An AQL inspection does not search an entire shipment for every defective unit. It uses a defined sample to make an accept-or-reject decision about a production lot.
That distinction matters. Buyers sometimes read “AQL 2.5” as permission for 2.5% defective goods. It is not. The result comes from a sampling plan, sample size, and acceptance numbers selected under a specified standard.
What does AQL mean in product inspection?
AQL is commonly expanded as acceptance quality limit in current sampling terminology. Buyers use an acceptance-sampling standard, such as ISO 2859-1 or ANSI/ASQ Z1.4, to connect:
- the presented lot size;
- inspection level;
- sampling plan and inspection severity;
- AQL selected for each defect class;
- sample size;
- acceptance and rejection numbers.
The inspector counts defects or defective units according to the agreed plan. The lot is accepted or rejected for that defect class based on the applicable numbers.
This is a decision procedure, not a prediction that the unsampled units contain the same defect percentage as the sample.
Start with a real, identifiable production lot
Sampling is meaningful only when the lot presented for inspection is the lot the buyer may release.
Before selecting cartons, confirm:
- purchase order and product revision;
- model, color, size, and other variants;
- total finished quantity;
- packing status;
- carton count and packing list;
- whether rejected, unfinished, or replacement stock is separated;
- whether the inspector can select across the available lot.
If production is incomplete, cartons are inaccessible, or the factory chooses every inspection carton, the mathematical plan may look correct while the physical sample is biased.
Choose the inspection level intentionally
General inspection levels determine how the lot size maps to a sample-size code. Level II is often used as a default for general workmanship, but “often used” is not a technical reason for every product.
A lower inspection level reduces the sample and the ability to detect problems. A higher level increases the sample, inspection time, and cost. The choice should reflect product history, process stability, order value, supplier performance, and the consequences of a missed defect.
Special inspection levels may be used for tests that are destructive, slow, costly, or impractical on the general sample. Examples might include a long cycle test, destructive seam test, or pack-out verification requiring substantial setup.
Record the level by test or defect group. Writing only “standard inspection” leaves room for different assumptions.
Classify defects before the inspector arrives
Defect classes describe commercial or safety significance. A practical scheme is:
- Critical: a condition that may be unsafe, unlawful, or fundamentally unacceptable for the intended use.
- Major: a defect likely to cause functional failure, return, significant customer dissatisfaction, or inability to sell as intended.
- Minor: a departure from the approved standard that does not materially reduce normal use.
The same visual issue can have a different classification depending on location and effect. A small mark hidden under a product may be minor; the same mark across a customer-facing logo can be major.
Create a defect list with photographs, diagrams, or boundary samples. Include known sample-stage problems. An inspector should not invent commercial severity on the factory floor.
Select AQL values by risk, not habit
Buyers frequently use different AQL values for critical, major, and minor defects. The exact values are a buyer decision informed by product risk, customer expectations, historical performance, and the sampling standard.
Do not copy values from an unrelated product. A cosmetic promotional item, food-contact component, children’s product, and load-bearing part do not have the same consequences of failure.
Some characteristics should not rely on ordinary AQL sampling at all. Safety, legal, regulatory, or high-consequence functions may require 100% production controls, validated process controls, accredited laboratory testing, traceability, or another risk-specific method. “Critical defects: 0” in a small sample is not proof that none exist in the lot.
How the sample-size decision works
The normal workflow is:
- determine the lot size;
- choose the inspection level;
- use the standard’s table to obtain a code letter;
- use the code letter and plan to find the sample size;
- use each selected AQL to find acceptance and rejection numbers;
- inspect the randomly selected sample;
- compare observed counts with the applicable decision numbers.
Do not build a permanent web table from memory or an unlicensed screenshot of a standard. The controlling standard, revision, and plan should be available to the qualified inspection provider using it.
Count defects consistently
Decide whether the plan counts defects or defective units, and train everyone on the defect list. One unit can contain more than one defect. How those observations feed the decision must match the stated procedure.
The report should show:
- sample size and selection method;
- inspected quantity by model or variant;
- each defect description and class;
- defect count and affected-unit count where useful;
- photographs linked to the finding;
- acceptance and rejection numbers used;
- result by defect class;
- non-sampling tests and their results separately.
Avoid a report that shows only “pass.” The counts are useful for trend analysis even when the lot is accepted.
AQL does not cover every inspection task
A pre-shipment inspection usually contains more than sampled workmanship.
Quantity reconciliation, shipping-mark review, document checks, carton assortment, measurements, function tests, barcode scans, packaging checks, and product-specific tests can use different sample quantities or 100% verification of available records.
For example, the inspector might use a general AQL sample for visual workmanship, a smaller special-level sample for a destructive test, and every selected carton for assortment. The inspection instruction should keep those populations clear.
Build the complete scope with the pre-shipment inspection guide, not from an AQL table alone.
What happens when an AQL inspection fails?
A failed result means the presented lot did not meet the agreed sampling decision rule. The normal response is not to delete a few photographed defects and keep the original pass/fail result.
Ask the supplier to:
- contain the affected lot;
- investigate the process and extent of the problem;
- sort, rework, or replace product under a documented instruction;
- record corrected and rejected quantities;
- address the root cause where practical;
- present the corrected lot for a new random inspection when required.
The buyer then decides whether to reinspect, accept a documented concession, hold shipment, or reject the lot under the applicable agreement.
Common AQL mistakes buyers can prevent
Treating AQL as the maximum defect percentage
The selected AQL helps determine sampling acceptance numbers. It is not a promise that the shipment contains no more than that percentage of defects.
Choosing the sample after seeing the goods
Changing the level or plan because the lot “looks good” undermines the pre-agreed decision rule.
Letting the factory prepare the sample
The inspector should control random selection across the accessible lot. Convenience samples can hide variation between lines, shifts, or carton areas.
Using vague defect categories
If buyers, factories, and inspectors classify the same mark differently, the acceptance numbers will not produce a stable decision.
Using AQL for compliance claims
Visual sampling does not replace destination-market research, approved construction, qualified testing, or production controls for regulated characteristics. Use the product compliance research checklist to separate these questions.
Frequently asked questions
How many units should be inspected for a China order?
The sample depends on the lot size, inspection level, sampling plan, and applicable standard. There is no fixed percentage that works across all order sizes. Ask the inspection provider to state the standard, revision, code, level, sample size, and decision numbers in the booking confirmation.
What AQL should I use for consumer products?
There is no universal consumer-product setting. Choose defect classes and values from the product’s use, failure consequences, customer expectations, supplier history, and other controls. Safety or regulatory characteristics may need methods outside normal AQL sampling.
Does passing AQL mean the shipment has no defects?
No. It means the inspected sample supported acceptance of the lot under the stated sampling rule. Unsampled units remain unsampled.
Can I change AQL after an inspection fails?
Changing the decision rule after seeing the result destroys the value of a pre-agreed acceptance plan. If the buyer grants a commercial concession, document it as a concession rather than rewriting the inspection result.
Write the sampling plan into the order
The purchase order or inspection specification should name the standard, revision, inspection level, AQL values by defect class, defect definitions, lot-readiness rule, selection authority, and failure process.
Connect those controls to the China purchase order checklist and preserve the approved comparison standard through the product sample process.
Sampling plans require competent application. Product safety, regulatory compliance, and contractual acceptance may require controls beyond AQL inspection; obtain qualified technical and legal advice where appropriate.