AQL 2.5 vs 4.0: Which Standard for Your Handbag Order

Which inspection standard should a handbag order run to — AQL 2.5 or AQL 4.0? AQL 2.5 for anything with a visible brand mark, a load-bearing strap or a retail customer who inspects; AQL 4.0 only for minor defects, or for a giveaway grade where the bag is a container rather than a product. The difference is not two numbers on a page: on a 1,200 piece lot drawn at 80 units, AQL 2.5 tolerates five major defects while AQL 4.0 tolerates seven, and against a process actually running at 5 per cent defective that gap turns a 4 per cent chance of rejection into roughly a 22 per cent chance. Both standards come from the same table in ISO 2859-1, so choosing between them is not a question of which is correct. It is a question of which failure you would rather pay for: a re-sort at origin, or returns at the destination.

AQL 2.5 against AQL 4.0, on one lot of 1,200 pieces

The One Line Answer, and the Two Exceptions

Default to AQL 2.5 on major defects and AQL 4.0 on minor defects. That pairing is the working shape of most handbag programmes and it is the shape to write into a purchase order unless something specific argues against it. The two exceptions are narrower than most people assume, and both are about the destination rather than the bag.

• Exception one: the bag is not the product. Promotional totes, conference giveaways and packaging-grade bags carry no brand promise and no resale value. Running them at AQL 4.0 on majors saves a re-sort on goods that will be handed out and forgotten.
• Exception two: the defects are genuinely cosmetic and the channel is forgiving. A market-stall line sold at a price point where a glue mark does not generate a return can run loser on minors, provided criticals stay at accept zero.

What is not an exception is the desire for a smoother inspection. A looser AQL does not make goods better and it does not make an inspection faster; it changes who absorbs the cost of defects found later. A sourcing partner that proposes AQL 4.0 on majors because AQL 2.5 is hard to hit is telling you something useful about the process, and the useful response is to ask which operation is producing the defects rather than to move the bar.

Side by Side: The Accept Row at Every Lot Size

Both standards are read from the same table. The code letter and the sample size come from the lot size and the inspection level; only the accept and reject pair changes when the AQL changes. Under General Inspection Level II with single sampling, the comparison looks like this across the range a private label handbag programme actually runs at — MOQ 500 pcs per style / colourway and upward.

Two things are worth reading out of that table. The gap widens with the order: two defects at 1,200 pieces, four at 10,000, seven at 35,000. And the sample size never changes between the two standards. Choosing AQL 4.0 does not reduce the inspector's work by a single unit; it only raises the number of defects they are allowed to record before the lot fails.

What Each Standard Does to a Process Running at Five Per Cent

The accept row is easy to read and easy to misunderstand, because it says nothing about how often a lot actually fails. That depends on the operating characteristic curve: the probability that a lot with a given true defect rate will be accepted by a given plan. Run the numbers for sample size 80 and the two standards diverge sharply in the range where real programmes sit.

Read the 5 per cent row again, because that is the row where the decision lives. A programme whose process average sits at 5 per cent major defects — not unusual on a first run of a structured body with a new lining — will have roughly one shipment in five rejected under AQL 2.5 and roughly one in twenty-five under AQL 4.0. On 1,200 pieces, each rejection costs a re-sort at USD 0.35-0.60 per piece, so USD 420-720, plus a second inspection and the schedule damage of a 3-6 day delay against a 35-50 day bulk window.

Now read the same row from the other side. Under AQL 4.0, nineteen out of twenty of those shipments sail through with 5 per cent of the goods defective. At 1,200 pieces that is 60 defective units arriving at a warehouse, and the return handling cost of USD 12-18 each puts the destination-side bill at USD 720-1,080 before a single customer is lost. The two standards do not differ in quality. They differ in where the bill lands.

A Six-Question Decision Checklist Before You Pick One

Work these six in order. The first two settle most orders and the last two are the ones people forget.

1. Does the bag carry your mark where a customer will see it? If yes, majors at AQL 2.5. A visible logo on a defective unit is a brand problem, not a quality problem.
2. Is there a load-bearing joint — strap, handle, harness? If yes, majors at AQL 2.5. Attachment failures are criticals or majors, and they are the defect that produces a safety complaint.
3. Who inspects on arrival? A retail or marketplace customer running their own incoming check effectively imposes their standard on you regardless of what your purchase order says.
4. What is the unit economics of a return? If return handling exceeds the re-sort cost per piece, tighten. Re-sort at origin is USD 0.35-0.60; return handling is USD 12-18.
5. Is this the first run of a new construction? If yes, hold AQL 2.5 and accept a higher rejection risk, because the re-sort data is what fixes the second run.
6. Can the destination absorb a delay? If no, consider a price adjustment clause for minors rather than a looser AQL, so a borderline lot ships on time with compensation.

Where AQL 4.0 Is Genuinely the Right Call

AQL 4.0 has honest uses, and pretending otherwise makes the whole comparison less credible. The common thread is that the bag's function is containment rather than representation.

• Promotional and event totes. A printed tote handed out at a conference is judged on whether it holds a folder. Running majors at AQL 4.0 keeps the unit price where the marketing budget needs it.
• Packaging and gift-with-purchase bags. Where the bag ships inside another product, the cosmetic bar can be looser provided the seams hold and there is no chemical non-compliance.
• Minors on any programme. AQL 4.0 is the normal setting for minor defects even on a premium line. A thread tail over 3 mm is not worth a re-sort on 1,200 units.
• Emergency re-shipments. Where a launch date is immovable and the defect is cosmetic, an agreed allowance at AQL 4.0 with a price adjustment beats a missed season — provided it is in writing and provided criticals remain at zero.

Where AQL 2.5 Pays for Itself

The case for AQL 2.5 is not moral, it is arithmetic plus brand. Three situations produce the clearest return, and all three are common in private label handbag programmes.

First, any programme with a repeat customer. The second order is where a handbag programme becomes profitable: sampling can be skipped where the specification is unchanged, removing 6-10 working days, and retained tooling amortises a second time. A defect escape on the first order damages the relationship that generates that repeat, and USD 420-720 of re-sort is a cheap insurance premium against it. Second, anything sold through a channel with a published acceptance policy. Marketplace and retail customers increasingly run their own incoming checks; a shipment that passed AQL 4.0 at origin can fail theirs, and the cost of that failure is a chargeback rather than a re-sort.

Third, anything regulated. Chemistry does not have an AQL. Nickel release above 0.5 micrograms per square centimetre per week is restricted under REACH Annex XVII whether one unit or fifty are affected, Proposition 65 carries a warning obligation in California independent of sampling, and CPSIA sets 90 ppm for lead in coatings on children's items. Those are document-based pass or fail conditions, and they sit outside the sampling argument entirely — which is why the critical class is accept zero under both standards.

Splitting the Difference: Class by Class

The most useful realisation in this comparison is that you do not have to pick one number for the whole order. ISO 2859-1 contemplates a separate AQL per defect class, and that is how a mature specification is written.

A premium line is worth calling out separately. Moving majors from AQL 2.5 to AQL 1.0 on a 1,200 piece lot drops the accept number from five to two, which is a material tightening; it is justified where the retail price puts the cost of a single visible defect above the cost of the re-sort. Below roughly USD 200 retail, AQL 2.5 remains the sensible default.

Recording the Choice So It Survives the Second Order

Whatever is chosen has to be written down in a form that can be checked a year later, by somebody who was not in the conversation. One clause does it, and it should name the standard, the level, the plan, the class and the counting method.

• Standard and level: pre-shipment inspection to ISO 2859-1, General Level II, single sampling, normal inspection.
• Class limits: critical accept 0; major AQL 2.5; minor AQL 4.0.
• Counting method: nonconformities per hundred units, stated explicitly.
• Defect definitions attached, with handbag-specific examples per class.
• Fail consequence: re-sort, re-inspection and a revised ship date in writing.
• Report format: issued to the buyer within 24 hours, with photographs and the sampling basis on the page.

Write it once and the switching rules take care of the rest: two of five consecutive rejected lots moves the programme to tightened inspection, where a 1,200 piece lot at AQL 2.5 tolerates three majors instead of five. Ten consecutive accepted lots can qualify for reduced inspection at 32 units. The AQL 2.5 arithmetic in full covers the sample table and the report contents, and the twelve-station inspection checklist turns whichever number you choose into a walk-through. The handbag quality control reference sets out where this decision sits in the wider programme.

Writing the Defect Catalogue So Two People Agree

An AQL is a number attached to a list, and the list is where most arguments start. Two people can read the same ISO 2859-1 table and disagree about whether a scuffed edge paint is major or minor, which changes the accept number by two and the outcome by a shipment. A written catalogue of defects, with each one assigned to a class and each class assigned an AQL, removes that argument before there is a lot to inspect.

How the Eighty Units Are Actually Drawn

The sample size is fixed by the standard and so is the way it is drawn, which is the part nobody checks. A draw that is not random is not an inspection under ISO 2859-1, whatever the report says, and a draw concentrated in the first cartons packed will systematically miss a defect that appeared halfway through the run. Three rules make the draw defensible.

• Select cartons before units. On a 1,200-piece order in 80 cartons, draw 8-10 cartons at random by number, then draw units from within them.
• Stratify by colourway. Where the lot holds three colourways of 400, draw proportionally rather than taking the nearest 80 units.
• Cover the run, not the stack. Ask for units drawn from the beginning, middle and end of the packing sequence, because a setting drift shows up only across it.
• Record the numbers. The carton numbers and unit positions belong in the report, so a re-inspection can be compared against the first.

What an AQL Does Not Catch

Sampling is a filter with a known mesh size, and three classes of problem pass straight through it. Naming them is not a criticism of the standard; it is the reason the specification, the chemistry file and the loading check exist alongside it. A buyer who thinks an AQL 2.5 report means the goods are good will eventually be surprised, usually by something the report never looked at.

• Chemistry. Chromium VI by ISO 17075 and nickel release are pass or fail on the substance, not on the sample. One affected unit is the same problem as fifty.
• Systematic drift. A strap drop moving 1 mm per 100 units stays inside tolerance across the sample and outside it by unit 400, which is why the in-line check exists.
• Carton-level faults. Wrong carton count, mixed colourways in one carton and a short shipment are packing faults, found at a loading check rather than at the inspection table.
• Latent failures. An adhesive bond that has not cured, or an edge paint that cracks after 24 hours, looks correct at inspection and fails in week three.

Frequently Asked Questions

People Also Ask