Handbag Manufacturers With AQL 2.5 QC: What That Really Means

What does AQL 2.5 really mean when a handbag partner offers it? It means a defined number of units gets pulled from your lot and inspected against a written defect list, using the sample sizes and accept-reject thresholds published in ISO 2859-1, and it fails the lot at a number rather than at an opinion. At MOQ 500 pcs per style / colourway the general inspection level II code letter is H, so 50 units get pulled and the fourth major defect fails the run, alongside 6-10 working days of sampling and 35-50 days of bulk elsewhere in the programme. AQL 2.5 is a statistical gate that says how many defects in how many inspected units trigger rejection — at 500 pcs that is 50 units with an accept of 3 and a reject of 4 major defects. Everything below follows from six questions buyers actually ask: how many get opened, what counts, who writes the list, what failure costs, where the method stops working, and what belongs in the purchase order.

AQL reference points

What Does AQL 2.5 Actually Promise?

AQL stands for acceptance quality limit, and the number is the percentage of defects the sampling plan tolerates in a lot judged acceptable. Two point five means roughly 2.5 defective units per hundred are tolerated at the plan's operating characteristic, not that 2.5% of your shipment will be bad. That distinction trips up almost every first-time buyer: the plan describes what it will accept, not what it predicts, and it says nothing at all about the units nobody opened.

The value of the standard is that it removes negotiation from the moment of inspection. Every term is fixed in advance — how many cartons to pull from, how many units from each carton, which defect classes exist, and what number fails the lot. Two inspectors running the same plan on the same lot reach the same verdict, which is why ISO 2859-1 appears in purchase orders worldwide and why vague phrases about being strict are worthless by comparison.

One further point matters more than the arithmetic: AQL governs a pre-shipment sample, so it is a gate rather than a description of the run. Insist on cartons selected across more than one packing day, or the verdict quietly describes one shift.

How Many Units Get Opened From a 500-Piece Lot?

Sample size comes from two inputs in ISO 2859-1: lot size and inspection level. Take general level II for ordinary goods, find your code letter from the lot-size band, and read the sample size for that letter. A 500-piece order falls in the 281-500 band, which is code letter H, giving a sample of 50 units. Moving up the ladder changes both the letters and what those 50 units can tolerate.

Two consequences follow. First, a 500-piece run is inspected to a proportionally stricter-looking threshold than a 3,000-piece run, because small samples tolerate fewer defects in absolute terms. Second, the plan can be run at level I or level III instead, which lowers or raises the sample size for the same lot; level III is worth requesting on a first run, and cost-effective on any style with a history of returns.

What Does Accept 3 and Reject 4 Mean in Practice?

It is a single-sampling plan: one draw of 50 units, one verdict, no second chance inside the plan. Zero to three major defects and the lot passes. Four or more and the whole lot is rejected — not those four units, the whole lot. That all-or-nothing structure is deliberate, and it is precisely why payment belongs against the inspection report rather than against a loading date.

What Counts as Critical, Major and Minor on a Bag?

The plan is meaningless without defect classes, and this is where most disputes start, because the checklist rather than the standard decides what a defect is worth. Three classes cover the field. Critical covers anything that breaches a legal limit or creates a safety hazard. Major covers what a customer would notice and return. Minor covers a deviation from specification that a customer would accept on the shelf.

• Critical. A lead or phthalate result above the CPSIA limit, a sharp burr on metal trim, a detachable small part on a children's item, or an undeclared Substance of Very High Concern under REACH.
• Major. Stitch density below the specified 7-9 per inch, edge paint cracked at a fold, a strap attachment without its reinforcement tape, plating rubbed through at a contact point, a zipper that catches the lining.
• Minor. A hangtag misaligned by two millimetres, a dust bag with a loose drawcord, or a carton mark printed slightly off register.

Most plans accept zero critical defects without argument, some part of the minor tally, and everything between them falls to how the AQL number is applied. A common arrangement is AQL 0 for critical, 2.5 for major and 4.0 for minor, and that triple is what should appear in writing. Without it, a single number covers all three classes, and an argument about whether a crocking failure on a dark lining is major or minor becomes a scheduling problem.

Which Tests Belong on the Defect List?

Crocking deserves a named entry on the list, because it generates the most expensive complaint a bag brand receives: colour transfer onto clothing. Ask for a rub fastness result tested to ISO 105-X12, reported on the grey scale, and treat anything below 3-4 on dark linings as major rather than cosmetic. Pair that with a nickel release figure to EN 1811 on metal trim and a chromium VI result to ISO 17075 on leather, both of which are documentation rather than opinion.

Who Writes the Defect Checklist?

The party that writes the list writes the outcome, so it should be you, or jointly you and the inspector. left to the standard alone, an inspection covers appearance, workmanship and quantity — plenty enough to fail a badly made run and nowhere near enough to catch the failure mode that would sink your season. Your own list is where brand-specific requirements go: embossing depth, branded zip puller alignment, buyer-owned tooling used only for your runs.

For most brands the list runs ten to twenty lines and takes an afternoon to write, drawing on whatever drove returns last season. Send it with the purchase order and ask for acknowledgment in writing, then ask who holds it at the inspection table. Third-party inspection can be booked through SGS or another agreed body, and the report is issued to you either way, which only helps if the report measures your list rather than a generic one.

What Happens When a Lot Fails?

A failed lot has three possible outcomes, and agreeing which applies before the run is cheaper than arguing afterwards. Re-sort means opening every carton and removing defective units, usually at the cost of whoever is found responsible. Rework means repairing the specific fault, which is feasible for some failures and impossible for others — an edge-paint problem can be refinished, a mis-cut panel cannot. A rejected lot means a fresh production slot, which inside a 35-50 day window can cost you the season.

The money question hiding inside all three is who pays, and it is unassigned in most quotations. Write it down: re-sort labour, rework material, additional inspection visits and air freight if rework pushes past the vessel cut-off, then check how that line reads in writing row nine into a quotation. Tie the balance to document completion so a borderline result cannot be resolved by shipping anyway.

How Do Inspection Levels and Switching Rules Change the Result?

ISO 2859-1 carries switching rules that move inspection between normal, tightened and reduced states based on recent history, and those rules explain why two apparently identical orders get different treatment. Tightened inspection, triggered after consecutive poor lots, keeps the sample size but lowers the accept number, so the next lot has to be measurably better to pass. Reduced inspection rewards a clean streak with smaller samples, and carries its own risk for the buyer.

This is the strongest argument against changing production partner for every season. A stable relationship accumulates its inspection history, and that history has value in both directions: a clean run of lots earns you statistical credit, while restarting with somebody new puts every order back into the strictest posture. Note it alongside partner versus vendor behaviour when weighing a 3% saving against a documented quality record.

How Do You Verify an AQL Claim Before Committing?

A party quoting AQL 2.5 should be able to answer four questions in one email, with documents attached rather than reassurances. If any answer is qualitative, treat the inspection as unconfirmed and plan your own third-party visit, because the gap between promising AQL 2.5 and running it is exactly where hidden quality cost lives. Answers of this shape also distinguish a production partner from an intermediary faster than any pricing question.

1. Which standard and which revision? The answer should name ISO 2859-1 and the inspection level, not merely quote a number.
2. Who performs the inspection, and who receives the report? Internal QC can be excellent, but the report must be issued to you, with photographs of the defects found.
3. Send the most recent inspection report, redacted if necessary. Reports show whether the plan is run or merely named, and whether defect classes are defined.
4. What are the accept numbers at my lot size? A party running the plan knows that 500 pcs means 50 units and a reject at four.

Where AQL Stops Working

Three limitations deserve stating plainly, because buyers who discover them at the inspection table tend to conclude they were misled. First, the plan inspects attributes, not variables: a unit either meets the line or it does not, so a strap that is 3 mm short and one that is 30 mm short count the same. Second, it is a sample, so a fault confined to cartons nine through fourteen can pass cleanly, which is why random carton selection across packing days matters.

Third, and most importantly, no sampling plan can substitute for the tests that only appear when you ask. Salt-spray exposure to ASTM B117, ISO 105-X12 rub fastness, EN 1811 nickel release and ISO 17075 chromium VI are laboratory procedures, not defects you can see on a bench. A lot can pass AQL 2.5 on appearance and still fail entry because nobody commissioned a REACH scope or checked Proposition 65 exposure points on plated hardware, both of which belong in the same compliance packet referenced in every quotation.

What to Write Into the Purchase Order

Six lines convert an inspection phrase into an enforceable obligation, and every one of them is a sentence rather than a clause. Put them in the order given, and have them acknowledged by reply rather than assumed, because silence at this stage is how programmes lose their leverage. Anything not written here gets negotiated at the worst possible moment: after production has finished and before your launch date.

Which Two Commercial Lines Make the Rest Bite?

Then add the two commercial lines that make the rest bite: the balance payment falls due against the inspection report rather than a calendar date, and any lot failing the plan is not loaded without your written instruction. Those two sentences turn a quality claim into a commercial position. Discuss an inspection clause with a private label coordinator before the first run, not after it.

Which Cartons Get Opened, and How Many of Each

ISO 2859-1 fixes how many units are drawn but says nothing about which cartons they come from, so that choice is made at the inspection table and it quietly decides what the verdict describes. Cartons taken from one packing day produce a result about one shift's work; cartons spread across the lot produce a result about the lot. Every rule below is a sentence in the inspection booking rather than a favour to ask for. The mechanics are set out at greater length in the plain-language guide to AQL 2.5 inspection.

• How many cartons. A widely used rule is the square root of the carton count, rounded up, plus one: twenty-five cartons gives six selected, sixty-four gives nine.
• Spread across time. Where a lot ran across more than one packing day, selections are taken from at least two days, otherwise one bad or good shift decides the verdict.
• Proportional units. Units are drawn in proportion to the contents of each selected carton rather than in equal numbers, so an inspector follows the packing plan.
• Never the top of the stack. Selecting six cartons from the top layer is the most common accidental bias at the table, and it is impossible to detect from the report.
• Count first. The physical carton count is reconciled with the packing list before selection begins; a discrepancy is recorded even when it does not change the verdict.

The Reference Sample, and the One That Stays Behind

A defect classification without a physical reference is a set of opinions waiting for an argument. Three artefacts settle it, and the question is rarely worth raising until the second season, by which point the first set has usually gone missing. The sample-type question is explained in the PP and golden sample comparison, and the retention rules below assume both exist.

• Three sealed references. One for the brand, one for the line, one held by whoever runs the third-party inspection, each sealed and signed with a date.
• Two units of each style. One to keep sealed and one to handle, which halves the number of arguments about whether this is really the approved bag.
• Storage away from light. References left in daylight no longer represent the lot; leather darkens perceptibly within twelve to eighteen months of display.
• Replacement cycle. Every twelve to eighteen months, or every third production lot, whichever arrives first, and the replacement is cut from a current bulk unit rather than a sample.
• Record the lot number. A reference that cannot be traced back to a hide lot or a plating batch settles a colour argument but not a material one.

What an Inspection Actually Costs, and What It Saves

The money side of inspection is rarely written down, which is why it gets argued about every season. Write it once and the argument ends. For a first order at 500 pieces the whole exercise costs less than one badly handled return wave, and the comparison becomes starker as the order grows. The going rates and what they buy are set out in the third-party inspection cost guide.

• Man-day rate. Third-party inspection in Fujian typically runs USD 250-350 per man-day, and a 500-piece lot at general level II occupies about one man-day.
• Per unit. Across 500 pieces that is USD 0.50-0.70 per bag, including the written report with photographs.
• Re-sort at origin. Where defects justify opening every carton, expect roughly USD 0.35-0.60 per piece plus two to four days before the vessel.
• Returns at the destination. Each returned unit usually costs USD 18-35 once inbound freight, replacement and handling are counted, and retailers charge these rather than negotiate them.
• Break-even. Fifteen returned units already exceeds the whole inspection; one thousand pieces inspected costs about the same as thirty returns.

Frequently Asked Questions

People Also Ask