When margins slip, the reflex is a headcount cut. It is an honest reflex: labor is the biggest controllable line on the P&L, the math is legible, and the savings land next quarter. Boards understand it. Banks understand it.

The problem is that the P&L cannot see why labor looks heavy, and in an operations-heavy business the answer is very often that the plant's bottleneck is starving everyone downstream. Cut those "underutilized" people and you have made the P&L smaller without making the system better, and frequently you have cut the exact capacity you will need the month after the constraint gets fixed.

We published a full case study on this pattern: a $14.5M components maker, four layoffs planned in finishing, where the actual problem was a CNC bottleneck losing 13% of its hours to changeovers. Finishing was not overstaffed. It was idle waiting for parts, which on a spreadsheet looks identical. Fixing the changeovers instead of cutting the people ended with EBITDA up $1.42M and the same 52 employees, absorbing 16% more volume.

Why the reflex misfires

A plant is a chain, and the chain moves at the speed of its slowest link. Downstream of the bottleneck, people will look underutilized, because they are being fed at the bottleneck's pace. That idleness is a symptom of the constraint, not a labor problem, and it is precisely the labor a cut removes because it is precisely the labor that looks cheapest to remove.

Then the constraint gets fixed, as it eventually must, and the recovered throughput has nowhere to go. You rehire, at market rates, with training time, into a labor market that remembers. The cut cost more than it saved, and the P&L that justified it never saw any of this coming because the P&L cannot see the plant.

The test before any cut

Three questions, one honest week of data:

  1. What is the constraint? The step with the queue in front of it and starvation behind it. If nobody can name it with evidence, that is the first project, not the layoff list.
  2. Is the "excess" labor downstream of it? If yes, the labor is idle because the constraint starves it. Cutting it locks in the bottleneck's pace as your permanent capacity.
  3. What is a constraint-hour worth? Revenue minus material, divided by the bottleneck's productive hours. In the Hartwell case it was $1,199. Against that number, changeover reduction, maintenance, and even overstaffing at the constraint stop looking like costs and start looking like the cheapest capacity you can buy.

None of this means labor is never the answer. Sometimes it genuinely is. A support function that outgrew its purpose, a layer of supervision the system no longer needs, capacity that demand will not return to. The discipline is the order of operations: name the constraint first, price its hours, and then look at labor with the system in view. A cut you can defend against those three questions is a real decision. A cut you cannot is a guess with severance costs.

Finance's job isn't just to report the score. It's to help design the operating system that creates it. If the margin pressure is real and the layoff spreadsheet is already open, talk to us first. One week of constraint analysis is cheaper than one wrong reduction, and we have watched it change the decision more often than not.