AQL 2.5 is the acceptance quality limit most mid-to-premium bag programmes apply at final random inspection under ISO 2859-1, single sampling, general level II: a lot passes when the defects found in the drawn sample sit at or below the published acceptance number. In plain terms, 2.5 tolerates roughly two or three minor faults per hundred units while still failing the batch on any critical failure such as a split load-bearing seam or hardware that breaks below its pull-test threshold. The number is a filter, not a promise, which is why dependable OEM bag quality assurance layers material intake checks, inline inspection and pre-shipment verification around the final draw.
What AQL 2.5 Actually Means for a Bag Order
Acceptance quality limit is often misread as a defect rate the factory is allowed to ship. It is not. It is the threshold at which a sampling plan, drawn at random from a finished lot, will pass or fail the entire batch. A 2.5 limit sits in the middle of the commercial range: tighter than the 4.0 commonly used for promotional soft goods, looser than the 1.0 some heritage leather houses insist on for full-grain exteriors.
The distinction matters commercially. A buyer who specifies 2.5 and then rejects a shipment because one unit in forty had a loose thread is misusing the standard, and a factory that ships consistent major defects while hiding behind the acceptance number is doing the same. The contract should state the AQL, the sampling plan, the inspection level and the defect definitions — four separate clauses that are too often collapsed into one line.
| AQL level | Typical application in bags | Risk posture | Cost implication |
|---|---|---|---|
| 0.65 | Full-grain leather exteriors, heritage constructions, gift-with-purchase for VIP tiers | Very conservative; high rejection probability on cosmetic variation | Extra sorting labour, higher unit cost, longer inspection window |
| 1.0 | Premium handbags with visible edge paint and exposed stitching | Conservative; natural leather marks still need a reference standard | Moderate uplift; requires agreed defect reference boards |
| 2.5 | Standard commercial default for fashion handbags, totes and backpacks | Balanced; accepts minor cosmetic variance, rejects functional failure | Baseline; no surcharge in most programmes |
| 4.0 | Promotional totes, giveaways, non-critical packaging items | Permissive; functional defects still classed separately | Lowest inspection cost per unit |
| 0 (critical) | Any child-safe, safety-critical or load-bearing component | Zero tolerance regardless of the cosmetic AQL | Mandatory inline verification and pull-test records |
One subtlety deserves emphasis. Critical defects carry their own zero acceptance number and are never folded into the 2.5 arithmetic. A bag with a needle broken off inside the lining, a detachable strap fitting that releases under load, or a magnetic closure that fails CPSIA small-parts rules in a children's line is a critical failure whatever the cosmetic tally says.
How the Sampling Plan Works Under ISO 2859-1
The standard converts a lot size into a sample size code letter, then converts that letter plus the chosen AQL into a sample size and an accept/reject pair. General level II is the default inspection level and the one most bag programmes adopt; level I halves the sample and is sometimes used for repeat orders from a proven line, while level III roughly doubles it and is reserved for first runs or high-value exotic materials.
Programme parameters at a glance
| Parameter | Commitment |
|---|---|
| Minimum order quantity | 500 pieces per style / colourway |
| Sampling lead time | 6–10 working days (12–15 days for complex constructions) |
| Bulk production lead time | 35–50 days after sample approval |
| Inspection standard | AQL 2.5 (ISO 2859-1), single sampling, general level II |
| Incoterms | FOB Xiamen — EXW, CIF and DDP available on request |
| Payment terms | T/T 30% deposit / 70% before shipment; L/C at sight for container loads |
| Verified production floor space | 4,950 m² (SGS-verified) |
| Team | 137 staff across 7 production lines |
| Equipment | 149 machines including computerised saddle-stitch and edge-painting units |
| Monthly capacity | 200,000 pieces |
| Entity | Founded 2014; principal has worked in bag development since 2004 |
A worked example makes the mechanics concrete. For a lot of 1,200 units at general level II, the code letter is J, which calls for a sample of 80 pieces. At AQL 2.5 the accept number is 5 and the reject number is 6, so six or more defects in that random draw fails the lot. The same lot at AQL 1.0 gives accept 2, reject 3 — materially stricter, and materially more likely to trigger a re-sort.
| Lot size (pieces) | Sample size (Level II) | Accept / Reject at 2.5 | Accept / Reject at 1.0 |
|---|---|---|---|
| 151–280 | 32 | 2 / 3 | 1 / 2 |
| 281–500 | 50 | 3 / 4 | 1 / 2 |
| 501–1,200 | 80 | 5 / 6 | 2 / 3 |
| 1,201–3,200 | 125 | 7 / 8 | 3 / 4 |
| 3,201–10,000 | 200 | 10 / 11 | 5 / 6 |
| 10,001–35,000 | 315 | 14 / 15 | 7 / 8 |
Buyers should also understand switching rules. ISO 2859-1 provides for tightened inspection after a run of rejected lots and reduced inspection after sustained good performance. Programmes that never invoke the switching rules are leaving a control lever unused — and in some cases quietly shipping worse quality than the stated AQL would suggest, because no one is tracking lot history.
Defect Classification: Critical, Major and Minor
Most disputes over a failed inspection are not really disputes about counting. They are disputes about classification. A scuff on a bottom-stud is minor to one inspector and major to another; a shade difference between the flap and the body is a defect to a brand that sells tonal dressing and a non-issue to one that sells contrast. The fix is a written defect dictionary, agreed at sample approval, with photographic reference boards.
| Class | Definition for bags | Representative examples | Handling |
|---|---|---|---|
| Critical | Renders the product unsafe, illegal or non-functional | Broken needle fragments, strap fitting release under load, lead or phthalate exceedance, choking hazard on a child's item | Zero acceptance; lot rejected and root cause required before rework |
| Major | Visible or functional fault a retail customer would return | Split seam at handle base, zip that jams, edge paint flaking on the flap, misaligned logo debossing, lining torn at the pocket | Counted against the AQL; above the reject number the lot fails |
| Minor | Workmanship variance below the return threshold | Single loose thread, light glue halo under a welt, small shade step within tolerance, dust mark removable by wipe | Counted against the AQL; high minor counts often reveal a process drift |
- Write the defect dictionary before sampling, not after the first failed lot, so both parties read the same standard.
- Photograph each agreed defect class on a reference board and keep one copy on the line and one with the inspection team.
- Separate cosmetic tolerance on natural leather from workmanship faults — grain variation, healed scars and shade steps are material characteristics, not defects.
- Set a functional testing protocol alongside the visual AQL: strap pull, zip cycle, colour fastness to rubbing, handle swing.
- Agree in writing whether inspection costs are borne by the buyer, the programme, or shared on a first-failure basis.
Inspection Stages Across the Production Cycle
Final random inspection is the last gate, not the only one. By the time cartons are packed, most of the quality outcome is already determined — by the leather selection, by the splitting and skiving accuracy, by whether the adhesive was given its full open time. Programmes that rely on a single end-of-line draw tend to alternate between panic rework and unexplained escapes.
| Stage | Timing | What is checked | Output document |
|---|---|---|---|
| Material intake | On receipt of hides, fabric, hardware, linings | Thickness, temper, shade against the approved swatch, coating adhesion, hardware plating and pull strength | Incoming inspection record with lot traceability codes |
| Pre-production | Before cutting bulk | Approved sample, confirmed BOM, pattern set, cutting dies, print or emboss tooling | Pre-production meeting minutes and signed counter-sample |
| First-piece / inline | First 20–50 units off the line | Seam allowance, stitch density, edge paint coverage, hardware alignment, assembly sequence | Inline inspection report with corrective notes |
| Mid-production | At 20–40% completion | Process stability, recurring faults, operator technique drift | Mid-run report and retraining log if required |
| Final random inspection | At 80–100% packed | AQL 2.5 visual and functional check, packing, labelling, carton count | FRI report with photographic evidence |
| Loading supervision | At container stuffing | Carton quantity, seal numbers, stacking, container condition | Loading report with photographic record |
On our side, the programme team works from the SGS-verified production floor in Quanzhou, Fujian, across 7 lines and 149 machines, with inspection built into the line rather than bolted on at the end. Our production partners build to your specification and the inspection protocol is written into the order file from the first sample, so the person checking the final carton is measuring against the same reference the cutter used.
Frequent Bag Defects and Where They Originate
Defect patterns are diagnostic. A recurring fault at a predictable location almost always traces to a process parameter rather than to operator carelessness, and treating the symptom by adding inspectors rarely works. The table below maps the faults we see most often to their usual origin and the control that prevents them.
| Defect | Typical root cause | Preventive control |
|---|---|---|
| Edge paint cracking or peeling | Insufficient drying between coats, paint applied over unsealed flesh side, low-grade acrylic in cold weather | Specify coat count and drying time; test flexibility on the actual leather; keep a retained edge sample per batch |
| Handle or strap seam failure | Stitch density too low, thread under-spec, insufficient seam allowance at the load point | Define SPI and seam allowance on the tech pack; run a pull test to an agreed newton threshold on every colourway |
| Zip waviness or jamming | Tape fed under tension, slider mis-matched to the chain, welt cut short | Relax-feed training; verify slider and chain are a matched set; check welt length at first-piece |
| Hardware plating wear or tarnish | Thin plating under the specified micron count, no lacquer seal, salt-spray exposure in transit | Specify plating thickness in microns; require a salt-spray or abrasion test report from the hardware vendor |
| Lining puckering or excess fullness | Lining cut to the shell pattern rather than eased, no allowance for turn | Develop a separate lining pattern; verify at first-piece and again at mid-run |
| Logo placement drift | No jig or template for debossing, heat-press, or embroidery hooping | Require a placement jig; measure from two fixed datum points and record on the inline sheet |
| Adhesive bleed at welts | Excess adhesive, insufficient open time, wrong viscosity for the substrate | Set adhesive gram weight; log open time; keep a clean-wipe station at each bench |
The Quality File Every Buyer Should Hold
Documentation is what turns a good shipment into a repeatable one. When a brand reorders twelve months later, the difference between a smooth repeat and a slow, expensive rebuild is almost always whether the original specification, tolerance and defect references were captured properly. The following list is the minimum we recommend keeping under your own control, not only in the factory's folder.
- Request lot traceability codes on every carton so a field complaint can be traced back to a hide batch and a line.
- Keep your own retained sample from the shipped lot, not only the pre-production approval — the two can diverge.
- Ask for the raw inspection sheets, not only the summary pass/fail page; the distribution of faults tells you where the process is drifting.
- Record which operator or station a recurring fault came from. Most persistent defects concentrate in one or two steps.
- Store the file in a shared location the programme team can access, so a reorder brief is a reference number rather than a re-negotiation.
Corrective Action: What Happens After a Failed Lot
A failed final inspection is expensive for everyone. Cartons are unpacked, units re-sorted, the line is disrupted, and the shipping window narrows. The productive response is not more inspection but a closed loop: identify the root cause, change the process or the input, then verify the change on the next run and record it. Programmes that skip the verification step tend to see the same defect return within two or three orders.
- Classify the failure. Separate critical, major and minor counts and identify whether one fault type dominates or the defects are diffuse.
- Trace the lot. Pull the material batch codes, the line, the shift and the operator station for the affected units.
- Confirm root cause on the floor. Walk the specific step with the supervisor and observe the method in use, not the method written down.
- Agree the corrective action in writing, with an owner and a date. Rework alone is not a corrective action unless the reworked units are re-inspected to the same plan.
- Verify effectiveness on the next run of the same construction, then close the entry in the corrective action log.
- Update the tech pack or defect dictionary if the failure revealed an ambiguous specification rather than a process error.
It is also worth agreeing the commercial mechanics in advance. Who pays for re-inspection, who absorbs the cost of a second sorting pass, and what happens to the shipping date when a lot fails are all easier to settle at contract stage than at 2 a.m. before a vessel cut-off.
Build your inspection protocol with us
Send us your tech pack and target market, and the programme team will return a sampling plan, defect dictionary template and inspection schedule matched to your construction. Reach us at service@junyuanbags.com or on WhatsApp at +86 177 5002 0688.
Frequently Asked Questions
What does AQL 2.5 mean in bag manufacturing?
AQL 2.5 is an acceptance quality limit applied during final random inspection under ISO 2859-1. For a given lot size and inspection level, the standard sets a sample size and an accept/reject pair; at general level II a lot of 1,200 pieces draws 80 units, with acceptance at five defects and rejection at six. It describes the threshold at which the statistical sample fails the batch, not a permitted defect rate for the shipment.
Is AQL 2.5 good enough for luxury leather handbags?
For most commercial premium programmes, 2.5 is an appropriate balance of cost and control. Full-grain leather exteriors with visible edge paint and exposed stitching often justify 1.0 or 0.65, because cosmetic variance is more noticeable. The more important question is whether a defect dictionary and reference boards exist, since classification disputes cause more friction than the numeric limit itself.
How many pieces are inspected for an order of 5,000 bags?
At general level II, a lot of 3,201 to 10,000 pieces draws a sample of 200 units, with an accept number of 10 and a reject number of 11 at AQL 2.5. Critical defects remain at zero acceptance regardless of sample size. If the programme has a history of passed lots, reduced inspection may apply; a history of failures triggers tightened inspection under the switching rules.
What is the difference between a major and a minor defect on a bag?
A major defect is one a retail customer would notice and return the product for, such as a split seam at the handle base, a zip that jams, or misaligned logo debossing. A minor defect is a workmanship variance below the return threshold, such as a single loose thread or a removable dust mark. Critical defects involve safety, legality or total loss of function and carry zero acceptance.
Can I specify a zero-defect requirement instead of AQL 2.5?
You can specify a stricter limit, and some heritage houses do, but a literal zero-defect visual requirement usually converts into 100% sorting rather than sampling. That is slower and more expensive, and it does not remove the need for inline control. A more workable structure is AQL 1.0 or 0.65 for visual defects with a separate zero tolerance for critical safety and functional failures.
Who pays when a shipment fails inspection?
Commercial practice varies, and the right answer is to settle it in the contract before production begins. A common structure is that the factory absorbs the cost of rework and re-inspection when the failure traces to workmanship, while the buyer absorbs changes arising from a specification revision made after approval. Delays to the shipping window should be addressed as a separate clause.
What inspections happen before the final AQL check?
A structured programme runs material intake checks on hides, fabric, hardware and linings; a pre-production review of the approved sample, BOM and tooling; and first-piece or inline inspection on the earliest units off the line, followed by a mid-run review. Final random inspection then verifies the packed lot. Each stage produces a written record with lot codes, which is what makes a later complaint traceable.
How long does a final inspection take for a typical bag order?
Most final random inspections for a single style in the 1,000 to 5,000 piece range are completed within one working day, with the written report issued within 24 hours. Complex constructions, multi-colourway lots, or orders requiring functional testing such as strap pull and zip cycle testing will extend that window. Book inspection at least five working days ahead of the packing completion date.
Does AQL 2.5 apply to packaging and labelling?
Packaging and labelling defects are usually counted as minor or major depending on their effect on retail readiness. A missing care card is typically minor; an incorrect fibre or material content label can be major because it carries regulatory exposure. Where labelling is subject to market-specific rules, verify it separately against the destination requirement rather than folding it into the visual tally.
What is general inspection level II and why is it the default?
ISO 2859-1 defines three general inspection levels that scale the sample size relative to the lot. Level II is the standard default and the reference point for published accept and reject numbers. Level I uses a smaller sample and suits proven repeat runs; level III uses a larger sample and suits first runs, high-value materials, or a lot with a prior failure history.
Should hardware be tested separately from the visual inspection?
Yes. Visual AQL sampling will not reliably detect a plating that wears through in six weeks or a fitting that releases below its rated load. Hardware should carry its own incoming verification: plating thickness in microns, a salt-spray or abrasion result from the vendor, and a pull test to a defined threshold on load-bearing fittings. Failures here are critical, not major.
How do I stop the same defect recurring on reorder?
Close the loop rather than only reworking the lot. Classify the failure, trace it to a material batch and a station, confirm the root cause by observing the step on the floor, then agree a written corrective action with an owner and a date. Verify it on the next run of the same construction and log the outcome. If the root cause was an ambiguous specification, update the tech pack.
Are natural leather marks counted as defects?
Genuine material characteristics — grain variation, healed scars, neck wrinkles, and modest shade steps between panels — are not workmanship defects, but the tolerance must be defined in writing and illustrated. Without a reference board, inspectors and brand teams will disagree on every unit. Agree the acceptable range at sample approval and keep a retained sample from the shipped lot.
What certification frameworks sit alongside AQL inspection?
ISO 2859-1 governs the sampling plan. ISO 9001 covers quality management and ISO 14001 environmental management at the facility. REACH and CPSIA set chemical and safety limits for materials and components, OEKO-TEX and Leather Working Group audits address input sourcing and tannery practice, and ASTM test methods define mechanical checks such as abrasion and seam strength. They are complementary layers, not alternatives.