Why Your Sample Doesn’t Match the Bulk Order

It is one of the most common and most expensive disappointments in textile sourcing. The sample arrives, everyone in the office passes it around, the design team signs off, you place the order — and eight weeks later a carton lands that is recognisably the same product and unmistakably not the same quality. The colour is fractionally flatter. It feels a little thinner. It is a centimetre short. Nothing is wrong enough to reject, and nothing is right enough to sell at the price you planned.

The instinct is to assume you were shown a “golden sample” — a deliberately better piece built to win the order. That happens. But in our experience running a mill since 2006, it is the minority. Most sample-and-bulk mismatches are not deception; they are the predictable result of five variables that are controlled in a sample and uncontrolled in production. Understanding which variable drifted tells you whether you have a supplier problem or a specification problem — and only one of those is fixable by arguing.

Five variables that drift between sample and bulk

1. The finishing hand

What changes: fringe knotting, edge straightness, and the final softness of the hand-feel are the most labour-dependent steps in the whole process. A sample is finished by one skilled worker; a run of 2,000 pieces is finished by a team across several shifts.

How to tell this is your problem: the bulk pieces vary among themselves — some fringes are tighter, some hems sit straighter. If the bulk is consistently slightly worse than the sample, that is a different cause. If it is unevenly worse, it is the finishing step.

The fix: ask for the fringe and hem method to be written into the PO as a named standard, and ask how many people will finish the order. A mill that owns its finishing can answer this; one that subcontracts cannot.

2. Dye lot

What changes: a sample is often dyed as a small batch, sometimes alongside other work, using yarn that was already in the dye house. Your production order is dyed as its own batch — and every dye batch has a slightly different result, even with an identical recipe.

How to tell: the bulk is uniformly a shade off from the sample, but internally consistent. This is the single most common complaint, and it is usually a real limit of the process rather than a mistake.

The fix: this is where a retained sample and a dye recipe matter. A mill that keeps a customer’s approved sample and shades the repeat against it will hold the difference under a perceptible threshold. A mill working from a photo will not. Ask the direct question: “do you keep a retained sample and the dye recipe for our approved lot?” If the answer is vague, expect drift on every reorder, not just this one.

3. Weight and GSM tolerance

What changes: GSM is a target, not a fixed value. A production run carries a tolerance band, typically ±5% and sometimes wider on jacquard. If your sample happened to land at the top of the band and your bulk at the bottom, that is a 10% difference in hand-weight — entirely within specification, and very easy to feel.

How to tell: weigh ten pieces from the bulk and compare to the sample. If the average is within 5% but the sample sat at the high end, you have a tolerance problem, not a quality problem.

The fix: specify the band, not just the target, and state that the approved sample defines the acceptable floor. Write it as: “350 GSM ±5%, and no piece lighter than the approved production sample.” That single clause removes the entire ambiguity.

4. Shrinkage, and when the measurement was taken

What changes: cotton peshtemal shrinks. A 100×180 cm woven piece commonly finishes around 95×175 cm after the first wash, and some constructions move more. A sample measured before washing and a bulk order measured after will differ by exactly that amount — and both parties will believe they are right.

How to tell: the difference is proportional on both dimensions and consistent across every piece. If the shortfall is uniform, it is shrinkage.

The fix: state whether your size is woven size or washed size in the PO, and ask the mill to confirm which one the sample represents. Most disputes about size are two people using the same words for different measurements.

5. Yarn lot and source

What changes: the cotton yarn itself varies between lots — fibre length, combing quality, and the mill that spun it. A sample may have been woven from a premium lot that was on the floor at the time. Production three months later draws on a different lot.

How to tell: the bulk feels slightly rougher or picks up a faint halo of loose fibre, and this becomes more apparent after two or three washes rather than immediately.

The fix: for repeat programs, specify the yarn source and combing standard, and ask to be told in advance if the source is changing. A supplier who tells you “the yarn lot is switching, here is a comparison piece” is protecting your brand; one who lets you find out from your customers is not.

The approval process that actually prevents this

The single most effective change most buyers can make is to stop approving a hand-made sample and start approving a production sample pulled off the running loom, under real production conditions. The sequence looks like this:

Step	What happens	What you sign off on
1. Development sample	Made to establish the design, colours and hand-feel	The direction — not the final spec
2. Spec sheet	Size (woven/washed), GSM band, yarn count, fringe method, dye reference, packing	Every measurable value in writing, in the PO
3. Production approval piece	One piece pulled from your actual run, finished and washed like the bulk	The real thing, under real conditions
4. Bulk continues	Only after that piece is approved in writing	Your approval holds the run
5. Pre-shipment QC	Measurements, weight and defects checked against the approved piece	The final gate before it ships

Steps 3 and 5 are the ones buyers most often skip, and they are the two that would have prevented almost every mismatch we have been asked to help resolve. A production approval piece delays the order by a few days. A bulk shipment that misses costs a season.

The specification sheet: what to put in writing

If you change only one thing about how you buy, change this: put every measurable value in the purchase order, and make the approved piece the legal definition of “correct”. A specification sheet for a peshtemal should carry all of the following:

• Size, stated as woven or washed, with the tolerance allowed on each dimension
• GSM with a band (e.g. 350±5%), plus “no piece lighter than the approved piece”
• Weave construction and warp/weft yarn count, and whether the cotton is combed
• Dye reference: yarn-dyed or piece-dyed, and the colour standard used (Pantone chip, or the retained sample)
• Fringe method and count, and hem type
• Label type, placement and sewing method
• Packing: folding, bag type, carton count and carton markings
• Defect tolerance: the maximum acceptable percentage, agreed in advance
• Retained sample: confirmation that the mill keeps your approved piece and dye recipe on file

The approved sample should not be a memory of what you hoped for. It should be the physical clause in your contract that says what you are willing to accept.

What to do when it has already gone wrong

If a shipment has already arrived off-spec, the useful questions are diagnostic ones, and they also tell you what kind of supplier you are dealing with:

1. Is it uniformly off, or unevenly off? Uniform points to dye lot, tolerance or shrinkage. Uneven points to finishing labour. The answer determines whether a rework is possible.
2. Measure and weigh, do not describe. “It feels thinner” gets you a polite reply. “Ten pieces average 318 GSM against a 350±5% specification” gets you a solution, because it is a fact rather than an impression.
3. Ask for the production record. A mill with records can tell you the actual GSM and the actual dye batch of your order within a day. A trader will not be able to produce it at all — which is itself the most useful piece of information you will get.
4. Decide on the relationship, not just the shipment. Every supplier will eventually have one bad run. What separates a partner from a vendor is whether the response is a rework plan and a process change, or an explanation of why it is within tolerance.

Frequently asked questions