260 cartons counted against an order for 300, and the 40 nobody could produce

The situation

A buyer in the US had ordered 300 cartons of upholstered dining chairs from a factory in Foshan. Flat-packed, four chairs to a carton, mixed models across four carton types. The order was confirmed, the deposit paid, and production reported complete two weeks before the booked loading date.

The factory sent the usual closing evidence: photographs of the finished stack, a photograph of the loaded container, and a packing list showing 300 cartons.

The container was sealed at the loading point on Friday afternoon. It reached the destination port eleven days later, and the receiving warehouse counted 260 cartons before the pallet even came off the truck.

Why the photographs looked fine

Nothing about the closing photographs was dishonest. The stack genuinely existed, and it genuinely held most of the order. A photograph of a warehouse stack shows depth and height, and a stack photographed from the front at forty cartons wide looks the same at forty cartons deep.

This is the structural problem with counting by photograph: **the third dimension is invisible.** A factory can photograph a stack that is 30% short and produce a completely convincing image, because the camera only ever sees two of the three axes.

• **Front elevation is not a count.** Depth is unmeasurable from a single viewpoint.
• **A packing list is a statement of intent.** It records what the factory believed it had packed. It is evidence that no error occurred, and no evidence that the error did.
• **Container capacity is not a control.** A 40ft high cube holds roughly 1,100–1,300 cartons of mixed flat-pack furniture. Loading a part-container is not physically difficult and does not look unusual from the loading dock.

What the carton count actually showed

Type D accounted for 34 of the 40 missing cartons. Those were the largest cartons in the order — two chairs each, upholstered, in the heaviest board — and they were the ones that would not stack with the rest.

The factory's explanation was that Type D was still in the upholstery queue and would be finished in three days. The loading date had been booked around the other three types, and Type D was scheduled for the following week.

**Nobody had compared the packing list against the actual stack, because nobody at either end regarded quantity as something to be verified.** It was treated as an administrative number: the factory produced it, the buyer received it, and neither party treated a discrepancy in it as a shipment defect.

The other batch in the same problem

So the goods in that container were sound and consistent. **The defect was not in the furniture; the defect was in the paperwork and the schedule.** That is a materially different finding from a quality failure, and it changes what the buyer is owed.

• **Fabric batch references across all four carton types** — consistent across the 260 counted. Single dye lot.
• **Carton labelling and model marking** — legible and correct on all types sampled. The cartons were properly identified.
• **Moisture and conditioning on the solid frame components** — within range for the destination climate.

What to add to your purchase specification

None of these five lines costs anything to write. They are the difference between discovering a 40-carton shortfall in a US receiving warehouse and discovering it in Foshan while the container is still open and the furniture is still in the factory.

• **Make the carton count a line in the PO, not a courtesy.** *Cartons to be counted and verified at loading against the packing list before container sealing. Quantity verification is part of the pre-shipment inspection scope.* Without this line, a loading-day inspection has no agreed basis for the count and the factory has no stated obligation to pause for it.
• **State the packing list is not the counting authority.** Write it plainly: *the packing list is the factory's declaration; the verified quantity of record is the count made at loading.* This one sentence removes the most common source of friction at the loading dock, because the two documents are otherwise treated as equivalent and they are not.
• **Specify the split-shipment rule in advance.** If the order is authorised to ship short: *partial shipment requires written approval before loading, with the make-up quantity, production week and batch reference stated at the time of approval.* This is the line that prevents the two-batch problem, because it forces the dye lot and the production date into the conversation before the first carton is packed rather than after.
• **Require batch references on every upholstered item and state that they may not be mixed.** *Fabric and finish batch references must be consistent across the shipment; goods from different batches may not be combined in one delivery without written approval.*
• **Set the tolerance that triggers a stop.** *A quantity variance greater than 2% or 5 cartons, whichever is lower, halts loading pending written instruction.* Without a stated tolerance, every short shipment becomes a negotiation conducted by the person nearest the container — the loading supervisor, who is not the person who owns the commercial relationship.