WORKWEAREU
Back to Blog
Quality & Compliance 2026-11-03 Denny Lin

AQL 2.5 Explained: What a Final Random Inspection Actually Catches — and What It Misses

Quick Answer

AQL 2.5 (Acceptable Quality Limit) is the standard sampling plan for garment final inspection: for a 500-piece order, an inspector checks 50 randomly selected pieces and rejects the lot if more than 3 carry major defects. "2.5" does not mean 2.5% defect-free — it defines the sampling math. The honest limitation: AQL inspection catches workmanship problems (stitching, sizing, labelling) but cannot verify what requires lab testing — waterproof ratings, flame resistance, retroreflectivity after washing. A passed inspection plus a test report is assurance; a passed inspection alone is a snapshot.

Open ten workwear factory websites and nine of them promise "AQL 2.5 inspection on every order". We make the same promise on ours. The difference between a promise and a system is whether anyone — buyer or factory — can explain what the number actually commits us to. Here is the explanation we give buyers who ask, including the part most factories leave out: what an AQL inspection cannot do.

What AQL 2.5 Actually Is

AQL stands for Acceptable Quality Limit — the worst average defect rate a buyer will tolerate across a production run. The working mechanics come from the sampling tables (ANSI/ASQ Z1.4, General Level II by default in garments):

- Order size sets the sample size. A 500-piece order draws 50 pieces; a 5,000-piece order draws 200. The inspector selects cartons and pieces at random — "at random" is the entire point.

  • Defects are classified three ways. Critical (safety or legal failures: a needle fragment in a garment, a missing compliance label) — AQL 0, one is too many. Major (functional or visible problems: broken stitching, wrong measurements, stained fabric) — AQL 2.5. Minor (cosmetic issues a wearer would not notice) — usually AQL 4.0.
  • The number is a threshold, not a percentage. At AQL 2.5 on a 50-piece sample, up to 3 major defects pass; 4 fail the lot. "2.5" describes where that accept/reject line sits in the tables — it does not mean the goods are "97.5% perfect".

    A failed lot does not go in the container. It goes back for 100% re-work or sorting, then re-inspection — which is exactly the outcome the system is designed to produce.

    The Part Factories Leave Out: What AQL Cannot Catch

    An inspector works with eyes, hands, a tape measure and a checklist. That is powerful for workmanship — and blind to performance.

    AQL inspection catches: stitching quality, seam slippage visible at the surface, measurement drift against the size chart, wrong or crooked labels, logo placement errors, colour mismatch against the approved swatch, packing and carton-marking errors.

    AQL inspection cannot catch: whether the rain jacket actually holds 5,000 mm of hydrostatic head (EN 343). Whether the reflective tape still performs after 50 wash cycles (EN ISO 20471). Whether the fabric is the FR fabric you specified or a lookalike. Whether the fluorescent dye meets the chromaticity coordinates of AS/NZS 1906.4. Every one of those requires instruments or a laboratory.

    This is why a certificate and an inspection report are not interchangeable documents. The lab report says the design and materials perform; the AQL report says the production run was built properly. Certified programs need both — a test report on production-intent samples, then AQL 2.5 on the bulk. A factory offering only the inspection is showing you half the picture.

    When to Inspect: The Timing Most Buyers Get Wrong

    The default — Final Random Inspection at 100% produced, 80% packed — is the latest acceptable moment, not the best one. By then the bad news arrives with the container already booked.

    For first orders, new styles, or any order where a mistake is expensive, the professional sequence is:

    1. DUPRO (During Production Check) at 20–60% completion. This is where systematic errors get caught while they are still cheap: the wrong seam-tape width applied to 800 jackets, a stitch density drifting off spec, a size label run printed wrong. Fixing it now costs hours; finding it at FRI costs weeks.

2. Final Random Inspection before loading. The statistical verdict on the finished goods, and the document your file needs.

Factory QC vs Third-Party: The Honest Trade-Off

We run our own AQL 2.5 final inspection on every order, with photos and defect lists in the report — and we still tell first-time buyers to consider third-party inspection (SGS, Bureau Veritas, QIMA, roughly US$300–600 per man-day). Not because factory QC is theatre, but because independence is the product: you are buying a report neither side controls.

The pattern that works over a relationship: third-party on the first order, factory report plus evidence on repeats, third-party again annually or whenever something changes — new fabric, new production line, new style. The pattern that fails, every time: no inspection because the pre-production samples were beautiful. Samples are made by the best hands in the building; bulk is made by lines. Inspect the lines' work.

What to Demand in the Report

A report worth filing contains: the sampling plan (order qty, sample size, AQL levels), the defect list with photos and classification, actual measurements against the size chart, photos of the random draw — cartons being opened, not a prepared table — and the verdict with corrective actions if failed. If your supplier's "inspection report" is a one-line pass stamp, ask for the defect list. The answer tells you whether an inspection happened at all.

We include the full breakdown as standard — sampling plan, defect photos, measurements, carton draw — on every order, ours or third-party witnessed. [Ask for a sample report](/#contact) and we will send a redacted real one, so you can judge the format before you judge the factory.

---

Fujian Tiancheng Holdings Garments I/E Corp., Ltd. is a workwear and PPE export company serving brands in the EU, US, Canada and Australia since 2000 — production runs on BSCI, SMETA and SCAN-audited partner factory lines in Fujian.

Frequently Asked Questions

What does AQL 2.5 actually mean on a garment order?+

AQL 2.5 sets the defect threshold for major defects in the sampling tables (ANSI/ASQ Z1.4, General Inspection Level II is the garment-industry default). For a 500-piece order the inspector draws 50 random pieces; the lot passes with up to 3 major defects and is rejected at 4. Minor defects usually run at AQL 4.0 (5 allowed on the same sample), critical defects at AQL 0 — a single critical defect (a needle fragment, a wrong safety label) fails the lot.

Is a passed AQL inspection a guarantee of quality?+

No — it is a statistical statement, not a guarantee. With 50 pieces checked out of 500, a lot carrying 4–5% defective units will still pass the inspection a meaningful share of the time. That is the designed trade-off: full 100% inspection costs several times more and still misses defects from human error. AQL’s real value is discipline — the random draw means the factory cannot stage a "golden sample" table, and the written report makes the result auditable.

What can AQL inspection NOT check?+

Anything that requires instruments or lab testing: waterproof hydrostatic head (EN 343), retroreflective performance (EN ISO 20471, AS/NZS 1906.4), flame resistance (EN ISO 11612), colourfastness ratings, fabric composition. An inspector can verify the correct labels are attached and workmanship is sound — they cannot verify the fabric performs to the certificate. That is why certified programs pair AQL inspection with lab test reports on production-intent samples; the two documents answer different questions.

When in production should the inspection happen?+

The standard is a Final Random Inspection (FRI) when 100% of the order is produced and at least 80% is packed — early enough to reject before shipping, late enough that the sample represents the real goods. For first orders or new styles, add a During Production Check (DUPRO) at 20–60% completion: it catches systematic problems (wrong stitch density, off-spec seam tape) while there is still time and fabric to fix them. An FRI alone tells you the bad news when it is nearly too late.

Should I hire a third-party inspector or trust the factory’s own QC?+

For a first order with a new factory, third-party (SGS, Bureau Veritas, QIMA — roughly US$300–600 per man-day) is cheap insurance: the report is independent and the factory knows it. On repeat orders with a proven factory, many buyers accept the factory’s own AQL 2.5 report plus photo/video evidence, bringing back third-party checks annually or when anything changes (new fabric, new line, new style). The pattern that fails: no inspection at all because "the samples were good."

What should an inspection report include?+

Five things: the sampling plan used (order quantity, sample size, AQL levels for critical/major/minor); the defect list with photos and classification; measurement results against the size chart (not just "checked"); on-site photos of the draw — cartons opened at random, not a prepared table; and the verdict with the factory’s corrective action if failed. A report that is only a pass/fail stamp with no defect photos is decoration, not documentation.

About Fujian Tiancheng Holdings Garments

Workwear & PPE export specialist since 2000. Production runs on BSCI, SMETA and SCAN-audited partner factory lines in Fujian — 1,000+ workers across the network. Serving brands in the EU, US, Canada and Australia.

Contact us for a quote →