Most brands hold a photo of their approved sample rather than the sample itself, & most factories keep the sample but not the paperwork that says what it was approved against. Both choices look harmless while nothing goes wrong. They stop looking harmless the first time a batch arrives that the brand is sure is different, because at that point the only useful object in the relationship is a sealed, dated, signed sample of the product that both sides agreed to make.
This article explains what a gold sample is, why two copies should exist rather than one, how each copy is sealed & labelled so that it can be relied on years later, what storage conditions it needs, which events bring it off the shelf, & when it should be retired. It is a short, practical piece about a small object that decides how quickly most quality arguments get settled.
What a gold sample is meant to prove
A gold sample, sometimes called an approved reference standard, is a physical unit of the product as agreed, retained by both parties, sealed & labelled so that it can be identified without reference to anyone's memory. It is not a marketing prop & it is not a gift. It is an evidence item, & a factory handles it exactly as it handles a retention sample.
What it proves is narrower than many brands expect. It shows what the product looked like, smelled like & weighed at the moment of approval, under whatever storage it has since received. It does not prove that a later batch is outside specification unless the specification itself was written down; without a specification sheet, the gold sample is a comparison, not a criterion.
The two documents therefore travel together. The gold sample is the physical half & the specification sheet is the written half. Where a brand holds only one of the two, the conversation about a disputed batch becomes a discussion about preference, & preference is the hardest thing to resolve across a language difference & a nine-hour time gap.
Why two copies exist rather than one
A single copy creates a structural problem. If the brand holds it, the factory cannot compare an in-process batch against the reference without asking for it to be shipped back, which introduces transit, heat exposure & delay. If the factory holds it, the brand cannot check an incoming delivery the day it lands in a warehouse.
Two copies also protect against the failure that neither side expects: loss. A sample shelf can be cleared by a move, a fire, a flood or simply a tidy-up by someone who did not know what the object was. Two samples in two buildings, ideally in two countries, reduce that to a nuisance rather than an argument.
A third reason is commercial. When a formulator leaves, a factory changes hands, or a brand moves to a new supplier, the copy held by the other side is the only continuous record of what was agreed. Succession is exactly the moment when institutional memory is weakest, & it is also the moment when a new supplier is most likely to be asked to match an existing product.
Where production is split across two sites, a third copy for the second site is often sensible, but only if there is a written rule about which copy governs when the three disagree. Two copies with a rule beats three copies without one.
Sealing, labelling, signing: making the sample evidence
A sample becomes evidence when it is sealed in a way that shows interference. A tamper-evident bag, a heat-sealed pouch or a sealed container with a signed tape across the closure all work; a box that has been opened three times does not. The seal should carry a signature or a stamp that crosses the closure, so that a break is visible.
The label carries the facts that make the object identifiable. Product name, formula version, batch or sample code, fill weight, pack format, date of manufacture or sampling, date of approval, & the name of the approving person. Where a fragrance or a colourant is involved, the raw material lot used for the sampled batch belongs on the label or in a paper record that names the sample, because those two attributes vary between lots & are the ones most frequently disputed.
Both copies should be labelled identically & approved in the same exchange. The practical method is to prepare both, seal both, label both, & have the brand's signature applied to both in the same document exchange that approves the product. Where a courier carries the brand's copy, the photograph of the sealed & signed sample before dispatch completes the chain.
One line on the approval document should state that the two samples are the approved reference & that both are held under the conditions described. That sentence is what turns a jar on a shelf into something that can be produced in a conversation with a customer, an assessor or a new supplier.
Storage conditions: what a gold sample can survive
A gold sample should be stored under the conditions the finished product's label claims, or better. That usually means cool, dark, dry, away from strong odours, & in the primary packaging where the pack affects the product. A capsule kept loose in a plastic bag is not stored under the conditions it was designed for, & whatever it shows after a year is evidence about the bag rather than about the product.
Practical storage varies with the format. Oil-filled capsules are sensitive to heat & light, & a warm office shelf accelerates shell softening & colour change. Emulsions separate if they are allowed to cycle through temperature swings, which is why a store room next to a loading bay door is a poor choice. Alcohol-rich products have their own behaviour in high heat, & the pack gaskets & pump components may be affected long before the liquid is.
The stored sample should be logged, not just placed. A register that ties the sample code to a shelf position means a request resolves in minutes rather than in an afternoon of searching. Where more than one product is in development, the register is what stops the wrong sample being produced as the reference in a discussion.
A useful discipline is to check the reference at each reorder & note the observation. A short line recording appearance, odour & fill weight at that moment builds a history of the reference itself, so that when it eventually changes, someone knows whether the change happened in the sample or in the product.
The triggers that bring a sample off the shelf
The first trigger is an incoming batch that looks or feels different. The reference is opened alongside the new units, compared in the same light, & the result recorded. That comparison takes fifteen minutes & it either closes the question or converts it into a technical one with evidence attached.
The second trigger is a reorder after a gap. Where a product has not been made for a year, the raw materials will have changed lots & possibly suppliers, & the process will have been run on different equipment or by different operators. Comparing the new first-article against the gold sample at that point is standard practice in food & pharmaceutical manufacturing & is just as useful in cosmetics.
The third trigger is a change of supplier or of formulator. A brand moving production, or a factory taking on a product it did not develop, needs the reference to reverse-engineer the target without relying on a description. A written specification is essential here, but the sample is what settles the sensory question that a specification cannot capture.
A fourth, less obvious trigger is an internal one: a claim or a pack change. Colourants, fragrances & even preservation systems can be forced to change by a regulatory update, & the reference is what shows whether the reformulated version is still recognisably the same product to the brand's own customers.
What happens when only one side holds a copy
When only the factory holds a copy, the brand's ability to check an incoming delivery depends on a third party's willingness & speed. A comparison request becomes a project, & a project can be deferred. The practical result is that small drifts are tolerated until they become large enough to generate complaints, by which point several batches have been produced.
When only the brand holds a copy, the factory works without a reference on the bench. Operators compare against the last batch they made rather than against the approved original, which is how gradual drift happens inside a specification. Nobody has done anything wrong; the control simply was not on the line.
When neither side holds a sealed copy, the argument is decided by whichever party has the better records & the louder voice. This is the situation most disputes in cosmetic contract manufacturing actually reach, & it is entirely preventable at the cost of one extra unit & twenty minutes of paperwork.
When a gold sample retires
A reference does not last forever, & pretending otherwise creates a different problem. The product itself changes over time even under good storage, & after the labelled shelf life plus a margin, the sample no longer represents anything that could still be on a shelf. At that point it should be retired formally, & a fresh reference drawn from a current approved batch.
A workable rule is to retain the reference for the labelled shelf life plus a defined margin, & to renew it whenever the formula version changes, the pack changes in a way that affects the product, or the raw material supply changes materially. Each renewal is a small exercise: prepare two, seal two, label two, sign two, & record the reason for the renewal.
Retirement should be recorded rather than implied. A note in the register saying which reference was replaced, by which, & on whose authority keeps the history readable. Without it, a drawer of unlabelled jars accumulates, & the value of the whole programme collapses into a storage problem.
Custody compared across sample types
The table below sets out how the main sample types used in a cosmetic project differ in what they represent & what custody they need. It is a comparison of roles rather than a fixed procedure, since a single project may not use every row.
| Sample type | What it represents | Who should hold one | Typical retention |
|---|---|---|---|
| Development sample | A formula direction under evaluation | Factory; brand if requested | Until the next round |
| Gold sample | The product as approved for production | Both sides, sealed | Shelf life plus a margin |
| Pre-production sample | The first batch on the real line | Both sides | At least one reorder cycle |
| Retention sample | Each commercial batch as released | Factory; brand where agreed | Per the applicable record rules |
| Stability sample | Behaviour under stated conditions | Factory laboratory | Until the study concludes |
| Pack retention | The artwork & pack as printed | Both sides | With the batch records |
| Reference for a claim | The unit behind published wording | Brand, with factory records | As long as the claim is used |
The retention column is a working norm rather than a legal rule, & the period attached to batch records in a given market may be longer than the period the physical sample can usefully survive. Where the two differ, keep the records for the longer period & the sample for as long as it remains representative.
Where custody arrangements fail
The most common failure is a sample that cannot be identified. It was sealed, it was stored, & the label fell off or never named the formula version. The object survives but the link to the approval does not, & an unidentified sample is worth very little in a discussion.
The second is a sample held in the wrong place. A reference kept in a sales office, a showroom or a trade show kit has usually experienced heat, light & handling that the product was never designed for. The sample still exists, but any observation drawn from it is about the journey rather than the product.
The third is an unsigned pair. Two samples were prepared & both were kept, but only one carries a signature, so the brand's copy cannot be established as the reference the factory approved. The fix is procedural & takes one document exchange, which is why it belongs in the initial approval rather than in a later correction.
The fourth is the quiet one: a reference that was never renewed after a reformulation, so the brand & the factory are comparing against different generations of the same product without either of them knowing it.
What to put in the brief
Settle custody while the project is still commercial rather than technical, because that is when both sides are motivated to agree. Name the person who will hold the brand's copy, name the storage location, & state the retention period you want. Ask the factory to confirm the same three things on its side, & ask for the reference to be renewed whenever the formula version changes.
- The formula version & sample code that the gold sample represents.
- Which party holds which copy, & at which site each copy is stored.
- The sealing method, & whose signature appears across the closure.
- The storage conditions, stated in terms of the labelling claim the product makes.
- The retention period for the reference, & the trigger for renewal.
- The procedure for comparing a new batch against the reference, & who records the outcome.
- What happens to the reference if the brand changes supplier or formulator.
- Who is authorised to retire a reference, & how that retirement is recorded.
Add one line about transport, since references occasionally travel. A sample sent by courier across a border may be opened by customs, which breaks the seal & ends its life as a reference. Where a copy has to move, send a second sealed copy in the same consignment & keep the first at home, so that a customs inspection costs a sample rather than the programme.
Where our capability fits
We prepare two sealed references for every approved product, sign both in the same exchange that approves the specification, & label each with the formula version, the sample code, the fill weight & the approval date. The copy that stays with us is logged against a shelf position, & the register is renewed whenever a formula version or a pack changes. Where a brand asks us to compare a new batch against the reference, we record the observation in writing rather than reporting an impression.
We work with brands on private label (your brand on a proven base), OEM/ODM contract manufacturing & full custom formula development, & the reference discipline is identical across all three. Our lines include a retinol capsule essence, a blister-packed argan hair oil capsule set & a copper peptide firming capsule essence, each of which is released against a retained reference as well as a batch record.
Tell us which product you want held as a reference, who will keep your copy, & how long you need it, & we will set the arrangements up before the first batch is released. Start at our inquiry page, email capsulebiotech@163.com, or message the right team on WhatsApp: Capsule Series (capsules only) +86 158-7655-5319, Cosmetics Series (serum, hair care, body care, fragrance) +86 158-7655-5319. How the reference is created during sampling is covered in our guide to the sampling process, & how references & records support batch consistency in our article on factory quality control.
General manufacturing information, not regulatory or legal advice. Record retention periods, sample storage duties & labelling requirements vary by market & by product, & the applicable requirements should be confirmed for each destination before a product is placed on sale.