Samples That Mean Something: Golden Samples, PPS, and Approval Discipline

Almost every sourcing dispute that ends with the phrase "but the sample was perfect" began months earlier, when the sample process was treated as a formality instead of a contract. Samples are only useful if each one answers a defined question, is approved in writing against defined criteria, and is physically preserved as the reference the production run will be judged against. This guide lays out the standard sample stages used in China manufacturing, a sealed-sample procedure you can implement immediately, the commercial norms around sample and tooling charges, and the real reasons production goods diverge from an approved sample along with the defenses that work.

The Sample Stages and What Each One Proves

Buyers get into trouble by treating "the sample" as a single object. Mature sourcing programs recognize a sequence of samples, each built under different conditions and each capable of proving only certain things. Approving a beautiful prototype tells you nothing about what a production line will produce at rate, and confusing the two is the root of most later disappointment.

Prototype and Proof-of-Concept Samples

The earliest samples exist to test the idea: form, ergonomics, basic function. They are often hand-built, machined from the wrong material, finished cosmetically to look right, or assembled from off-tool parts. A prototype can validate a design direction. It cannot validate a supplier's manufacturing capability, and it should never be the reference for a purchase order.

Engineering Samples

Engineering samples are built to the actual specification, usually with production-intent materials and, where tooling exists, off real tools, but assembled in small numbers by skilled technicians rather than by the line. This is where dimensional reports, material certifications, and functional test data belong. Expect several iterations, and document what changed between each round, because an undocumented "small improvement" at this stage is a specification drift you will rediscover in mass production.

Pre-Production Samples

The pre-production sample, often abbreviated PPS, is the critical gate. It should be produced on the production line, with production tooling, production materials, production operators, and the final packaging. Its entire purpose is to answer one question: can this factory, using the exact process it will use for your order, produce a conforming unit. If any element is substituted, borrowed, or hand-finished, it is not a PPS, whatever the supplier calls it. Do not release mass production, and ideally do not release the production deposit, until a true PPS is approved. The linkage between sample approval and payment milestones is worth building into your contract explicitly, and our guide to payment terms with Chinese suppliers covers how approval gates and payment stages should interlock.

The Golden Sample

Once the PPS is approved, it becomes the golden sample: the physical reference standard for the order. Everything afterward, in-process checks, pre-shipment inspection, warranty claims, is judged against it. A golden sample is not a marketing object; it is the arbitration evidence you will rely on if the shipment arrives wrong.

Sealing Samples: A Written Procedure That Holds Up

A golden sample that lives loose on someone's desk is not evidence; it is a souvenir. The sealing procedure used across Chinese manufacturing, often called a sealed sample or approval sample process, is simple and worth copying exactly. Prepare at least three identical approved units. Attach a sealing label to each, recording the product name and number, purchase order reference, sample stage, approval date, an explicit list of any approved deviations, and the signatures or chops of both the buyer's representative and the factory's authorized quality manager. Then distribute custody: one sealed unit stays at the factory quality department, one goes to the buyer or the buyer's local office, and one goes to the third-party inspection company or agent who will perform the pre-shipment inspection.

Three details make the process robust. First, the label must state what the sample governs and what it does not; if color is approved but a cosmetic blemish is noted as a deviation, write it down, because silence will later be read in the factory's favor. Second, sign across the seal so tampering is visible. Third, record the sample in the purchase contract by reference, with a clause stating that the sealed sample plus the written specification together define acceptance, and that where they conflict, the written specification prevails. That last clause matters because samples age, finishes oxidize, and a physical object cannot capture tolerances.

Sample Charges and Tooling: The Commercial Norms

Buyers frequently misread sample charging as a signal of bad faith in either direction. The prevailing conventions in China are more mundane. For existing catalog products, factories commonly charge the sample cost plus courier freight, and many will credit the sample charge back against a subsequent production order; asking for that credit is normal and rarely refused. For customized products requiring engineering time, a genuine sample charge is reasonable and even reassuring, since a factory that invests real hours will want to recover them, and a factory offering everything free may be planning to recover the cost elsewhere in the quote. Courier charges are almost always the buyer's account, and providing your courier account number is standard practice.

Tooling is a separate negotiation with its own conventions. Whoever pays for a mold has the strongest claim to it, so if you fund tooling, put ownership in writing: a tooling agreement identifying the mold, stating that it is the buyer's property, prohibiting its use for other customers, and specifying its guaranteed life in shots and the conditions for transfer or release. Amortizing tooling cost into the unit price is common but weakens your ownership claim; if you accept amortization, define in the contract the quantity at which ownership transfers. None of this requires hostile drafting, but all of it must be written before the tool is cut, not after a relationship sours.

Why Good Samples Become Bad Shipments

The pattern buyers describe as "the sample was great, the bulk was a disaster" has identifiable causes, and almost all of them are preventable at the process level rather than the trust level. The most common is that the approved sample was never a true production sample: it was a technician-built showpiece, and the line simply cannot replicate it. The second is material and component substitution, where a costed-down resin, coating, chip, or fastener enters between approval and production, sometimes at a subcontractor the buyer never knew existed. The third is process drift: a different operator, a worn tool, a changed cycle time to hit a deadline. The fourth is subcontracting of part or all of the order during peak season, which silently replaces the factory you audited with one you did not. The fifth is ambiguity the factory resolved in its own favor because the specification was silent.

The defenses map one-to-one. Insist on a genuine PPS from the production line before release. Lock the bill of materials at approval and require written change approval for any substitution, with the sealed sample and specification cited in the purchase contract as the acceptance standard. Use in-process inspection on longer runs, not only a final random inspection, so drift is caught while it is correctable. Prohibit undisclosed subcontracting in the contract and verify capacity claims before ordering; researching a supplier's actual scale and production profile through a resource like our manufacturer directory is a reasonable first filter, and sector-level context from our industry reports helps you judge whether a claimed capability is plausible for a factory of that size. Finally, write the specification as if the reader were motivated to misunderstand it, because under margin pressure, someone eventually will be.

Key Takeaways

• Treat samples as a staged sequence, and never let a prototype or engineering sample stand in for a true pre-production sample built on the real line.
• Seal at least three approved units with signed labels and split custody between factory, buyer, and inspector, referencing the sealed sample in the contract.
• Expect to pay for custom samples and courier freight, ask for sample charges to be credited against production orders, and settle tooling ownership in writing before the tool is cut.
• Bulk failures usually trace to fake PPS units, material substitution, process drift, or hidden subcontracting, all of which are addressed by written gates rather than trust.
• Tie sample approval to payment milestones so that money and quality evidence move together through the project.

This guide is editorial reference material, not legal or transactional advice. Verify supplier claims and regulatory requirements independently. See our data sources and editorial policy.