The Buffer

Walk into a full warehouse and everything in you relaxes. Racks to the ceiling, weeks of cover, every order shipping from a shelf. Nothing in a plant looks as reassuring as a full rack, and few things are quietly costing more.

Look closer and the stock is working. A supplier runs two days late and nobody ever hears about it; the shelf absorbs him. A changeover overruns by an hour — absorbed. A machine loses an afternoon, and the order still ships on time from the rack. Each one is a problem solved silently, by inventory, at full price, and nobody is asked.

None of this is mismanagement. The buffer is doing exactly what it was bought to do, which is to absorb variation. It just cannot tell the difference between the variation you accept and the signal you need. A problem the shelf catches is a problem nobody escalates, and a problem nobody escalates gets fixed by no one. So the operation pays twice: once for the stock, and once for every weakness the stock keeps comfortable.

Taking the cover down is uncomfortable by design. Done in steps, the buffer starts giving its problems back: small stops surface that used to vanish, and for a few weeks the plant feels worse while it is getting better. Changeovers have to shrink so that small batches are affordable, and losses have to be read by the hour, because a monthly average is nothing anyone can act on. At a plant I ran, delivery climbed as the stock came down. The problems had finally become expensive enough to fix.

So the question on the next warehouse tour is not how many weeks of cover are left. It is this: what is this stock apologising for? Every full rack is a receipt for a problem somewhere else in the plant, one that stopped announcing itself the day the cover arrived and has been on the payroll ever since.

Where this shows up: Production turnaround — getting delivery, cost and quality back under control.