01The exercise
Two weeks. Four numbers, recorded daily, on paper if necessary.
- Pallets received. Under inbound product, plus any purchased empty.
- Pallets shipped. Under outbound product, plus any sold or collected.
- Pallets scrapped. Broken, dumpstered, given away, burned by somebody who should not have.
- Pallets on hand. A physical count on day one and day fourteen.
Closing stock should equal opening stock, plus received, minus shipped, minus scrapped. It will not. The gap is your loss, and it is almost always larger than anyone in the building expects.
02What the number usually is
Before the count, when we ask people to guess, the answers cluster between 4 and 6%. After the count, the results look different.
| Operation type | Typical measured annual loss |
|---|---|
| Single site, own fleet, own vehicles | 3–7% |
| Single site, third-party carriers | 8–14% |
| Multi-site, internal transfers | 10–18% |
| Multi-site with customer deliveries | 14–22% |
| Seasonal or project-based | 15–30% |
A 17% annual loss on a 20,000-pallet fleet is 3,400 pallets. At twelve dollars each that is $40,800 a year, and it is invisible because it appears in the purchasing line rather than a loss line. Nobody is stealing them. They go out under product and do not come back.
03The observer effect, which is fine
Something predictable happens during the count. The loss rate falls. Not because anyone changed a policy, but because for the first time somebody at each site knows the pallets are being counted.
This is sometimes raised as an objection — the measured number is not the 'real' number because measuring changed it. I would argue that is the point. If two weeks of counting reduces loss, then counting is the intervention, and the correct response is to keep counting.
“The baseline count was the uncomfortable part. We thought we were losing maybe five percent. It was seventeen. Everything after that was easy, because nobody could argue with the number.”
04What to do with the answer
- Below 5%: you are running a tight operation. Focus on repair rate and grade matching instead.
- 5–10%: identify which lanes or sites the loss concentrates in. It is rarely evenly distributed.
- 10–18%: there is a specific mechanism — a customer who scraps them, a carrier who does not return them, a site with no collection. Find it before spending on anything else.
- Above 18%: stop any plastic pallet evaluation immediately. You would be buying $110 assets for a system that loses $12 ones.
In every case, the count comes before the solution. We run this as the first stage of any managed programme, and it is the stage customers are most reluctant to do and most glad to have done.
05Why it is never done
Because pallets belong to nobody. They are not stock, not equipment, not consumables in most chart-of-accounts structures. They arrive under things, leave under things, and no single department owns the number.
That ownership gap is the whole problem, and the count is how you close it. Eight hours of somebody's time, spread over two weeks, to find the largest uncounted cost in the building.
