Sampling and test flow in food contact testing

Key takeaways

1. Random selection across the packed lot is the step that lets the result speak about the shipment rather than about a showcase sample.
2. Sealed samples with batch linkage and a written sampling record preserve identity from warehouse to laboratory bench.
3. Laboratory conditions must mirror the harshest foreseeable use so the verdict reflects real kitchens. ^[2]
4. Acceptance sampling logic turns the laboratory verdict into a lot decision with understood residual risk. ^[1]
5. Read the verdict as a bounded statement about the lot and confirm corrections with retests under identical conditions.

Buyers spend heavily on laboratory testing and surprisingly little on the sampling that gives those tests meaning. The laboratory can only describe what arrived in its receiving room; everything about which goods those pieces represent was decided earlier, in the warehouse, by whoever chose the cartons. A perfect test on hand-picked samples proves the factory can make one good box.

Professional sampling closes that gap with unglamorous discipline: random numbers against carton counts, selection by an independent party, sealed bags with batch codes, photographs, and a written record signed at the scene. None of this requires expensive equipment. All of it requires planning before the visit.

The flow below runs end to end: preparing the draw, executing it in the warehouse, protecting samples in transit, what happens at the laboratory, and how the verdict converts into a shipping decision.

Prepare the draw before the visit

Sampling starts on paper. Confirm the total lot quantity and carton count, decide how many cartons to open and how many pieces to take, and define which variants must appear in the set: every colour family, the most decorated version, each size extreme. Write the plan down and share it with the factory so the visit runs on procedure rather than negotiation.

Bring the materials that make integrity visible: tamper-evident bags, labels, a camera, and a sampling form capturing purchase order, SKU, batch codes, carton numbers, quantities, dates, and signatures. Integrity that is not documented at the moment cannot be reconstructed later, and a laboratory receiving anonymous pieces can only test orphans.

Schedule the draw when the lot is genuinely complete and packed. Sampling a half-finished lot selects from the easy early production and misses the rushed tail where defects cluster. Completeness of the lot is a precondition of the draw, not a courtesy to the factory.

Execute random selection in the warehouse

1. Verify the lot is completeCount packed cartons against the declared quantity and confirm production has finished before opening anything.
2. Generate carton numbers randomlyUse random numbers across the full carton range rather than accepting factory-suggested cartons or convenient top-layer boxes.
3. Open and draw across variantsTake pieces from the selected cartons covering all required variants, noting any carton that cannot be opened or any variant that is missing.
4. Seal and label immediatelyBag, seal, and label each sample with SKU, batch, carton number, and date at the sampling point, and photograph the sealed set.
5. Record and signComplete the sampling form with all counts and codes, have both sides sign, and keep custody of the samples or seal them for courier.
6. Retain referee piecesHold back sealed duplicates under stated conditions so any later challenge can retest the same lot rather than new production.

Protect identity in transit

Samples change hands between the warehouse and the bench, and every handover is a chance for mix-ups. Sealed labelling with batch codes, a courier record naming the seals, and prompt dispatch keep the chain short and provable. Perishable or condition-sensitive articles need the transport conditions noted as well, since heat or crushing in transit can alter what the laboratory receives.

Brief the laboratory in writing: the article description, the material, the foreseeable contact conditions, the standard or limits to test against, and the contact for technical questions. A laboratory told only to run a generic migration test may choose conditions that fit neither the product nor the destination market. Precise instructions are part of sampling, not an optional extra.

Food contact testing food contact testing commissioned this way arrives at the bench as identified evidence rather than as anonymous objects. That identity is what later lets the buyer, the supplier, and if necessary a second laboratory argue about the same goods instead of talking past each other.

What the laboratory does with the samples

At the bench the samples meet defined simulants, times, and temperatures chosen to mimic the harshest foreseeable contact, and the laboratory measures what migrates under those conditions. These conditions are not arbitrary severity; they translate the product's real use into reproducible settings, which is why the buyer's briefing about contact conditions matters so much.

Technical regulations of the destination market set the limits the measurements are judged against, and such measures apply across origins under the international framework governing technical barriers to trade. ^[2] The laboratory's report therefore states method, conditions, limits, values, and uncertainty together; each element without the others loses meaning.

Turnaround runs from days to weeks depending on the program, and long-duration storage tests extend further. Build that calendar into the order plan from the start rather than discovering it when the container is booked.

Turn the verdict into a lot decision

The verdict describes the tested pieces; acceptance sampling logic extends it to the lot with understood risk. The theory shows how sample size and acceptance numbers bound the chance of shipping a poor lot or rejecting a good one, as summarized in the standard overview of acceptance sampling. ^[1] A pass on a sound random sample supports shipment, while a pass on a compromised sample supports very little regardless of how clean the numbers look.

Read failures as lot statements too: the lot as drawn does not meet the requirement, which triggers containment of the affected portion, diagnosis, correction, and retesting under identical conditions. Retest samples follow the same draw discipline, because a correction proven on hand-picked pieces proves nothing about production.

File the sampling record next to the report in the SKU dossier. Months later, when a customer or authority asks what a report covers, the record answers in one page. Reports without sampling records age into orphans that defend nothing.

Frequently asked questions