sends it to the biuro@silesiaprofit.pl email and to the owner at 7:00 AM every Monday. This also allowed for better raw material purchase planning. We reduced frozen capital in the warehouse by 92,600 PLN because the system precisely calculated exactly how much aluminum we needed for the next 11 business days, taking into account the current work pace.

We reduced reporting time from 26 hours to 47 minutes. That's 3 full business days recovered for one manager.

Staff resistance and how we broke it

Introducing changes in a company that has worked the same way for 12 years always causes fear. People on production initially thought automation was just a pretty word for an announcement of layoffs. We had to tell them honestly: we aren't looking for savings in staffing, but in the time you waste on nonsense. CNC operators got a simple tablet at their station where, instead of writing on sheets, they click three buttons at the end of the shift. This takes them 6 minutes less than before, and we are sure the data is real.

The key to success was showing foremen that thanks to this data, it would be easier for them to fight for their goals and bonuses. When they saw on the screen that their efficiency was growing thanks to better machine service planning, they stopped treating the system as an enemy. Industrial report precision became a support tool for them, not a whip over their heads. Even now, 8 months after implementation, no one in the plant has returned to paper notes in sandy notebooks.

What next? Maintaining financial discipline

Automation is not a one-time shot you can forget about. It's a process that requires oversight. Once a quarter, we drop by the plant in Katowice to check if the algorithms still calculate the margin correctly with drastically changing gas and electricity prices. In March 2024, we updated the logistics module, which allowed shaving another 2.1% off LTL transport costs. Zero redundant tables – only what actually translates into a bank transfer.

Owners of other plants in Silesia often ask us if such play is worth it with 150 people. The answer at Silesia Profit Solutions is simple: the investment in this specific system paid for itself after 158 days. Since then, every hour of personnel time saved and every percentage point of energy costs shaved off is pure profit that stays in the company. No modern fairy tales, but hard evidence that allows Mr. Marek to sleep peacefully and plan expansion by another 2 production lines.