Acceptance Threshold
Statistical sampling schemes designate the maximum percent defective or the maximum number of defects per hundred units that can be considered satisfactory as a process average during batch inspection. When applied to finished garment export consignments, aql functions as the formal dividing line between an accepted delivery and a rejected production lot under standards such as ISO 2859-1 or ANSI/ASQ Z1.4. The parameter establishes the producer risk by defining how frequently lots of acceptable baseline quality will pass verification.
It ceases to guarantee zero defects across uninspected units, as batches passing an inspection schedule still harbour nonconforming garments within the defined probabilistic tolerance.
Operating Dynamic
Inspection protocols map the lot size against defined inspection levels, typically General Inspection Level II, to yield a specific sample code letter. This code letter dictates the sample size pulled randomly from finished packing cartons across colourways and size distributions. Evaluators count nonconformities across minor, major and critical defect categories, checking the tally against matched acceptance and rejection numbers.
When the count of major stitching breaks or dimension drifts reaches the rejection threshold, the lot fails unconditionally. Factory sorting and total re-inspection become mandatory before presenting the goods for a secondary audit.
Defect Classification
Commercial purchasing contracts split nonconformances into separate tiers to prevent cosmetic deviations from carrying the weight of structural faults. Critical defects such as exposed needle fragments or broken snaps that present immediate safety hazards typically carry a zero-tolerance limit. Major defects that impair garment saleability or structural integrity, including skipped stitches and torqueing side seams, commonly register at an aql of 2.5.
Minor blemishes such as loose raw edge threads or slight shade variations across internal facing panels sit at 4.0. Setting these tiered values distributes quality verification pressure across the garment production line.
Contractual Boundary
The threshold establishes only a statistical sampling plan and does not constitute a warranty of uniform excellence across bulk production. Sampling tables operate on random distribution, meaning clustered processing faults can escape detection if localized to isolated packaging cartons. Mills sometimes claim a lot is defect-free because it cleared verification, yet the mathematical reality implies that a specific proportion of defective units enters commerce legally under the agreed score.
Legal title transfer remains tied to this statistical gate, but consumer returns for systemic stitching failure bypass sampling agreements entirely.