A workshop with 10 tabletops but only 20 of the 40 table legs it needs cannot finish more than 5 tables — no matter how much spare wood is sitting in the yard. A workshop with one machine and a growing crew eventually stops getting more output per extra worker. A machine run too slowly wastes as much fuel or electricity per unit as one run too fast. These are not three unrelated problems. They are three distinct production models, and knowing which one you are actually dealing with is what tells you where to fix the real bottleneck — not just where it feels busiest.
It's not a skill problem — it's knowing which model you're in
Most workshop owners diagnose slow output the same way: buy more material, hire more people, or push the machine harder. Sometimes that works. Often it doesn't — and the reason is that each of these three situations behaves differently, and the fix that works for one makes no difference, or even hurts, in another.
1. Fixed-proportions inputs: when spare stock can't save you
Some inputs have no flexibility between them — a table needs exactly 1 tabletop and 4 legs, always in that ratio. This is called a fixed-proportions (Leontief-style) production relationship: you cannot compensate for a shortage of one input by having more of another.
The practical trap: a manager sees full shelves of one material and assumes the line is well-stocked, while the real constraint — a different, cheaper-looking part — is quietly capping output. The maximum output is always set by whichever input is scarcest relative to its required ratio, never by the average of what's in stock.
- Check every input against the required ratio, not just against its own reorder point.
- A shortage here is a hard ceiling — no amount of overtime or extra shifts on other stations fixes it.
- Most industrial production is closer to this model than people assume — it's the reason generic "add more of X" advice so often fails on a shop floor.
2. Diminishing returns: when "just add more people" stops working
Now picture the opposite situation: one machine, and you keep adding workers around it while everything else — the machine, the space, the layout — stays fixed. This is the classic law of diminishing returns, and it moves through four predictable phases.
- Phase I: each additional worker adds more than the last one — output accelerates.
- Phase II: output still rises, but each new worker adds less than the one before (point A is where the acceleration peaks).
- Phase III: total output keeps climbing toward its ceiling, but average output per worker is now falling (point B, where the two curves cross, is where you should already be reconsidering).
- Phase IV: you've added so many people that they start getting in each other's way — total output actually falls (point C is the absolute ceiling).
The point most people miss: the ceiling (C) is not the target — point B is. Past B, each additional hire earns you less output per head than the ones already on the floor, which usually means the money is better spent on a second machine, not a fourth person.
3. Machine operating speed: why too slow costs as much as too fast
Take a car: at 1,500 rpm in fourth gear it burns more fuel per kilometer than it does at a moderate, steady speed — and above a certain speed, fuel burn per kilometer climbs again. The relationship between operating speed and input consumed per unit of output is U-shaped, not flat. The exact same logic applies to a production machine.
- Every machine has a fixed technical setup (its size, its design — this doesn't change day to day) and a variable operating speed (how fast you choose to run it, which does change shift to shift).
- Running a machine too slowly to "be careful" often wastes more energy or material per unit than running it at its designed speed.
- Running it too fast past that same optimal point usually increases scrap, wear, or energy use per unit again — speed is not a free lever.
The lesson: "run it faster" and "run it slower" are both sometimes the wrong answer — the right question is where your specific machine's optimal operating point actually sits, and that's a measurable number, not a guess.
Which one is your bottleneck
Before changing anything on the floor, work out which of these three situations you're actually in — the fix for one can be irrelevant, or even harmful, for another. Fixed-ratio shortages need inventory planning against the real ratio, not more of everything. Diminishing returns on labor need a second machine or a redesigned layout, not another hire past point B. Speed problems need the optimal operating point measured, not a blanket "faster is better" instruction. This is exactly the kind of sequencing question we work through before recommending what to build first.
Nguyễn Hải Minh
I build custom software and data solutions for manufacturing ERP systems, including INFOR, for clients in Germany. As a Staatlich geprüfter IT-Techniker (Fachrichtung Informatik) and Informationselektroniker, I combine deep technical skill with business-systems thinking to help manufacturers automate operations and optimize cross-border import and export.
More about the author →