AQL Sampling Inspection

AQL Sampling Inspection Explained for Importers Working With Chinese Suppliers

Published June 2026 · Updated September 2026 · By SSTI Inspection Team

AQL — acceptable quality level — is the statistical framework most widely used in pre-shipment inspection to determine how many units to inspect from a production batch and what defect rate signals a failing result. When buyers receive an inspection report that references AQL 2.5, Inspection Level II, or a pass at a specific sample size, understanding what those figures mean in practical terms allows the report to be used as a genuine decision-making tool rather than a stamp of approval.

AQL inspection is not complicated in its core logic, but its value depends on buyers understanding what it measures, what it does not measure, and how to set it up correctly before the inspection visit takes place. This guide explains the key concepts in practical terms for importers who want to use AQL-based inspection effectively in their sourcing from China.

The Basic Logic: Sampling Instead of Checking Every Unit

Inspecting every unit in a large production batch — thousands of garments, electronic components, or assembled goods — is not feasible within the cost and time constraints of a standard pre-shipment inspection. AQL sampling addresses this by inspecting a statistically representative subset of the batch and using the results to draw a conclusion about the quality level of the full quantity.

The underlying principle is straightforward: if a carefully drawn sample from a batch shows a defect rate below a defined threshold, it is statistically reasonable to conclude that the full batch is at or below that defect level. If the sample shows a defect rate above the threshold, the batch fails. The AQL number itself defines that threshold — the maximum percentage of defective units that is considered acceptable in the shipment.

AQL 2.5, for example, means that up to 2.5% of units in the batch may contain the relevant defect type while still being accepted. AQL 1.0 applies a stricter threshold — only 1.0% defective is acceptable before the batch fails. Lower AQL numbers represent stricter standards. The appropriate AQL for a given product depends on the product category, the consequences of defects reaching the end customer, and the buyer's quality requirements.

How Sample Size Is Determined

The number of units inspected is not chosen arbitrarily. It is determined by the batch size and the inspection level, using the reference standard ISO 2859-1 (formerly used as MIL-STD-105E in some markets). This standard produces a lookup table: given the total quantity in the batch and the chosen inspection level, a code letter is assigned, and that code letter determines the sample size.

Inspection levels define how intensively the batch is sampled relative to its size. General Inspection Level II is the standard used for most pre-shipment inspections. Level I uses a smaller sample and is appropriate for situations where less discrimination is needed or cost constraints are tight. Level III uses a larger sample and provides more confidence in the result, but at higher cost and time. Special inspection levels are used for specific types of testing that require a different sampling approach.

As an example: for a batch of 2,000 units inspected at General Inspection Level II, the sample size is 125 units. For a batch of 10,000 units at the same inspection level, the sample size is 200 units. Note that the sample size does not increase proportionally with batch size — the statistical relationship between sample and batch becomes more efficient as batch size grows. This is why AQL-based inspection is particularly suitable for large-volume orders.

The sample should be drawn randomly from the full batch — selecting units from different cartons, different positions in the storage area, and different points in the production run where identifiable. A sample drawn only from the most accessible cartons or from goods at the top of the pallet does not represent the full batch and reduces the reliability of the inspection result.

Why the Same Batch Passes at One AQL and Fails at Another

AQL is not applied once. It is applied separately to each defect category, because not all defects carry the same consequence for the buyer or the end customer.

Critical defects — safety hazards, regulatory violations, products unfit for their intended use — are normally held to zero tolerance: one in the sample fails the batch. Major defects, the ones buyers and their customers are likely to reject, are commonly inspected at AQL 2.5. Minor defects, cosmetic or dimensional departures unlikely to trigger a customer complaint, usually sit at AQL 4.0. A batch can therefore pass on minors and fail on majors in the same visit.

Which specific defects fall into each category for your product should be defined by you before the inspection, not left to the inspector's general judgment. That decision does more to determine the result than the AQL number does — see how defect classification works for who decides, how disputes with suppliers get resolved, and how to read the defect summary in a report.

What AQL Inspection Does Not Tell You

Understanding the limitations of AQL inspection is as important as understanding what it confirms. A passed AQL result does not mean that every unit in the batch is defect-free. It means that the defect rate found in the inspected sample is below the acceptable threshold defined by the chosen AQL level. The full batch will contain some defective units — the question is whether their proportion is acceptable.

Statistical sampling carries inherent uncertainty. The sample is not the same as the population. A batch that genuinely has a 3% defect rate will occasionally produce a passing sample result at AQL 2.5, and a batch with a 1.5% defect rate will occasionally produce a failing sample result. These are statistical realities of sampling, not inspection failures. The sample size and inspection level together determine how reliably the sample reflects the true batch quality level.

AQL inspection also does not detect defects that are not on the inspection checklist. If a buyer has not specified that a particular dimension, functional criterion, or labelling element should be checked, the inspector will not check it. The inspection covers what was defined in the scope — no more. Buyers who provide detailed, product-specific inspection checklists get inspection results that are relevant to their actual requirements. Buyers who rely on a generic checklist receive results that may miss the most important criteria for their specific product.

Defining AQL Criteria Before the Inspection Visit

The most common reason AQL inspection results are less useful than they should be is that the criteria were not defined in enough detail before the visit. When defect classification is left to general judgment, when the inspection checklist is not product-specific, or when the AQL levels are not confirmed in advance, the result may be technically valid but not aligned with the buyer's actual requirements.

Before scheduling an inspection, buyers should confirm the inspection level to be used (normally Level II for standard pre-shipment inspection), the AQL values for critical, major, and minor defects, the specific defects that fall into each category for the product, any checkpoints that are specific to the buyer's requirements and not covered by a generic checklist, and whether any items should be checked at 100% — such as barcodes or critical labelling that cannot be assessed by sampling alone.

Sharing the inspection brief with the factory before the visit is generally advisable. It does not compromise the independence of the inspection — the inspector still draws a random sample and applies objective criteria. It does give the factory a clear picture of what is being checked, which tends to improve the preparation of goods for inspection and reduces avoidable failures on criteria the factory was simply not aware of.

When 100% Inspection Is Worth It Instead

Sampling is the right default for most orders, but there are situations where the buyer needs unit-level results rather than a batch-level judgment.

The strongest case is post-rework verification. If a batch failed an AQL inspection and the factory has reworked the defective units, sampling the reworked batch can produce a misleading picture — reworked units are often distributed unevenly, concentrated in specific cartons or production positions. Checking every unit is what confirms they were actually corrected.

Small-to-medium batches are natural candidates simply because the arithmetic changes: for 200 or 300 units, the time difference between checking every unit and checking an AQL sample may be marginal, while the certainty is substantially higher. High-value products where a single defective unit reaching the end customer carries significant financial or regulatory consequence can justify the cost — as can a situation where customer complaints on previous shipments mean the buyer needs the actual defect rate across current stock, not an estimate of it.

Products whose defects are clearly visible in a brief visual check lend themselves to full inspection because per-unit time is low and the sorting benefit is direct. Goods requiring extensive individual handling or functional testing do not, at any meaningful scale.

The Cost, Time, and Risk Tradeoff

Full inspection costs more and takes longer, and the gap widens with batch size: for 500 units the difference may be modest, for 3,000 units it can be several times the cost of AQL sampling. The time matters too, if inspection sits on the critical path between production completion and the vessel loading date. How this translates into a quote is covered in what third-party inspection costs in China.

There is a middle option most buyers underuse: apply AQL sampling as the primary method, but check 100% of the highest-consequence criteria within the sampled units. Scan every sampled unit's barcode; verify label content on every sampled unit; apply AQL criteria to workmanship and appearance. This buys targeted completeness where errors are disproportionately expensive without paying for full inspection across the batch.

Using AQL Results to Make the Shipment Decision

A passed AQL inspection provides documented, quantified evidence that the batch quality level is within the acceptable threshold the buyer defined. This is the factual basis for approving shipment. A failed result provides the same type of documented evidence for the opposite conclusion — with specific defect counts, categories, and photographs that the buyer can reference when communicating with the factory about corrective action.

When results are close to the threshold — where the number of defects found in the sample is at or just above the acceptance number — the inspection report should describe the nature of the defects in enough detail for the buyer to make a judgment about proportionality. A batch that fails at AQL 2.5 because of three minor surface marks may warrant a different response than one that fails because of systematic functional defects across the sample.

Repeat inspection results across multiple orders from the same supplier allow buyers to track quality trends over time. A supplier who consistently passes at AQL 2.5 with low defect counts provides a different quality assurance picture than one who repeatedly passes at the threshold, or fails and reworks before passing on a second inspection. This longitudinal view is one of the most practical benefits of maintaining a consistent inspection programme across repeat orders.

For information on how SSTI applies AQL sampling in pre-shipment inspection and production monitoring visits, see the Services page. Buyers who want to discuss the appropriate inspection criteria for a specific product can contact SSTI through the Contact page.

Where AQL Results Become a Quality Record

A single AQL result is a shipment decision. A series of them is something more useful: a record of how a supplier actually performs. A factory that passes at AQL 2.5 with two or three defects in the sample, order after order, is a different proposition from one that passes at the acceptance number every time, or that fails and reworks before passing on a second visit. Neither pattern is visible from any single report — which is the practical argument for keeping inspection criteria consistent across orders rather than adjusting them each time. What the report itself should contain to support that comparison is set out in what a pre-shipment inspection report should show, and if a result comes back as a fail, the options are in what to do when an inspection fails.

Back to Insights