When labor gets more expensive and orders stagnate: how to boost productivity without layoffs
Costs · 5 min read · 17.09.2026

A machine shop with 50 people, one month: roughly 8,000 paid hours. On one project, we pieced them together from attendance records and timesheets. It came out to 640 hours of vacation and sick leave, 240 of meetings and training, 580 of waiting for material, 340 of downtime with service interventions, and 260 of defect rework. That left 5,940 hours during which actual cutting took place. 74 percent. No one on that shop floor was lazy. The company simply never counted where its hours were going.
Meanwhile, the eurozone reports declining industrial production and, at the same time, a rising labor cost index. The margin of CZ/SK manufacturers is being squeezed from both sides. Each order brings in less, each hour costs more, and raising prices for customers isn't an option, because they have the same problem. That leaves a third path that most companies have never seriously explored: squeezing more output from hours they've already paid for. Based on our projects, a reserve of 15 to 25 percent is hidden in them. And decisions on overtime and headcount for the fourth quarter are being locked in over the coming weeks. So finding out what we're really paying for pays off right now.
Paid hours versus hours that cut
A basic audit is something a supervisor can handle with a spreadsheet in a single month. It doesn't require software or an outside person — it requires honesty with the numbers. One sum and four subtractions, that's the whole procedure. Add up all paid hours of the department, overtime included. Subtract absences: vacation, sick leave, doctor visits; these are paid, but you can't influence their volume in the short term. Subtract planned non-production activities, that is, meetings, training, inventory counts, cleanup. From the remainder, subtract everything during which the part wasn't moving forward — waiting for material, waiting for inspection, downtime, service, searching for tools and paperwork. What's left are production hours.
Post their share of paid hours on the department board every week, separately for each shift. One rule applies without exception: change nothing during the measurement month. Diagnosis is not a competition. When supervisors know they'll be evaluated, the numbers get dressed up and the month is wasted. A poor ratio is not the department's fault. It's the first map of where to look.

The three biggest deductions from paid hours
Waiting for material
The machine is set up. The blank isn't. The operator waits and the clock keeps running. Kanban loops for the ten most-used items, two to three cards per item, and replenishment twice per shift typically cut this waiting in half. At the machine shop mentioned above, it freed up 310 of 580 hours monthly. At a full hourly rate of €18, that's €5,500 a month. Without a single new hire.
An unbalanced line
The order's takt time is seven minutes and the third operation takes eleven, so the line produces at the pace of the slowest station while everyone else waits on the clock. Rebalance the work to takt: move a few steps to neighboring stations and put material within the operator's reach. On a four-person assembly, we raised output by 20 percent from the same 640 paid hours. No new line. No overtime.
Reworking defects
A defect doesn't get paid for once, but three times: inspection, repair, the second move. So introduce first-pass yield for every operation — the share of parts that pass the first time — and record the reason for each defect. Address the three most frequent causes, not the three loudest complaints. At one stamping shop, first-pass yield rose from 88 to 94 percent in three months, and rework hours dropped from 260 to 120 per month.
The market sets our wage rate. We set our output per hour ourselves.
Hours per part as the Monday KPI
The ratio of paid to production hours is a diagnosis. What you can actually manage is a number that changes week to week: hours per part, meaning the department's paid hours divided by good parts. 8,000 hours and 6,400 good parts gives 1.25 hours per part; at a full rate of €18, that's €22.50 of labor in every part. Every deduction from the previous chapter shows up here within a week, and a supervisor can read this number without any spreadsheets.
The ritual is fixed: Monday, 15 minutes, supervisors and the planner at the board. Three questions. What were hours per part in each product family? Which deduction was the biggest? What one action are we taking this week and who owns it? The meeting needs nothing more. And it couldn't manage with less.
This also includes the rule that saves the most on projects: no overtime without a named bottleneck. When the hours aren't going to the operation that limits output, overtime is paid waiting at a premium. One company used this rule to cut overtime from 300 to 180 hours a month. And output didn't suffer.
And once a department lives by this number, it also carries answers to sales questions. Can the order carry a price that covers hours per part, the full rate, and material? When a customer pushes the price down, we know exactly which part has to get cheaper and whether we can manage it without a new person or a new line. The answer then becomes a matter of the spreadsheet. Not a gut feeling.
Key takeaways
- Measure the ratio of production to paid hours over one month — an attendance export, timesheets, and a spreadsheet will do; change nothing during the measurement.
- Attack the three deductions we see in every workshop: waiting for material (kanban loops), an unbalanced line (balancing to takt), and defects (first-pass yield by operation).
- Introduce hours per part as the department's weekly KPI — paid hours divided by good parts — and review it with supervisors every Monday for 15 minutes.
- Assess overtime and hiring through the bottleneck and hours per part: overtime exclusively for a named bottleneck operation, orders only at a price covering hours, the rate, and material.
We can't reverse rising labor costs, and nobody guarantees us orders. European industry is moving to where investment happens, and we won't catch up with cheaper labor. We'll catch up with hours per part. A company that knows them can tell by the end of the day which order it can afford, which one to turn down, and when overtime pays off. Pick one department, open the attendance records, and start with it on Monday. If you want to make measurement, kanban, line balancing, and first-pass yield permanent, take a look at how we work on cost reduction.