Cut the spare capacity and an operation improves on every number that travels upward. Fewer people carrying the same output. Less stock on the shelf. Crews and machines running closer to full. Each of those moves reads as good management, and most get rewarded as such at the review. The bill for them lands later, in a form the efficiency drive never has to sign for.
That is the part worth sitting with. Efficiency is booked the moment you capture it. The fragility that came with it is paid for on a date nobody has told you yet.
The date is closer than most plans assume. The average company can expect a disruption to production lasting a month or longer roughly once every 3.7 years, and over a decade a serious shock can wipe out more than 40% of a single year’s profits (McKinsey Global Institute, August 2020). A one-in-four-year event is a standing feature of operating life. The only thing you cannot know in advance is which quarter it lands in.
When a business tunes itself for maximum efficiency, what it usually strips out is slack: the spare crew, the buffer stock, the person who was not fully loaded, the day in the schedule with nothing booked into it. On a calm week all of that looks like waste, and a good deal of energy goes into removing it. The removal is real money, and it shows up fast. What goes with it is much harder to see.
Where operational efficiency misleads
Eliyahu Goldratt built a career on an unfashionable observation: the efficiency of any single part of a system tells you almost nothing about what the whole system produces. A machine, a crew, a department can run flat out, look magnificent on its own numbers, and bring the business no closer to earning a pound. Being busy and making money are different things. A resource held at full load is often just producing inventory and delay for someone further down the line to deal with.
This is why an efficiency programme that chases utilisation everywhere so often makes the operation worse while every local dashboard turns green. Each part is tuned on its own terms. The flow between the parts, which is where the customer actually gets served, is where everything then seizes up.
You can watch this happen with a single utilisation target on a field force. Tell every engineer they must be productive for ninety-odd per cent of the day and the number will come good, because people are resourceful. What the number will not show you is the job left half-finished so the next one could start on time, the customer visited once when the fault needed two trips, the quiet decision to skip a check that was not strictly billable. Every local clock reads green. The rework, the repeat visits and the complaints all land somewhere the utilisation figure never looks, and together they cost far more than the idle time the target was set to chase out.
The slack was doing a job
Absorbing variation is the quiet job spare capacity does. Every real operation runs on variation: demand that spikes, a supplier who slips a week, two key people off sick at once, a batch that has to be redone. Slack is what lets the operation take those in its stride without the whole schedule lurching. Pull it out to lift utilisation and the work does not run smoother. It runs as a queue with nowhere to go the moment anything deviates from the plan.
In my years rebuilding the operating side of businesses, I observed operations running at what looked like textbook utilisation, every crew booked, every day full, and found them unable to absorb one ordinary bad week. A single depot down, one unusually large order, one run of weather, and the schedule fell over, because there was no reserve anywhere in it to take the hit. The people running those operations were proud of how tightly everything was packed. That tightness was the fault. Every hour was already spoken for, so any hour that went wrong had to be stolen from another job, and one bad morning rippled through the plan for a fortnight.
Why the bill stays hidden
The habit survives because of an accounting mismatch, and it is worth being precise about it. The saving from cutting slack lands this quarter, on a line every manager can point to. The cost of having cut it arrives in some later quarter wearing a disguise: a supplier failure, a demand shock, a key person walking out the door. It gets filed under bad luck. Almost nobody traces the lost fortnight back to the buffer that was taken out two years earlier to make a margin look tidier. By the time the shock arrives, the manager who made the cut has often moved up on the strength of it, and whoever holds the operation now has inherited a fragility they did not build and can barely see. So the efficiency drive is credited every time and debited never, and it runs again the following year.
What good operators do on purpose
The move that separates a resilient operation from a merely efficient one is deliberate, and it looks wrong on a spreadsheet. Good operators run parts of the operation below full capacity on purpose. Where a part is not the constraint, loading it to the ceiling buys nothing except queues and rigidity, so they leave it with room to breathe and place the reserve where a shock would otherwise stop everything. Goldratt made the same point in a factory decades ago: reducing the efficiency of a non-critical resource can lift the output of the whole. Spare capacity sitting exactly where the next hit would land is what lets an operation take a punch and keep serving customers.
None of this is a case for carrying fat everywhere. Some slack is genuine waste and should be cut without ceremony. The skill is telling the two apart: which spare capacity is protecting the flow and which is only cost. Standard efficiency programmes are built to find slack and remove it. Very few are built to ask what each piece of it was holding up before it goes.
A test for your own operation
Here is a plain question to put to your operation. When did you last have a genuinely bad week, and how long did the disruption actually last? If a single supplier slip or one demand spike still costs you a month of recovery, you are carrying all of the efficiency and none of the protection. The McKinsey clock is running on you whether or not the risk has been priced. It is worth finding the points where a shock would actually stop your operation while things are calm enough to do something about them.
The honest version of that question is an uncomfortable one, because a smooth run of good weeks tells you very little. An operation shows what it is made of only when something goes wrong, and by then the reserve is either already in place or it is absent. Building it in the middle of a shock is the most expensive moment to try, and usually far too late to matter.
An operation built only for efficiency is tuned for the conditions it has already met. An operation built to take a hit and keep running costs a little more on a quiet day and is still standing on the bad one. The bad one is coming. The only choice you have is whether to pay for the reserve in advance, on purpose, or to pay for its absence later, by surprise.
References
- McKinsey Global Institute. “Risk, resilience, and rebalancing in global value chains.” August 2020. Read Article
