Should You Inspect at AQL 1.5, 2.5 or 4.0 — and Does 1.5 Change Anything?

What Does AQL Actually Mean in ISO 2859-1 and ANSI Z1.4?

AQL stands for acceptance quality limit, and it is the worst tolerable process average when a continuing series of lots is submitted for acceptance sampling. That definition carries three qualifications that buyers routinely skip. It is a process average, not a guarantee about any single carton. It applies to a continuing series, not to a one-off shipment. And it is a limit on the sampling plan, not a statement about what is in the box — the plan decides whether to accept the lot, and it decides that from a sample.

ISO 2859-1 and ANSI/ASQ Z1.4 are the two names you will see on an inspection report, and for practical purposes they describe the same scheme: sampling plans indexed by lot size and inspection level, with accept and reject numbers set by the chosen AQL. A report written to either one is readable by anyone who reads the other. What matters is not which standard is named but whether the report states the four things that make the verdict interpretable: the lot size, the inspection level, the AQL, and the sampling plan type.

The sampling plan is a pair of numbers. Ac is the accept number: the largest number of defects in the sample that still lets the lot pass. Re is the reject number: at or above it, the lot fails. They always differ by one, and that single unit of difference is where most commercial arguments about inspection actually happen. Everything else on the page is context around those two numbers.

Which AQL Levels Are in the Standard Tables, and Which Ones Get Used?

The standard prints an array of AQL values running from very tight to very loose. That array includes 1.5. Stating otherwise would be wrong, and a buyer who repeats the wrong claim to a vendor loses the argument immediately. The real distinction is between the values that exist in the table and the values that the industry actually inspects against, and those are two different sets.

General merchandise inspection runs on three numbers, and bags are not an exception. Critical defects: zero tolerance, because a critical defect is one that makes the product unsafe or unlawful and there is no acceptable rate for those. Major defects: AQL 2.5, because a major is a defect a customer would notice and return the product for. Minor defects: AQL 4.0, because a minor is a departure from specification that a customer might see and would not send the bag back over. That triple is what a working inspection report on a bag programme shows.

Four other AQL values appear in specialist contexts and are worth naming so nobody thinks they were overlooked. Tighter values in the region of 0.65 and 1.0 turn up in automotive and medical component supply, where the consequence of a defect is a safety recall rather than a return. Looser values such as 6.5 and 10 turn up on commodity goods where the defect is cosmetic and the unit value is low. Neither end describes a branded bag sold at retail with a warranty behind it.

Is AQL 1.5 a Working Inspection Level or a Marketing Number?

It is a real number in a real table, and it is used as a marketing number. Both halves of that sentence matter, because the argument is won on the second half rather than by denying the first. A vendor quoting AQL 1.5 is quoting a plan that exists. What the quote does not say is what it costs, what it changes, and — most importantly — whether the inspection level and the defect classification behind it are the ones you assumed.

Here is the mechanical fact that decides the conversation. Under single sampling at General Inspection Level II, the sample size is set by the lot size through the sample size code letter. The AQL does not set the sample size at all; it sets the accept and reject numbers for that sample. So "AQL 1.5 means we inspect more pieces" is simply false for any normal lot. The same 500-piece order draws the same 50 units whether the plan is written at 1.5 or at 2.5. What changes is the number of defects tolerated inside those 50 units.

There is one narrow case where a tighter level genuinely draws more units. At the smallest sample size code letters the standard tables carry a down arrow, which sends the inspector to a larger sample size with a different accept number, because a two-piece sample cannot discriminate between 1.5 and anything else. That is real and it applies to tiny lots, not to commercial orders. Above a sample of 13 units the sample size is identical and only the accept number moves.

What Does AQL 1.5 Actually Do to Your Sample Size?

Almost nothing, and that is the point most buyers get wrong in both directions. The sample size for a 500-piece lot at General Inspection Level II is 50 units under single sampling, and it is 50 units whether the plan is written at AQL 1.5, 2.5 or 4.0. If you have been told that a tighter AQL means a bigger sample, you have been told something the standard does not say. What a tighter AQL changes is the accept number inside that fixed sample.

At a sample of 50, the accept and reject pair is 2 and 3 at AQL 1.5, and 3 and 4 at AQL 2.5. One unit. On a 500-piece order that single unit is the entire practical difference between the two plans: a lot with three major defects in the drawn sample passes at 2.5 and fails at 1.5. Everything downstream — the re-work, the re-inspection, the argument about whether that third defect was major or minor, the delay to the booking — comes from that one unit.

The effect on total inspection volume over a programme is real but it is indirect, and it is worth being precise about because this is where the cost actually hides. A tighter accept number means more lots fail. A failed lot is re-worked and resubmitted, and a resubmitted lot is inspected again, usually at a full sample. It also means the programme never satisfies the consecutive-acceptance run that permits a switch to reduced inspection, which is the only mechanism in the standard that genuinely reduces sample size. So the long-run effect of 1.5 is more inspection hours, not bigger samples.

What Does Sample Size Do to Cost, Calendar and Dispute Rate?

Sample size is a cost and a calendar item before it is a quality item, and it should be planned that way. Drawing 50 units from a 500-piece lot means unpacking, inspecting, repacking and resealing cartons, and on a retail-ready pack that is a half-day job for two people. Drawing 125 units from a 3,200-piece lot is a full day. That time sits inside the inspection window at the end of the bulk schedule, and it is one of the reasons the 35-50 day window carries a corrective cycle rather than a single inspection date.

Read the first five rows carefully, because they settle the argument. Up to and including a lot of 90 pieces, AQL 1.5 and AQL 2.5 produce exactly the same verdict: any defect at all rejects the lot. A vendor advertising 1.5 on small orders is advertising a number that cannot change the outcome. From 91 pieces upward the numbers separate, and they separate by one unit of acceptance at every code letter up to the largest lots in the table.

How Do You Read a Sample Size Code Letter and an Ac/Re Pair?

Two lookups, in this order. First, take the lot size and the inspection level to the code letter table, which returns a letter. Second, take that letter and your AQL to the sampling plan table, which returns the sample size and the accept and reject pair. The letter is the index; the AQL is the column. Confusing the two is why people believe a tighter AQL means more pieces, when the letter decided the piece count before the AQL was ever consulted.

A worked example for our standard order. A lot of 500 pieces at General Inspection Level II returns code letter H, which is a sample of 50 units. At AQL 2.5 for major defects the pair is accept 3 and reject 4, so three majors in the sample passes and four fails. At AQL 4.0 for minor defects the same 50 units carry accept 5 and reject 6. Critical defects are not sampled at an AQL at all; the plan is zero acceptance, and a single critical defect fails the lot regardless of what the major and minor counts did.

The inspection level is the second lever and it moves the sample harder than the AQL does. General Inspection Level I draws roughly 40% of the Level II sample, and Level III draws roughly 60% more. A vendor quoting AQL 1.5 at Level I is running a plan that is looser in absolute terms than AQL 2.5 at Level II, because the smaller sample with a zero-or-one accept number discriminates less than a larger sample with a higher accept number. When you compare two inspection offers, compare the level first and the AQL second.

What Are Critical, Major and Minor Defects, and Who Defines Them?

The classification is the part of an inspection that decides the outcome, and it is the part most often left to the inspector on the day. An AQL number with no written classification attached is not a specification; it is an invitation to argue. The definitions below are the ones used on bag programmes, and they belong in the purchase order rather than in the inspector notebook.

Who defines them is the question that matters, and the answer should be the buyer. A vendor-written classification drifts toward classifying defects down, because a defect reclassified from major to minor moves it from the 2.5 column to the 4.0 column and frequently from a fail to a pass. Write the classification yourself, with worked examples from your own product, and attach it to the purchase order. It takes an afternoon and it is worth more than the AQL number it sits next to.

Why Does Defect Classification Matter More Than the AQL Number?

Because classification moves a defect between columns and the AQL only sets where the column boundary sits. Take a 500-piece lot with 50 units drawn and four defects in the sample. If all four are major, the lot fails at AQL 2.5 and fails at AQL 1.5. If two are classified minor, the lot passes at AQL 2.5 and fails at AQL 1.5. If three are classified minor, it passes at both. The verdict moved twice and the physical bag never changed.

This is why a vendor quoting a tighter AQL is not automatically offering a stricter inspection. A plan at 1.5 with a generous classification and a Level I sample can pass a lot that a plan at 2.5 with a written classification and a Level II sample would fail. The three levers — level, classification, AQL — have to be read together, and the AQL is the weakest of the three. It is also the only one that gets quoted, which tells you something about what quoting it is for.

What Does the AQL 1.5 Claim Actually Mean for a Buyer?

sunteam publishes AQL 1.5, and it is worth taking that at face value rather than dismissing it. The question is what a buyer receives from it. On the standard bag order the answer is: the same 50 units drawn, one fewer defect tolerated, a higher probability that the lot fails, and — because the plan is statistically tighter — a higher probability that a lot of perfectly ordinary quality is rejected by chance alone.

That last point is the one that costs money and it has a name worth knowing. Every sampling plan has an operating characteristic curve: the probability that a lot with a given true defect rate will be accepted. Tightening the AQL shifts that curve left and it makes the middle of the curve steeper, which sounds beneficial and is not. In the steep middle zone, two inspectors drawing two different samples from the same lot reach different verdicts. That is the dispute zone, and a tighter AQL widens it. Disputes cost re-inspection days, and re-inspection days cost the booking.

The second cost is the one that never appears on a quote. A tightened accept number makes the switch to reduced inspection unreachable for most programmes, and reduced inspection is the mechanism that cuts the sample to roughly 40% of normal after a clean run. A programme that fails occasionally never earns it. So the buyer pays for the tighter number twice: once in the unit price, where the vendor prices in the expected re-work, and once in the calendar, where the re-inspections land.

What Happens When a Lot Fails, and Who Decides What Next?

A failed lot is not a rejected order. It is a decision point with three possible outcomes, and the outcome should be pre-agreed in the contract rather than negotiated in the week it happens. The three are: re-work and resubmission, a buyer-signed waiver, or rejection. Each has a cost and each has an owner, and the argument usually starts because nobody wrote down which one applies to which defect class.

The waiver deserves the most attention because it is the one buyers sign without reading. A waiver is a legitimate commercial decision: the launch date matters more than three major defects in a sample, and shipping is the right call. It is only safe when it is written, dated and signed by the buyer, and when it records the defect count and class. An unwritten waiver is how a failed lot becomes an accepted one with nobody having decided anything, and it is the document that decides a later claim.

Does a Tighter AQL Actually Deliver Better Cartons?

Not by itself, and this is the part of the argument that the whole industry gets wrong. An acceptance sampling plan is a filter, not a control. It decides whether to accept what has already been made. It does not change what gets made. A line running at a 2% major defect rate will produce some lots that pass at 1.5 and some that fail, and over a year the cartons arriving at your warehouse will look the same as they would at 2.5, with more of the borderline lots sent back for re-work.

What actually improves cartons happens before the final inspection and it is unglamorous: a first-off check at the start of the run, in-line quality gates at defined points during it, and a corrective action on the cause of any defect class that appears twice. Those three reduce the process average. Lowering the accept number does not touch the process average at all — it only changes which side of the line the current average falls on. This is the same argument that reorder consistency makes about drift: the control is in the process and the document, not in the verdict at the end.

There is a place for a tighter level, and it should be named honestly. If a product carries a safety consequence, if a defect is undetectable until the customer finds it, or if a customer contract imposes a lower outgoing quality level, then 1.5 or tighter is the right answer and the cost is justified. A branded bag with a warranty and a return path does not meet any of those three tests, which is why 2.5 is where the industry settled and why it is where our programmes sit.

What Are the Switching Rules, and Why Do They Matter?

ISO 2859-1 is a continuing-series scheme with three states, and it moves between them on the recent history of lots. Ignoring the switching rules is the most common way a well-written AQL clause quietly stops working, because a programme that never leaves normal inspection is missing both the tightening that should follow failures and the reduction that should follow clean runs.

For a buyer the practical consequence is one sentence: ask which state the programme is in. A report written under reduced inspection draws a much smaller sample, and a smaller sample with the same AQL is a weaker statement about the lot, not a stronger one. A report written under tightened inspection tells you the previous lots failed, which is information you want before you accept this one. Neither is visible from the AQL number alone.

When Is AQL 4.0 the Right Choice?

For minor defects on almost every bag programme, which is the standard position and worth stating plainly: 4.0 for minors, 2.5 for majors, zero for criticals. The reason is that a minor defect is by definition one a customer would not return the bag over. Holding minors to 2.5 means rejecting lots over untrimmed thread ends, and the cost of that re-work lands in the unit price or in the calendar, neither of which the customer who would not have returned the bag is asking you to spend.

AQL 4.0 as the single level for everything is a different decision and usually the wrong one, because it removes the major class from the plan. Majors and minors are not the same thing commercially: a major generates a return, a minor generates a shrug. A plan that treats them identically is a plan that will pass a lot with a failed zipper if the sample happens to be kind. The three-class structure exists because the consequences differ, and collapsing it is the most common quiet loosening in the industry.

There is a legitimate case for 6.5 on specific low-value, high-volume items where the defect is cosmetic and the unit economics cannot carry a re-work cycle. It is not the case for a branded bag sold at retail with a warranty. If a vendor proposes 6.5 for a product of that kind, the question to ask is what the defect classification says, because a loose AQL paired with a strict classification is often the honest version of the same plan.

How Should AQL Be Written Into the Purchase Order?

In six lines, in the body of the order rather than in a quality appendix that nobody reads, because the appendix is where a classification goes to be quietly reinterpreted. The lines are short and they remove every argument that an AQL number on a quote leaves open.

1. Name the standard and the scheme. Sampling by attributes to ISO 2859-1, single sampling, normal inspection at the start of the programme. State it if double sampling is agreed instead.
2. Name the inspection level. General Inspection Level II. Say it explicitly, because Level I with a tight AQL is a looser plan than Level II with a moderate one.
3. Name the level per class. Zero acceptance for critical defects, AQL 2.5 for major defects, AQL 4.0 for minor defects.
4. Attach the defect classification as a schedule. A written list with worked examples from your own product, and a note that unlisted defects are classified by the buyer.
5. Name the switching state and who moves it. Start at normal; state who decides a move to tightened or reduced and on what evidence.
6. Name the consequence of a failure. Re-work and resubmission, a buyer-signed waiver, or rejection, with who pays for each and what it does to the delivery date.

MOQ

500 pcs per style / colourway (300 on a repeat)

Sampling

6-10 working days (12-15 for complex builds)

Bulk production

35-50 days from sample approval

Inspection standard

ISO 2859-1 (ANSI/ASQ Z1.4 equivalent), single sampling

Inspection level

General Inspection Level II

Levels by class

Critical 0 / Major AQL 2.5 / Minor AQL 4.0

Payment

T/T 30/70

Incoterm

FOB Xiamen (CIF / DDP available)

Operating since

2014

Frequently Asked Questions

People Also Ask