When full inspection beats sampling
Full inspection beats sampling when the cost of one escaped defect exceeds the cost of checking every unit. Safety implications, regulatory exposure, high unit value, and unknown supplier history all push toward checking everything, while stable low-risk lots with good track records belong to sampling. Buyers who compare escape cost against checking cost make this call correctly. The framework below turns that comparison into a repeatable decision.
Key takeawaysSampling bounds the risk of accepting a bad lot but never proves every unit is good, which is why method choice starts from escape cost. · Full inspection wins where single defects carry safety, regulatory, or key-account consequences that dwarf checking cost. · Stable low-risk lots with good history are the natural territory of sampling, not of checking every unit. · Whichever method is chosen, defined responsibilities and records decide whether findings drive improvement. [1] · Test evidence cited in reports should rest with demonstrably competent laboratories under mutual recognition. [2]
Sampling inspects a fraction of the lot and full inspection checks every unit, and each method wins in different situations. Sampling is fast and economical for stable processes with low defect rates. Full inspection, painful as it sounds, is the rational answer where the cost of one escaped defect dwarfs the cost of checking everything.
The wrong choice wastes money in both directions. Sampling a lot that should have been fully checked lets dangerous defects reach customers. Fully checking a lot that sampling could have judged doubles quality cost for no additional protection. The decision deserves a structured comparison rather than habit.
This article sets out that comparison: what sampling can and cannot prove, the situations where full inspection wins, the economics of each, and how to run whichever method is chosen without creating new problems.
What sampling proves and what it cannot
A sampling verdict is a risk statement about the lot, not a certificate for each unit. When a sample passes agreed limits, the buyer learns that the lot is unlikely to be worse than a defined quality level, with the residual risk set by the sample size and limits chosen. That is genuinely useful information, and for stable processes it is all the protection needed.
What sampling cannot do is catch isolated defects scattered thinly through a large lot, nor can it judge characteristics hidden inside every unit without destructive testing of each one. A passed sample and a perfect lot are different things, and confusing them leads to blaming the method for what was actually a method-selection error.
This is why sampling works best with history. A supplier with several clean lots behind it has demonstrated process stability, so the sample confirms continuation rather than discovering the unknown. A new supplier, a new product, or a lot produced under rush conditions lacks that foundation, and sampling alone may leave too much risk uncovered.
Situations where checking every unit wins
The clearest case is safety. Where a defect could injure users or trigger regulatory action, letting even a small fraction escape is unacceptable, and only checking each unit reduces that risk to the practical minimum. The same logic covers high-value goods where one replacement shipment costs more than the full check, and orders for key accounts where a single failure endangers the relationship.
The second case is unknown quality. First lots from new suppliers, lots after material or process changes, and lots with a recent failure history all carry uncertainty that sampling cannot resolve. A full check or a full sort converts that uncertainty into known units, each individually judged. Buyers often procure this as third-party product inspection in China third-party product inspection in China, specifying full-check scope for the risky lot while keeping sampling for routine ones.
The third case is containment after a failure. When a sample fails and the lot must still ship on time, sorting the full lot under supervision separates good units from bad ones. Sorting is containment rather than improvement, but it rescues schedules that remake cannot meet.
Cost and risk comparison
| Factor | Sampling | Full inspection or sort |
|---|---|---|
| Direct cost | One inspector-day typically covers several lots | Multiple inspector-days scaled to lot size |
| Speed | Verdict usually same day | Days to weeks depending on quantity |
| Escape risk | Bounded but nonzero residual risk | Minimal for checked characteristics |
| Best lot profile | Stable supplier, low-risk product, good history | New supplier, safety relevance, failure history |
| Process insight | Confirms stability, weak on root causes | Finds every visible defect but still not the cause |
| Follow-up need | Corrective action on failures | Corrective action plus method review for next lots |
Running a full check without creating new problems
- Define the check per unitList every characteristic to be checked on each unit with measurable pass and fail thresholds.
- Protect goods during handlingConfirm handling steps that prevent inspection itself from scratching, soiling, or damaging goods.
- Tally every unitRequire countersigned tallies of checked, passed, and failed units so the reconciliation is auditable.
- Segregate failures at onceMark or segregate failed units immediately so they cannot drift back into the shippable pile.
- Record under disciplineApply the same documented discipline of responsibilities and records that mature quality systems require of any control. [1]
Do not let full inspection become a permanent substitute for process control. If every lot needs checking unit by unit, the process is not capable and the checking cost repeats forever. Use the defect data from the full check to drive the corrective action that returns future lots to sampling.
Weighing test evidence behind either verdict
Inspection verdicts sometimes rest partly on laboratory results, for example for restricted substances or performance claims. Those results deserve the same scrutiny as the visual check: which laboratory ran the test, whether its accreditation covers the method, and whether samples were sealed and traceable. Mutual recognition of accreditation across borders is organized through the international laboratory accreditation cooperation framework. [2]
Evidence from outside recognized accreditation is not automatically wrong, but it carries less weight and may need confirmation before it supports a shipping decision. Build this evidence check into the inspection scope for product categories where lab results decide pass or fail.
Sources
- Wikipedia: ISO 9000 —— Overview of the ISO 9000 family of quality management standards.
- ILAC: International Laboratory Accreditation Cooperation —— ILAC homepage describing the international cooperation on laboratory and inspection body accreditation.
Frequently asked questions
How do buyers calculate whether full inspection pays off?
Compare the cost of the check against the cost of one defective unit reaching the customer, multiplied by the expected defect frequency. Where a single escape means a safety incident, a major recall, or losing a key account, full inspection usually pays. Where escapes mean minor rework, sampling is normally the economical choice.
Does sorting a lot fix the underlying quality problem?
Yes. Sorting identifies the defective units so the lot can ship, but it does not fix the process that produced them. Without corrective action at the source, the next lot needs sorting again. Use sorting as containment and process work as the cure.
What management discipline matters regardless of method?
Documented procedures with clear responsibilities for checking, recording, and segregating nonconforming product, applied consistently regardless of method. A full check run without records is barely better than no check.
What does accreditation have to do with inspection reports?
It means the test findings carry weight only where the testing competence is demonstrated, which is why buyers check the accreditation behind cited results. The global accreditation cooperation framework explains how laboratory competence is recognized across borders. Test evidence from outside that system needs extra scrutiny before it supports a shipping decision.