AQL Inspection: Pass Good Lots, Reject Bad Ones, Protect Your Margin
In short: AQL 2.5 (Acceptance Quality Limit) is the quality level at which a sampling table is indexed — not a license to ship 2.5% defective goods. On a 5,000-unit order at General Level II, inspect 200 units and accept 10 major defects; an 11th rejects the lot. One critical defect fails the lot regardless of the major count.
This guide covers the two-table lookup, the random carton pull, and the contract clauses that decide who pays when a lot fails. Worked examples use the published tables; the clauses reflect standard OEM purchase-order practice.
🛡️ Our Verdict: Passing an AQL check reduces risk; it does not guarantee zero defects. It replaces opinion with a repeatable rule, and the rule is only as good as the lot and the defect definitions locked before sampling.
“The inspection does not measure the lot. It tests whether the lot is good enough to accept under a rule you agreed to in advance.” — David Wu, Senior Apparel Production & QA Consultant, LeelineWear
Does AQL 2.5 Allow 2.5% Defects?
“AQL 2.5 means 2.5% of this shipment can be defective.” That sentence appears in supplier emails every week, and it is the most expensive misreading in apparel quality control.
AQL stands for Acceptance Quality Limit. It is a property of the sampling plan, not a tolerance applied to individual garments. The 2.5 refers to the quality level at which the ISO 2859-1 table is indexed. It decides which acceptance number (Ac) applies once the sample size is known.
So what does the report actually certify? Only that the number of major defects found in the sample did not exceed that acceptance number. It did not measure the true defect rate of the lot. Sampling is a probability tool, and a genuinely worse lot can still pass.
Small orders are where this bites hardest. On a 100-unit lot at General Level II, the plan calls for a 20-piece sample: one major defect is accepted, two reject the lot. That is the inspection standard actually being bought.
Never accept “AQL 2.5 passed” as a summary. Demand the sample size, the acceptance number, and the rejection number on the report. Our guide to defect classification covers which faults count as major.
How Do the Two Tables Work?
ISO 2859-1 uses two lookups. Table 1 converts lot size and inspection level into a code letter. Table 2 converts that code letter and the chosen AQL into a sample size, an acceptance number, and a rejection number.
The default plan for most apparel orders is single sampling, normal inspection, General Level II. The rows below show what it produces at AQL 2.5 for major defects.
| Lot size | Code at General II | Sample size | Ac at AQL 2.5 | Re at AQL 2.5 |
|---|---|---|---|---|
| 91–150 | F | 20 | 1 | 2 |
| 151–280 | G | 32 | 2 | 3 |
| 281–500 | H | 50 | 3 | 4 |
| 501–1,200 | J | 80 | 5 | 6 |
| 1,201–3,200 | K | 125 | 7 | 8 |
| 3,201–10,000 | L | 200 | 10 | 11 |
| 10,001–35,000 | M | 315 | 14 | 15 |
Critical defects use a separate zero-acceptance rule: Ac 0 and Re 1. That rule belongs in the purchase order, not in ISO 2859-1, which has no critical class of its own. Minor defects usually run at AQL 4.0. The full matrix is in ISO 2859-1; its US twin is ANSI/ASQ Z1.4.
Two rules prevent most lookup errors. Read the lot-size band first, never the sample size. If a cell shows an arrow, follow it: the plan moves to a different sample size.
The Six-Step Inspection
📋 Checklist: Confirm the approved sample, spec sheet, measurement tolerances, packing list, defect glossary, and prior CAPA history. Pack a ruler, calibrated scale, camera, carton map, and needle detector. Without written criteria, an inspection is a guess with a clipboard.
Step 1: Lock Down the Production Lot
An inspection should never start on a moving target. Open cartons, loose repacks, and workers taping boxes at the last minute all indicate a lot that is still changing. The standard assumes sampling from a defined, static lot; if the lot changes, the statistical result is compromised.
Verify the carton and unit count against the purchase order, then segregate the goods by PO, style, color, and size. Define the defect classes in writing before the first box is opened: critical, major, and minor, with zero acceptance for critical unless the contract says otherwise.
The pre-production inspection is where those definitions should be settled. Arriving at final inspection with an unsigned defect list leaves every judgment open to challenge.
Step 2: Select the Inspection Mode and Sampling Level
Use General Level II as the default for routine final inspections. Move to Level I only for suppliers with a long, stable record, and Level III when the order is high-risk or the process has just changed.
Special levels S-1 to S-4 are for destructive or expensive tests such as seam-strength pulls, zipper cycling, and carton drops. Keep single sampling as the default; double sampling is faster for experienced teams but harder to audit.
Switching rules matter. Start on normal inspection. Move to tightened after two rejections in five consecutive lots, and return to reduced only after sustained good quality.
⚠️ Safety First: Authorize destructive tests in writing before the inspection. Enforce safe carton-cutting, and apply ESD precautions only when inspecting electronics.
Step 3: Read the ISO 2859-1 Tables
Work through the lookup once, slowly, until the sequence is automatic. Start with Table 1 and find the lot-size band. For a 5,000-unit order, the row is 3,201 to 10,000. Move right to the General Level II column and take the code letter: L.
Now open Table 2 and find row L. Read one column right for the sample size: 200 units. Then trace across the row for each defect class. Critical uses zero acceptance, Ac 0 and Re 1. Major at AQL 2.5 gives Ac 10 and Re 11. Minor at AQL 4.0 gives Ac 14 and Re 15.
⚡ Power Move: Keep a printed AQL chart at the inspection table. A calculator or app works as a speed check, but the printed table is the authority when numbers are disputed.
Step 4: Pull a Random Sample
Convert the sample-size requirement into a physical pull plan. A 5,000-unit lot packed 50 per carton means 100 cartons; select 20 cartons and draw 10 units from each. Do not simply open four boxes.
Generate the carton numbers from a random list and spread the selection across pallet positions. Stratify by size and color when the SKU mix requires it. Record carton IDs, pallet zones, and unit counts so another auditor can reconstruct the same sample.
If the factory offers only front-row cartons, stop and reselect. The sample must be one you controlled, not one the floor prepared.
Step 5: Inspect and Classify Defects
Place the approved sample beside the workspace. Feel the garment first, then measure it, then compare it against the spec sheet and the defect glossary. Check zipper resistance, seam slippage, measurements, labels, and any regulated markings.
Record one code per defect on a tally sheet, one line per unit, with a photo linked to each code. Do not double-count the same defect on the same unit.
Classify consistently. Critical covers safety and regulatory failures: a needle fragment, a banned azo dye, a failed flammability test, an illegal drawstring on children’s wear.
Major covers functional failures: an open seam, a failed closure, a wrong measurement, a missing label. Minor covers cosmetic faults: a loose thread, a light mark, a slight print shift.
If the inspection covers electronics or hard goods, swap in product-specific defect libraries. The logic is the same; the defect list is not.
The garment quality control checklist should already define fabric, seam, and measurement faults; color fastness testing covers the shade and staining defects a tally sheet can miss.
Where a garment carries a performance claim, check UPF testing standards and the eco-friendly fabric certification before accepting the tag.
Step 6: Make the Disposition Decision
Total each defect class and compare it with the acceptance and rejection numbers from Step 3. In the 5,000-unit example, a result of 0 critical, 7 major, and 12 minor passes. A result of 1 critical, 11 major, or 15 minor fails the entire lot.
Possible dispositions, in order of escalation: accept, hold for corrective action, rework, run a 100% sort, split the lot, or reject. Record the decision, the time, and the sign-off before any goods move.
What Happens When a Lot Fails?
Nothing leaves the factory. Quarantine the lot away from passing stock and attach a written failure notice to the report. A second inspection is then booked under tightened sampling; rechecking a sorted lot on the original plan simply repeats the same risk.
The PO decides who absorbs the cost. It should state that the supplier pays for the sort and the second inspection, set a correction deadline, and carry documented delay costs. Without those terms, the negotiation starts after the launch date has already slipped.
The quality dispute SOP covers the paper trail. Once goods have shipped, after-sales support covers what still goes wrong.
Why Does One Critical Defect Reject the Whole Lot?
Critical defects are not weighed against major and minor counts the way a defect rate is. The reading is absolute: one critical fails the lot, however clean the rest of the sample looks. That strictness is contractual, so it belongs in writing instead of being settled by a floor argument.
The economics justify it. A major defect costs a rework; a critical defect costs a recall, a retailer chargeback, or a legal claim. Our CPSIA compliance guide covers the children’s-wear rules that make this tier non-negotiable in the US market.
Three Process Failures That Sink an Inspection
Most inspection failures come from process, not statistics. Three recur across apparel lines.
Arguments Over Defect Severity
Without visual standards, classification becomes a debate. Compare against the approved sample, point to the signed defect photo board, and escalate edge cases before the final tally.
Refusal of Destructive Testing
If a factory hides testing tools or refuses an authorized test, halt the audit, record the inspection as a failure, and do not downgrade the scope.
Lot Passes AQL but Customers Complain
Standard sampling can still release too many functional flaws on complex activewear. Tighten the inspection mode, add special levels for functional tests, and move high-risk SKUs to 100% inspection. The sportswear certification guide maps each performance claim to the right test method.
Take the Checklist and Clauses to the Factory
The six steps above condense to one printable sheet. Use it at the inspection table, and attach the five PO clauses before the deposit.
- 📋 Inspection Pack Checklist: documents, tools, sample-plan fields, defect-class sign-off, and report contents.
- 📝 PO Clause Template: approved-sample control, factory access, the AQL plan, failure remedies, and the balance-payment gate.
Open the printable AQL inspection pack — print it or save it as a PDF before your next final inspection.
Conclusion
AQL inspection is a two-table lookup wrapped in a control process: lock the lot, set the plan, read the tables, pull a random sample, classify defects, and record the decision. The rule protects the buyer only when the lot and its defect definitions are fixed before sampling begins.
For floor-level support, third-party inspection services apply the same steps to your order.
Frequently Asked Questions about AQL Inspection
1. What does AQL 2.5 actually mean?
AQL 2.5 is the quality level at which the sampling plan is indexed, not a 2.5% defect allowance. On a 200-piece sample it sets the acceptance number for major defects: 10 or fewer is accepted, 11 rejects the lot. The report describes the sample, not the true defect rate of the lot.
2. Is AQL the same thing as ISO 2859-1?
No. AQL is the quality level you choose, such as 2.5 for major defects. ISO 2859-1, with its US twin ANSI/ASQ Z1.4, supplies the sampling tables, code letters, and switching rules. The AQL figure has no meaning on its own; it only works inside one of those tables.
3. How many units are inspected on a 100-piece order?
At General Level II, a 91–150-unit lot uses code F, so the sample is 20 pieces. At AQL 2.5, one major defect is accepted and two reject the lot. A 100-unit drop is therefore inspected far less thoroughly than a 5,000-unit order, which takes a 200-piece sample.
4. What happens if a shipment fails an AQL inspection?
The entire lot fails, not just the sampled units. Quarantine the goods on the factory floor, issue the failure notice in writing, and require a 100% sort. Book a second check under tightened inspection before release, and never release a failed lot on the strength of the factory’s own recount.
5. Who pays for the rework and reinspection?
The factory should pay, but only if the purchase order says so. Negotiate the cost of the sort, the reinspection fee, and documented delay before the deposit. Without that clause the bill defaults to the buyer, and the argument usually arrives two weeks before launch.
6. Can a manufacturer refuse a random AQL inspection?
A factory can refuse entry if access was never written into the contract, and some use internal safety policies to keep auditors off the floor. Put factory access and third-party inspection rights into the purchase order before the deposit, and treat any refused inspection as a failed lot rather than a scheduling problem.
Related reading: Screening factories? Our guide to clothing manufacturers in China covers ten named options and the verification steps to run before an order.
Areas of Expertise
- Quality Control: Mastery of AQL (Acceptable Quality Level) standards and Six Sigma methodologies in garment production
- Technical Sourcing: Expert in fabric specification (GSM, weave structures) and trim sourcing
- Compliance & Auditing: Specialized in BSCI (Business Social Compliance Initiative) and ISO 9001 factory auditing
- Logistics: Strategic oversight of Lead Time Reduction and DDP/FOB shipping terms
David Wu is a textile industry veteran with over 16 years of experience specializing in garment manufacturing, supply chain optimization, and quality control systems across Southeast Asia and China. His career is defined by implementing rigorous AQL 2.5/4.0 inspection protocols for mid-to-large-scale private label brands. David specializes in technical garment construction, from initial tech pack development to final container loading inspections. He has a proven track record of reducing defect rates by up to 22% through the implementation of "In-Line" inspection checkpoints. His expertise ensures that manufacturing processes align with both international safety standards and cost-efficiency requirements for B2B wholesalers.