Why bulk drifts from the sample — and the four causes

Every sourcing manager has opened a carton and found a garment that is recognisably the style they approved and unmistakably not the same garment. The colour is half a tone off. The shoulder sits differently. The hem is a centimetre short on the size that sells most.

Drift is not usually dishonesty and it is rarely carelessness on the line. It is almost always one of four causes, and knowing which one you are looking at tells you whether to fix the factory, the paperwork or the pattern.

1. The sample was made by the best tailor in the building

A sample room is four or five people who have been doing this for twenty years, sewing one garment slowly with the pattern in front of them. Bulk is a line of forty people doing one operation each at a set rate. The same specification produces a slightly different garment through those two routes, and the sample-room version is always the flattering one.

The correction is a pre-production sample sewn on the actual line, in the actual fabric, by the operators who will run the order. It takes two extra days and it is the single highest-value approval in the calendar, because it shows you the garment you are going to receive rather than the garment the sample room can make.

Where a style has an operation the line has not run before — a bound edge on a slippery viscose, a topstitch on a soft French terry — we also run a pilot of 30 to 50 pieces before releasing the full quantity. It is cheaper to lose fifty pieces than four thousand.

2. Fabric came from a different lot

The sample was cut from a two-metre swatch. Bulk is cut from 4,000 metres dyed in three batches. Shade varies between batches, and it varies within a batch from the beginning of the roll to the end. Weight varies. Shrinkage varies — and a 2% difference in shrinkage on a 100cm dress is a centimetre of length you did not order.

This is manageable and it is not fully eliminable. What a serious factory does is shade-band: group the rolls into bands, cut each band separately, and keep one garment's parts inside one band so a sleeve never meets a body from a different lot. Then we send you the band range, not the best roll, so you know the actual spread before goods ship.

The buyers who get bitten here are the ones who approve a lab dip against a strike-off in a hotel room. Approve it under a light box, or approve it against a physical standard you keep.

3. Nobody wrote down the tolerance

A tech pack says chest 52cm. Bulk arrives at 52.8cm. Is that a defect? It depends entirely on a number that was never on the sheet.

Half of the disputes we see between brands and factories are not quality failures, they are unstated tolerances. Everyone assumed a standard, and the two standards were different.

• Give every critical measurement a tolerance, not just a target — chest, waist, shoulder, sleeve length, body length.
• Say which measurements are critical. If eight of thirty points of measure decide whether the garment sells, we will guard those eight hardest.
• State the shade tolerance and the light source you will assess under.
• State the AQL you will inspect at, so we inspect at the same one rather than a friendlier one.

Four lines of paperwork. It removes more argument than any amount of goodwill.

4. The pattern was never graded for production

This is the expensive one, and it is invisible until bulk. A pattern that works beautifully in a sample size can grade badly: an armhole that stops being comfortable at the top of the size range, a waist that hits differently on the size that carries a third of the order. Bulk then looks correct on the size 10 you approved and wrong on the size 16 your customer buys.

The fix is fitting more than one size before approval. Fit the sample size and one extreme — usually the largest — and check the graded nest rather than the single pattern. On an ODM range we do this as standard because the pattern is ours; on an OEM order we will ask for it, and we would rather have the conversation in week two than in week six.

What to ask a factory about drift

You will not get a useful answer to 'do you have good quality control'. Every factory says yes. Ask instead:

1. Will the pre-production sample be sewn on the line that runs the order?
2. Do you shade-band, and will you send me the band range before shipment?
3. Which measurement points are you treating as critical on this style?
4. Do you run a pilot before releasing full quantity, and how many pieces?

Drift is decided in weeks two and three, not in the finishing hall. By the time you can see it, it has already been paid for.

None of this makes bulk identical to a sample. Cloth is not a printed sheet and it never behaves like one. It makes the difference small, known in advance and agreed — which is the actual standard worth holding a factory to.