Importers talk about "the sample" as if it were one object. A factory does not. Between a 3D print and the unit that comes off the real line there are several different things, each proving something different — and approving the wrong one is how a container of unusable goods gets ordered with everyone's honest agreement. This article is for buyers ordering custom or branded products in commercial volumes, from roughly 500 units, and explains what to ask for at each step and what an approval actually commits you to.
Why there is more than one kind of sample
Each sample answers one question: does it look right, does the tooling work, does the production line reproduce it, and is this the exact unit the batch will be judged against? Confusing those questions is the root of the classic failure described in the sample is fine, the batch is defective. The sequence below sits in the middle of the wider process laid out in how manufacturing in China actually works.
1. Counter sample (off the shelf)
An existing product the factory already makes, sent so you can judge its general workmanship. Useful for assessing a supplier, useless as a reference for your own product — it was made to someone else's specification. Treat it as a work sample of the factory, not a preview of your order.
2. Prototype
3D-printed or CNC-machined, made without production tooling. It proves geometry, ergonomics and appearance, and it is the right stage for changing your mind — changes here cost hours, the same change after tooling costs thousands. A prototype proves nothing about the production material, strength or surface finish, because it was not made by the process that will make your goods.
3. Tooling sample (T1, T2 trial shots)
The first units out of the new mould or die. This is where the tooling itself is validated: dimensions, shrinkage, flash, sink marks, whether the part releases cleanly. Expect two or three rounds — T1 rarely passes, and the corrections are normal toolmaking, not a bad sign. Who pays for tooling modifications at this stage should already be settled in your agreement; the whole question is in tooling and moulds: cost and ownership.
4. Pre-production sample (PP sample)
The one that matters. A PP sample is made on the real production line, with the real material, by the real workers, in the real packaging. It is the only sample that predicts what your batch will look like. A hand-finished unit polished by an engineer in the sample room is not a PP sample, however good it looks — and asking whether a sample was made in the sample room or on the line is one of the most useful questions in the entire process.
Approve the PP sample only after checking it against the document, not against your memory of the prototype: dimensions, material certificates, function tests, colour, markings and packaging. What that document should contain is in the technical spec sheet for a Chinese factory.
5. The golden sample
Once a PP sample is approved, it stops being a sample and becomes the reference. The golden sample is the physical standard the entire order is measured against — by you, by the factory and by the inspector.
Make it real, not symbolic:
- Three identical units, approved at the same time: one for you, one sealed at the factory, one for the inspection agency.
- Signed and dated by both sides, sealed so it cannot be swapped.
- Photographed from all sides at high resolution, with the seal visible.
- Referenced by name in the contract as the acceptance standard, alongside the spec revision number.
- Handed to the inspector before the pre-shipment inspection, so rejection has a physical basis. How that inspection works is in quality control when importing from China.
What samples cost, and why they are sometimes "free"
Expect to pay sample charges plus courier, often with the cost credited against the order. A tooling sample costs whatever the tooling round costs. A factory that insists samples are free for a first-time customer is usually pricing them into the unit cost — the same logic as the "free mould," and worth knowing about rather than being charmed by.
One rule holds regardless: never let a sample charge become a reason to skip the PP stage. It is the cheapest quality control money can buy.
The mistakes that cost a container
- Approving a sample room unit as if it were a production unit. Different people, different material batch, different care.
- Approving verbally or by photo. With no signed, sealed physical reference, "not as agreed" is an opinion, not a claim.
- Approving a sample in a different material or colour "because the real one will be the same." It will not.
- Keeping no sealed copy yourself. If the only reference sits in the factory, the reference is whatever the factory says it is.
A sample is not a gift or a formality. It is the only moment in the entire order when the product exists, you can hold it, and the money is not yet spent. Everything after that point is measured against the unit you signed.
Running samples the way a buyer should
For our clients we specify each sample stage, check PP samples against the spec rather than against expectations, seal and hold the golden sample, and use it as the acceptance standard at pre-shipment inspection — the protocol behind 3,500+ deliveries and 340+ verified suppliers over seven years. Real projects are in our case studies.
Start with product sourcing, or order a standalone quality inspection against your existing golden sample.
