The sample arrives, and it's right. The colour is right, the fit is right, the pattern sits where it should.
Six weeks later the cartons arrive and something is off. Not dramatically — the neckline is a centimetre wider, the embellishment has drifted a few millimetres, the fabric feels slightly lighter. Nothing you can point at and call "a different product." Just enough that it isn't the garment you approved.
When you raise it, the factory sends the photo of the sample. It matches. Both sides are being honest, and the order has no answer — because the approval conversation was a set of messages rather than a signed object.
A sample is where a specification stops being a document and becomes a physical standard. This is how the rounds work, how many you should expect, who pays, how long it adds, and how to approve one so that it still means something three months later.
What a sample proves — and what it can't
A sample answers one question: can this factory make this garment to this specification, at this level of finish?
It does not answer four others, and buyers get burned by assuming it does.
- It doesn't prove consistency. One sample is one unit, often made by the sample room rather than the production line. Production is where a pattern drifts when it's cut a hundred times.
- It doesn't prove the fabric. Unless you specify the exact mill, weight and composition in the tech pack, the sample fabric and the bulk fabric can be two different things that look the same in a photo.
- It doesn't prove timing. Sampling capacity and production capacity are different queues. A fast sample says nothing about a fast bulk run.
- It doesn't prove the price. A sample made by hand by a sample maker is not a cost structure. The bulk price depends on the construction you approved — that's why the tech pack has to be right before the quote, not after the sample.
What the sample does prove is worth a great deal: it removes the last excuse. Once you've approved a physical garment, "we didn't understand" stops being available to either side.
The sample types, and what each one answers
Not every order needs all of these. But it helps to know which conversation you're having, because "sample" is used to mean at least six different things.
Proto sample. The first attempt. Usually made from the sketch and the tech pack, often in whatever fabric is closest at hand. Its job is to answer: is the silhouette right? Is the proportion right? Nobody should be judging quality on a proto.
Fit sample. Refines the pattern. Goes on a body or a form, and answers: does it hang correctly, is the armhole right, does the neckline sit where you want it. Expect measurement comments and a revised pattern after this.
Salesman sample / photo sample. The presentation version — the one you photograph, show to buyers, or take to a trade show. It has to look right, which is a different standard from measuring right.
Pre-production (PP) sample. This is the one that matters. Made in the actual bulk fabric, with the actual trims, on the production pattern, by the production team. It is the last checkpoint before the line starts, and it is the sample that should become the reference standard for the whole run.
Size set / grading samples. One garment per size, to verify that the grading rules produce the right measurements across the range. Skipping this is how you end up with a perfect medium and an unwearable large.
TOP (top of production) sample. Pulled from the first units off the line. Answers: did production actually match the PP sample? This is the cheapest early warning you will ever get.
If you take one thing from this section: the PP sample is the contract. Everything before it is a design conversation. Everything after it is a compliance question.
How many rounds is normal
It depends on how far the design is from anything the factory has made before.
- A basic tee in a standard construction: one proto, one fit revision, then PP. Two rounds, sometimes three.
- An embellished or structured garment — rhinestone placement, beading, lace panels, a shaped bodice: expect three to four rounds, because placement and adhesion need physical checks that a drawing can't provide.
- A new silhouette or a new fabric: add a round. New fabric behaves differently after washing and pressing, and that only shows up in cloth.
- A repeat order in an unchanged specification: often no new sampling at all, or a single PP confirmation. If a factory wants to re-sample from scratch on an exact repeat, ask why — sometimes the answer is legitimate (new fabric lot), sometimes it's an internal process that will cost you weeks.
A useful rule: every round after the second is usually a specification problem, not a factory problem. If you're on round four and still arguing about the neckline, the tech pack is probably ambiguous rather than the sample room being incompetent.
Who pays for samples
The convention in Asian apparel manufacturing is straightforward, and worth understanding as a structure rather than a number.
You pay a sample fee, which covers the materials and the sample-room labour. Sample rooms are not production lines: a single garment made by hand costs many times what the same garment costs in a run of a thousand, and the factory can't absorb that on a speculative order.
The fee is usually credited against the bulk order. Not always, and not always in full — but on a confirmed production order it is common for the sample charge to be deducted from the invoice. Ask about this explicitly rather than assuming it. It's a normal thing to ask, and the answer tells you how the factory thinks about the relationship.
Repeat rounds may be charged again. A revision caused by a change on your side is a new sample. A revision caused by the factory failing to follow the approved spec should not be charged to you — and that distinction belongs in writing before the first round, not after the third.
Courier costs are usually separate and usually yours. Sample shipments move fast and small; they're not consolidated with bulk freight.
One thing that is not normal: a sample fee that is disproportionate to the number of samples, or a request for a large "development fee" before any specification exists. A development fee is legitimate when there is real pattern-making and material development behind it. It is a red flag when it arrives before anyone has asked you what you want to make.
How long it adds
Our standard is seven days per sample round from an approved tech pack. That's a working figure for a simple garment in available fabric; embellished styles and special-order fabrics take longer, and the clock only starts once the specification is settled.
The part buyers underestimate isn't the sample — it's the compounding.
- Round one: one week
- Your review and comments: two to five days, in practice
- Round two: one week
- Courier each way: three to seven days depending on destination
Three rounds with courier time is easily five to six weeks before a single production unit is cut. That is why sampling decisions made in a hurry in week one turn into delivery arguments in week ten. If you want the full timeline including production and shipping, this breakdown of where the calendar actually goes is the companion piece.
It's also why a clear tech pack is worth the effort: most extra sampling rounds are rework caused upstream, not factory slowness.
How to approve a sample properly
Approval is not "looks good, go ahead." It is a six-step routine, and skipping any step is how the story at the top of this article happens.
1. Measure it, don't just look at it. Use the points of measure from your tech pack, with the tolerances you specified. Write the actual numbers next to the spec numbers. A garment can look perfect and be two centimetres out at the chest.
2. Check it in the conditions it will be sold in. Wash it, or ask for a wash test. Embellishment adhesion, shrinkage and colour fastness are the three things that look fine on arrival and fail in the customer's hands. For stone and bead work, this is not optional.
3. Write comments on the garment, not in a separate message. Pin a numbered tag to each point you're commenting on and reference the number in your email. "The neckline is a bit wide" is unactionable. "Point 3: neckline 1cm wider than spec, reduce by 1cm" is a work instruction.
4. State clearly whether the sample is approved, approved with comments, or rejected. Ambiguity here is the single most expensive habit in sampling. "Looks good but maybe adjust the sleeve" means the factory will do whatever is easiest and you will discover which one in six weeks.
5. Sign, date and photograph the approved sample. All three. The photo set should include the full garment, the label, the care label, the inside seams and a close-up of any embellishment.
6. Keep a sealed counter-sample. See below — this is the step almost everyone skips, and the one that settles arguments.
The sealed counter-sample
When you approve a PP sample, ask the factory to make or retain an identical unit, seal it, sign and date across the seal, and keep it at the factory. You keep a second one.
That sealed garment is now the reference standard. If a dispute arises at final inspection, nobody argues about what the sample "looked like" or which photo is current. You open the sealed sample, put it next to the bulk unit, and measure both.
This single habit converts most quality conversations from opinion into fact — which is the same principle behind written acceptance criteria at inspection. A defect is only a defect if there is an agreed standard to compare it against.
Ask for it in the sampling email, before production starts. It costs the factory one garment and costs you nothing, and a factory that resists it is telling you something.
Approve with comments, or reject?
There is a third option people forget: approve with comments and require a second PP sample.
- Approve when the garment meets spec on measurement, construction and finish, and any remaining differences are cosmetic and within tolerance.
- Approve with comments when the differences are real but fixable, and you want written confirmation they'll be corrected on the production run. Make the comments specific and numbered, and require a revised PP sample if any of them affect fit or construction.
- Reject when the construction is wrong, the fabric is not what was specified, or the finish is below the standard you agreed. Rejecting costs a week. Accepting a wrong sample costs the whole order.
The asymmetry is the point: a week of sampling is cheap compared to a container of garments you can't sell.
The mistakes buyers make
1. Approving "close enough." Close enough on a sample becomes the new standard for the run. If it isn't right, say so while it costs one garment.
2. Approving without measuring. The most common failure, and the easiest to avoid. Eyes are bad at centimetres.
3. Changing the specification after approval. Any change after the PP sample resets the clock and often resets the price. If you must change, treat it as a new round and expect a new sample.
4. Not keeping the sealed sample. Then the argument is between a photo and a memory.
5. Judging a proto sample on quality. A proto made in the nearest available fabric by the sample room is a shape test, not a quality test. Judge it on silhouette and proportion, and hold quality judgement for the PP sample.
6. Letting sampling delay the production booking. Sample rounds and production slots compete for the same calendar. If your launch date is fixed, tell the factory at the start, so sampling can be compressed deliberately rather than accidentally.
The short version
A sample is where the specification becomes physical. The approval is only worth something if it's written, measured and signed.
- The PP sample is the contract. Make it in the bulk fabric, on the production pattern, by the production team.
- Two to four rounds is normal, depending on construction and embellishment. Extra rounds are usually a tech pack problem.
- Sample fees are normal; crediting them against the bulk order is a conversation worth having.
- Seven days per round is a working figure — the compounding with review and courier time is what people underestimate.
- Measure, photograph, sign, and keep a sealed counter-sample. That last step ends most disputes before they start.
If you're about to start sampling and want to know what a realistic round count looks like for your design, write to lemon@mtdmj.com with the style and the embellishment and we'll tell you straight — including when the answer is "this will take three rounds." Our sample room runs on a seven-day standard from an approved tech pack, with a minimum of 100 pieces per style for production and 30/70 payment terms. The questions buyers ask most are on our FAQ page, and this walk through the full production process shows where sampling sits in the run. You can also start at mtdmj.com.
MTDMJ (Senhean Garment Co., Ltd.) is a custom apparel manufacturer in Guangzhou, China. We have been making garments since 2003 — twenty-three years and more than 900 clients later, in our own factory, with our own sample room, cutting room and export team. We work on a quote basis, with a minimum of 100 pieces per style and 30/70 payment terms. We make custom T-shirts, dresses, shirts and hoodies.
Keep the conversation going. We share sourcing notes, factory-floor detail and buyer guides on LinkedIn, Facebook and Medium. If you're planning an order, write to lemon@mtdmj.com.
0 comments