Almost every quality dispute in this industry starts the same way: the buyer holds an approved sample in one hand and a production unit in the other, and the two are not the same. Sometimes the difference is real. Sometimes it is a difference in how the two were made, not in what they are.
Understanding which is which decides whether the conversation is about a defect or about a misunderstanding.
The sample was not made on the production line
Approved samples are usually made in one of three places: a dedicated sample line, the production line run slowly under supervision, or the production line during a trial run. Each produces something slightly different from a full-speed production run.
At low speed, adhesive has more time to wet out, sealing is more uniform, and the core forms with less turbulence. Run the same materials at production speed and the same settings produce a marginally different product — not defective, but not identical.
This is not a hidden trick. It is a physical consequence of running slower. What matters is whether the factory knows the difference and sets production parameters to close it, or whether it simply assumes the sample is representative.
Where the visible differences usually come from
- Colour. Nonwoven and film are produced in batches. Lot-to-lot variation of a printed or tinted material is normal and is usually within an agreed tolerance — if a tolerance was agreed.
- Thickness and loft. Core formation at speed differs from core formation at low speed. Small differences in thickness change how the pad feels and how it packs.
- Print registration. Packaging film at speed registers differently from packaging film run slowly.
- Fold and seal appearance. Cosmetic, and the difference buyers most often notice first.
- Odour. Adhesive and film can carry a faint process odour immediately after production that dissipates over days. Comparing a fresh unit with a three-week-old sample is not a fair test.
What a sample approval should actually fix
An approved sample is only useful if it is accompanied by the parameters that produced it. Otherwise it is an artefact with no instructions attached.
- Retain signed samples on both sides. One set at the factory, one with the buyer, both dated and signed. Photographs are not a substitute.
- Record the specification, not just the sample. Basis weights, dimensions, absorbency targets, adhesive weights, pack materials. The sample shows what was achieved; the spec shows what must be reproduced.
- State tolerances explicitly. Dimensions, weight, colour, print position. A tolerance that is not written down will be argued about.
- Agree the comparison method. Under what light, at what distance, with what instrument for anything measured.
The gap that cannot be closed
Some variation is irreducible. A run of two hundred thousand units will not be perfectly uniform, and no factory should promise that it will be. What a factory can promise is that the distribution sits inside a stated band, and can show measurements to prove it.
That is why process capability data is more useful than a single golden sample. A sample tells you what the line can do on a good day. Capability data tells you what it does across a run.
If you are preparing an approval, the sample revision and approval guide sets out a workable sequence, and the sample to mass production page explains how we carry parameters from trial into the production run.
Both are cheaper to read before the container ships than after.
