Global buyer checklistBefore mass production

Golden sample and sample approval

Approve a traceable sample version, not an impression. Record what arrived, how you checked it, any differences and what work your decision allows.

· · Source notes below

Prepared by Bao L. Zhou, Currawong’s China-side verification desk for international buyers.

Also available in Español, Português and Tiếng Việt.

SAMPLE BASELINE / 01

Make the sample review repeatable

Another reviewer should be able to identify the same sample, repeat the key checks and see exactly what was approved.

01

Identify the sample and its source

Record sample type, supplier, workshop, model, ID, revision, quantity, dispatch date, receipt date and reviewer. Photograph labels and features that distinguish it.

Give each version its own ID. Mark older versions as superseded; “latest sample” is not a stable reference.

02

Fix the reference documents

List the specification, drawing, bill of materials, colour, artwork, labels and packing instructions by ID and revision. State which controls if they disagree.

Define the product specification before judging the sample. Save the reviewed files so later changes remain visible.

03

Use measurable criteria

Record intended use, key functions, dimensions, tolerances, materials, assembly checks and test methods. Note equipment, conditions, units and reviewer where they affect the result.

Use written criteria for measurable requirements. A physical sample helps with appearance and feel; it cannot replace material, performance or destination requirements.

04

Check appearance and packing

Check finish, colour, fit, moving parts, accessories, artwork, labels and packing against the agreed scope. Photograph the reviewed state and record lighting or other relevant conditions.

List temporary materials, manual finishes and prototype-only features. Record each accepted difference, its owner and the condition for closing it.

05

Match test reports to this sample

Match report number, sample ID, model, material, dates, method and result. Check that the laboratory’s scope covers the required test and that the report answers the intended product question.

Laboratory competence does not prove who supplied the item or how future lots will perform. Check certificates and reports.

06

Record differences and the decision

Log each difference, required value, observed result, impact, evidence, owner, due date and action. Choose a clear state: rejected, revision required, conditional or approved reference.

State what conditional approval allows. Keep required hold points open until an authorised buyer records their closure.

07

Label and keep the reference

Where practical, keep matching buyer and supplier samples. Label them with ID, revision and approval date. Record the custodian, storage, seal condition, photographs and limits caused by ageing or damage.

The written specification still controls characteristics that one physical item cannot preserve, such as composition, hidden construction or long-term performance.

08

Connect approval to production and inspection

Link the approved sample to the PO, specification, supplier response and inspection brief. Name the revision, site, quantity or stage, remaining holds and person who may release work.

Review changes to the site, materials, tooling, design, packing, test evidence or destination. Decide whether a new sample or test is needed before affected work starts.

One approved sample and a later lot-sampling decision answer different questions. Use ISO 2859-1:2026 only where the agreed lot-sampling plan calls for it.

Next: purchase order, production inspection and pre-shipment inspection.

APPROVAL RECORD / 02

Minimum fields for a reviewable sample decision

Control areaRecordEvidenceHold when
IdentitySupplier, site, SKU, sample type, ID, revision and datesLabels, dispatch record and dated photographsThe reviewed item cannot be distinguished from another version
Comparison packControlling specifications, drawings, materials, artwork and packaging revisionsVersioned files and conflict ruleTwo references disagree or can be replaced without approval
Function and measurementIntended use, methods, equipment, tolerances and resultsCheck sheets, measurements and reviewer recordA critical requirement has no objective acceptance method
Appearance and pack-outFinish, assembly, accessories, branding, labels and packagingDated photographs and approved exceptionsA prototype-only or temporary feature is not declared
Test evidenceReport, tested-item identity, method, result and laboratory scopeSource-checked report and sample matchThe report cannot be connected to this sample and decision
ExceptionsRequirement, observed difference, impact, owner, due date and actionIssue log and closure evidenceAn open exception is hidden inside a general approval
Retained referenceBuyer/supplier copies, labels, custodians, storage and conditionSigned labels, seal and custody photographsThe reference is missing, changed or no longer comparable
HandoffDecision state, authorised owner, released work, PO and inspection linksSigned revision and supplier acknowledgementProduction can start from an ambiguous or superseded version

DECISION STATES / 03

Separate sample acceptance from production release

StateMeaningPermitted next action
Development / rejectedThe item does not yet meet the comparison pack or intended-use criteriaRequest a corrected, newly identified sample
ConditionalNamed differences are accepted only for evaluation or a limited purposeClose the written conditions; do not infer mass-production release
Approved referenceThe exact sample revision is accepted as one controlled comparison referenceComplete the PO, evidence and release gates
Production releasedAn authorised owner has linked the approved revision to defined work and closed mandatory holdsStart only the named stage, quantity and revision
Superseded / reapproval requiredA relevant change or later revision has replaced the prior baselineHold affected work until impact review and a new decision

Check where the sample was made

Branch and investment records do not prove where a sample or production lot was made. Ask for both site names and addresses; review any difference before production release.

Branch and investment records returned for 45 companies. Queried 21–22 August 2026.
Registered structureCompaniesWhat was returned
Branch records returned8 of 45Branch registration records
Investment records returned18 of 45Investee company records
Neither type returned24 of 45Neither record type in this query
Branch and investment records returned for 45 companies. Queried 21–22 August 2026. Branch records returned: 8 of 45; Investment records returned: 18 of 45; Neither type returned: 24 of 45.
Branch and investment records returned for 45 companies. Queried 21–22 August 2026.

The groups overlap: 5 companies returned both branch and investment records. “Neither” means neither type was returned in these queries; it does not prove one address or no other site.

Source: 45-company NHTSA-linked sample, licensed-platform queries on 21–22 August 2026. This selected group does not represent all suppliers.

METHOD / 04

How this checklist was prepared

The ISO sources below cover version control, clear buyer requirements, laboratory competence and lot sampling. The eight checks and decision states are Currawong’s workflow for recording sample reviews.

We have not inspected your sample or production. This guide is not product testing, a legal or compliance opinion, supplier approval or production release.

PRIMARY SOURCES / 05

Official sources used for this guide

Source notes reviewed on 5 September 2026. Recheck applicable requirements for your product and destination.

COMMON QUESTIONS / 06

Common questions

What is a “golden sample”?
Teams often use the term for an approved physical comparison reference. The label alone is not a control: give the item a unique ID and revision, connect it to written criteria, record custody and state exactly what decision it supports.
Can the sample replace the product specification?
No. It can represent selected visible or tactile characteristics, but written requirements should control dimensions, tolerances, materials, hidden construction, performance and other measurable or destination-specific criteria.
Does a passing laboratory report approve the sample?
Only the named tests and results can support the decision, and only when the tested item is matched to the reviewed sample. The buyer must still assess the rest of the specification and destination requirements.
Can I approve with minor differences?
You can record a conditional state, but list every accepted difference, impact, owner and closure condition. Do not let conditional acceptance silently become production release.
What happens after a material, component or factory change?
Record the impact, supersede the affected baseline and require a new revision or reapproval before affected work proceeds.

BUILD THE HANDOFF

Connect the sample decision to the order and inspection record.

Choose the checks your order needs. Sending a request does not charge you or approve the supplier.

How we checked

Fill rates come from running our report process on real companies and counting how many of the twelve dimensions returned data on the date stated. The most recent query date on this page is 22 August 2026. Where a table carries its own date, that date governs.

Need supplier records? Choose the checks you need in the report menu, packs from $26.55. Delivery follows your order confirmation.

Payment next? Use the free pre-payment checklist.