On a large order, "did you check the quality?" cannot mean "did you inspect every chair." Pull and unpack 50,000 finished chairs to inspect each one and you would damage some, delay all of them, and still miss things because tired inspectors miss things. The grown-up answer is statistical sampling under AQL — the Acceptable Quality Limit — and it is worth understanding because it is also how your own third-party inspector will check us.
What AQL actually is
AQL sampling comes from ISO 2859-1, the same tables as ANSI/ASQ Z1.4. You pick a sample size based on the lot size and an inspection level — General Inspection Level II is the normal default — and the table gives you a sample quantity and an accept/reject number. Then you sort defects by severity. The widely used furniture spec is zero critical defects allowed, a major-defect AQL of 2.5, and a minor-defect AQL of 4.0. In plain terms, an AQL of 4.0 on minors means the lot passes if the number of minor defects in the sample stays at or below the table's limit — for a typical sample that is on the order of a dozen-odd — and fails above it. The point is a defined, agreed line, not a vague "looks fine."
Why sampling beats 100% inspection at volume
Sampling is not the cheap option pretending to be rigorous — it is the statistically sound one. A correctly sized sample tells you the lot's defect rate with known confidence. Trying to 100%-inspect a huge order introduces its own errors: inspector fatigue, handling damage, and a false sense that "we looked at everything." Where we do inspect every unit is function-critical points that a sample cannot cover safely — for example a quick load and swivel check on chairs, because a structural miss is a critical defect, not a cosmetic one. Cosmetic items — a stitch, a scuff, a label — go to AQL sampling.
The trade-off: tighter AQL costs money
You can ask for a tighter AQL — say 1.5 on majors instead of 2.5 — and we can build to it. But tightening the limit means more rework and more rejected lots, and that cost lands in the unit price. For a value retail line, AQL 2.5 / 4.0 is the sensible balance. For a premium contract chair carrying your brand into a corporate fit-out, a tighter major-defect limit is worth paying for. We will quote either; what we will not do is claim a tight AQL and then inspect to a loose one. If your buyer requires a report, decide early whether the inspection is in-line, at final random, or both, because that changes the timeline and the cost.
Critical, major, minor — sorting defects is the real work
The AQL number is only half the system; the other half is how you classify a defect, and that is where disputes actually happen. A critical defect is one that makes the chair unsafe or unusable — a cracked weld, a base that fails a load check. Zero are allowed; one critical finding fails the lot outright. A major defect is something a customer would likely reject or return: a wobbly armrest, a seized tilt, a wrong color. A minor defect is a cosmetic blemish a customer would probably accept — a faint scuff, a slightly loose thread. The fights between buyers and factories are almost always about classification, not counting: is that mark a major or a minor? We settle it before production by photographing limit samples for the borderline cases, so the inspector is checking against pictures both sides agreed on, not against an adjective.
Inspection levels and why they change the cost
ISO 2859-1 also lets you pick an inspection level. General Level II is the normal default and balances cost against confidence. You can go to Level III for a larger sample and tighter confidence, or to a reduced level once a supplier has a long clean record. A larger sample catches more but costs more inspector-hours and slows the gate before shipment. For a first order with a new factory, a larger sample is money well spent — you are buying information about whether the supplier can actually hold the spec. For the fifth repeat order from a line with a clean history, a reduced level is reasonable. We will run to whatever level your buyer specifies; what matters is agreeing it in writing before the lot is built, alongside the golden sample and payment milestones.
We build and test to BIFMA / EN methods and third-party lab testing can be arranged per order — that covers durability; AQL covers conformance of the production lot. Both matter on a big order. Tell us the quantity and your AQL spec through the export desk, read how QC folds into private-label runs on the OEM / ODM page, or see the chairs these checks run on across the product range and the office-chair category.