How Many Cartons to Open: Practical Sampling at Receiving

You cannot open every carton. It would take days, damage the packing and often spoil the goods for resale. But you cannot simply guess either, because "it looked fine" is not a method and will not stand up when the supplier disagrees. Sampling is the middle path: open a defined share, chosen without bias, and write down the rule you used. This article gives a practical way to set the size of a sample and to widen it when a defect appears. It is a working method, not a quality standard. Formal sampling plans exist, and if your contract names one, follow it.

What a sample can and cannot tell you

A sample tells you what was in the cartons you opened. It also gives a signal about the rest of the lot: if nothing is wrong in a fair sample, there is probably little wrong, and if several cartons show the same fault, there is probably more. It does not prove the condition of the cartons that stayed closed, so write "in the sample opened" and never "the whole delivery is".

The count is a different matter. Every package is counted, as usual. Sampling is for the contents: model, size, colour, quantity per carton, hidden damage and labels.

Setting the size

Agree the rule before the shipment arrives, ideally with the supplier. A rule that is easy to apply has two parts:

1. A floor: at least one carton from every pallet and from every item line. This makes sure that nothing is skipped.
2. A share: a percentage of the cartons in each line, set by risk. For example 3% for low-risk goods, 5% for medium and 10% for fragile or expensive goods, or goods that have caused trouble before.

Take the larger of the two for each line, and never fewer than a small number such as 5 cartons per line. The percentages above are examples to illustrate the idea, not a standard. Set yours from your own experience and the value of the goods.

Choosing without bias

People open the carton in front of them. That is the easiest way and the least fair. Use a rule instead:

• Systematic: every fifth or tenth carton along the pallet row.
• Spread: from the top, the middle and the bottom layer, and from the front, the middle and the back of the container.
• Random: number the pallets on slips of paper or use the random-number function of a phone, and open the cartons at the positions drawn.
• Targeted extras: add the cartons that showed a clue (a rattle, a stain, a bulge), but record them apart so that they do not distort the result.

Write down the position of each carton opened: pallet, layer and place.

When to widen the sample

Decide the trigger in advance. A simple version:

• If no defect is found, stop at the planned size.
• If one carton shows a defect, open 4 more from the same pallet.
• If two or more cartons from the same pallet show the same defect, open every carton on that pallet.

Keep the results for the opened and the unopened cartons apart in the table, so that a reader knows how firm each figure is.

A conditional example

A line has 1,000 cartons on 25 pallets of 40. Your floor is one carton per pallet, 25 cartons, and your share at 5% is 50. You open 2 from each pallet. Cracked items turn up in one carton from pallet 8 and in two cartons from pallet 7. By the rule you open 4 more cartons from pallet 8, all of which are sound, and all the remaining 38 cartons on pallet 7, where 5 more show cracks. The record: "Sample 50 of 1,000 cartons, 2 per pallet. Defect in 3 of 50 (6%). Pallet 7 opened in full: 7 of 40 cartons with cracked items. Pallet 8: 1 of 6 opened. Other pallets: no defect in the sample." The 6% belongs to the sample and not to the shipment. The defect is concentrated on pallet 7, and that is what the supplier will see.

In short

A sample is only as good as the rule behind it, and the rule has to be agreed before the first carton is opened. CargoChek agrees the sample rule with you before arrival, opens the cartons and reports what it found in the sample next to the full count.