Every factory has a house standard for what passes. None of them is your standard. When a range is produced across a seamless mill, a cut-and-sew line and a swim specialist, the customer receives one brand — and notices immediately that it was inspected by three different sets of eyes. This is how to impose a single standard across factories that do not share one.
The short answer
- Each factory applies its own house standard; unless a single specification is imposed, a mixed range arrives to several definitions of acceptable.
- AQL 2.5 defines sample size and accept/reject numbers for a lot — it is a sampling plan, not a description of quality.
- The decisive question is who employs the inspector: a factory inspecting its own output is performing self-assessment.
- Cross-category consistency — color, logo, labeling, packaging — fails more often than individual product quality.
What AQL 2.5 actually specifies
AQL is a sampling plan, not a quality level in the everyday sense. For a given lot size it fixes how many units are drawn and the maximum number of defects at which the lot is still accepted. AQL 2.5 for major defects is the common commercial benchmark in apparel; tighter numbers exist and cost more time.
Two things it does not do. It does not define what counts as a defect — that comes from your specification and an approved reference sample. And it says nothing about whether the inspection actually happened as described.
So the useful question is never “do you inspect at AQL 2.5”. Everyone says yes. It is: against which approved reference, with what defect classification, at what production stage, and who signs the report.
Why mixed ranges fail on consistency, not defects
In a multi-factory range, the individual products are usually fine. What fails is agreement between them. The same black reads warmer on the bra than on the leggings because two dye houses matched to two different references. The logo sits three millimetres lower on the swim piece. Care labels are formatted three ways because nobody sent one template.
None of that is a defect at any single factory — each one made exactly what it was asked to make. It becomes a defect only at the level of the range, which is the only level your customer experiences.
That is why cross-category items need to be specified centrally and inspected against one reference: color, logo placement and size, label and packaging format, hangtag position, polybag spec. Write them once, apply them to every factory, and inspect them as a range rather than as four separate orders.
Who the inspector works for
In many arrangements the producing factory inspects its own output and issues the report. It may be done competently. It is still a self-assessment, and it fails in a predictable place: when a marginal lot has to ship this week to hold a vessel booking, the party deciding whether it passes is the party that loses money if it does not.
Buyer-side inspection changes the incentive. The inspector is employed by the party representing your interest, applies your specification rather than the factory's house norm, and reports findings you may not want to hear before goods leave the building.
This is worth asking about in plain terms: who employs the inspector, and who decides whether goods ship after a failed inspection. Our own final inspections are carried out by our QC team on the buyer's behalf, at AQL 2.5, and a failed lot does not become a shipped lot because a date is tight.
The documents that make a standard portable
A standard only travels between factories if it exists on paper. Three documents do most of the work. The sealed reference sample: one physical approved standard per style, with the same reference used at every factory producing that style's components. The specification sheet: measurements with tolerances, construction notes, defect classification. And the inspection report format: the same template from every factory so results are comparable.
Tolerances matter more than most buyers expect. A measurement chart without tolerances invites each factory to interpret its own acceptable range, and interpretations differ by more than the tolerance would have.
Agree these before sampling, not after the first inspection dispute. Once bulk exists, every conversation about standards becomes a negotiation about who pays.
Where to inspect, and when
Final random inspection at 100% production and at least 80% packed is the standard checkpoint, but for mixed ranges it is often too late to be useful — a failure found there delays the whole consolidated shipment, not just the failing category.
In-line checks matter more here. A first-output check at the start of a run catches a systematic error while it affects a hundred units rather than the full order. For color-critical or fit-critical categories, an additional check when the first sealed pieces come off the line is cheap insurance.
The practical rule for multi-category programs: inspect the earliest-produced category early enough that a problem can still be fixed without moving the launch date for everything else.
Quick answers
Is AQL 2.5 strict enough for activewear?
It is the normal commercial benchmark for major defects and is adequate for most programs, provided the defect classification and reference sample are specific. For fit-critical or compliance-critical items, tighten the classification rather than the AQL number — defining the defect precisely does more than sampling more units.
Can I inspect a mixed-category order in one place?
Only if goods are physically consolidated before inspection, which is a decision to make at planning stage. Otherwise each category is inspected at its producing factory against the same specification, with the range-level items — color, logo, labeling — checked against one shared reference.