ZIABRIDGE
Quality You Can Trust. Global Solutions You Can Rely On.
Reference illustration — will be replaced with a real photo.
Checking every unit in a shipment of, say, 10,000 garments would take too long and cost too much to be practical. AQL (Acceptable Quality Limit) sampling solves this: an inspector checks a much smaller, randomly selected sample, and uses the result to make a statistically sound decision about the whole shipment.
It's important to understand what AQL is not: it isn't a defect rate the buyer agrees to tolerate in production. It's a decision rule applied to a sample, used to judge whether the full lot is likely good enough to ship.
A worked example: for a lot of 1,200 garments at General Inspection Level II, the sample size is commonly 80 pieces. At AQL 2.5 for major defects, that might mean an accept number of 5 and a reject number of 6 — find 5 or fewer major defects in the sample of 80, and the whole lot passes; find 6 or more, and it's rejected. The exact numbers depend on the lot size and AQL level chosen, so always confirm against the current reference table rather than memorising a single set of numbers.
The sample has to be pulled randomly by the inspector, not handed over by the factory. A sample chosen by the factory tends to show only the best-looking pieces, which defeats the entire purpose of sampling — the result is only meaningful if it reflects the real condition of the whole lot.
This is the most common practical question buyers have, and the honest answer is: it depends on how much risk the product can tolerate. As a general starting point that many buyers use:
Common for higher-value items, children's wear, or anything where appearance and safety tolerance is low. Fewer defects allowed per sample.
The most widely used level for everyday garments — t-shirts, casualwear, basics. A reasonable balance between cost and strictness.
Sometimes used for minor-defect tolerance on lower-cost, high-volume basics where small cosmetic issues matter less.
These are common starting points, not fixed rules — many buyers set a stricter level for critical/major defects while allowing a slightly more relaxed level for minor ones on the same order. The right choice depends on the product's price point, end use, and your own brand's tolerance for returns.
AQL sampling gives a pass/fail answer for one shipment. But factories also track quality continuously, on the production floor, using a different metric: DHU (Defects per Hundred Units).
DHU = (Total defects found ÷ Total garments checked) × 100
For example, if a checker inspects 250 garments in a shift and finds 35 defects in total (some garments may have more than one), the DHU is (35 ÷ 250) × 100 = 14. A lower DHU means better running quality. Unlike AQL, DHU isn't a pass/fail gate on a single shipment — it's a day-to-day performance number, often tracked by line, by style, or by buyer.
OQL (Observed Quality Level) is a closely related idea: the actual defect rate observed in a sample, expressed as a plain ratio rather than a pass/fail result. The point of OQL is that two suppliers can both "pass" the same AQL inspection while having very different real defect rates — OQL surfaces that difference, which a simple pass/fail doesn't.