AQL 1.5 vs 2.5 vs 4.0: Which Level Should You Specify?

The choice of AQL column is presented in quotations as though it were the main quality decision in the order. It is not. Once the lot size and inspection level are fixed, the entire difference between the tightest and loosest of these three columns is two units of tolerance on fifty. Picking one takes a minute; writing a defect classification that makes the column mean something takes an afternoon, and only one of those has much effect on what arrives.

What Actually Differs Between the Three Columns?

Only the accept and reject pair, and nothing else. The plan takes its sample size from the lot size and the inspection level, then reads one column to get the pair. Column 1.5 gives a lower accept number, column 4.0 a higher one, and the sample sitting underneath them is identical in every case.

Stated numerically, on the standard export lot sizes: 500 pieces draws 50 units, accepting 2 defects at 1.5, 3 at 2.5 and 5 at 4.0. At 2,000 pieces the sample is 125 and the accept numbers are 5, 7 and 10. At 10,000 pieces the sample is 200 and the numbers are 7, 10 and 14. The proportion of tolerance hardly moves as lots grow.

What Does Choosing AQL 1.5 Buy You?

One or two units of tolerance, and a considerably higher probability that a borderline lot fails. That is genuinely worth having when the cost of letting a defective unit through is much larger than the cost of failing a passable lot: high unit value, a retail chargeback regime, a safety-adjacent feature, or a defect that only surfaces once the product is in use.

What it does not buy is more inspection. This is the misunderstanding that drives most of the argument, in both directions. Buyers ask for 1.5 expecting more units to be opened; vendors quote it knowing the sample will not change. If more evidence is what you want, General Inspection Level III is the instruction that delivers it — at additional cost, which is why nobody offers it for free.

What Does Choosing AQL 1.5 Cost You?

Three things, all predictable. More lots fail on counts that would previously have passed, each failure costs a re-work cycle plus a full second inspection, and the programme finds it harder to reach reduced inspection. For a programme shipping monthly, those second visits and re-work cycles are recurring costs rather than one-off events.

There is a subtler cost on the switching rules. Tightened inspection is triggered by rejected lots, and a tighter accept number makes rejection more likely on the same underlying quality. A programme can therefore spend most of its life in tightened inspection, where the effective number is no longer the one written in the purchase order — an outcome nobody chose.

When Is AQL 4.0 the Right Answer?

For minor defects, on almost every general merchandise programme. That is the conventional pairing with major at 2.5, and it reflects what the class actually contains: faults a customer can see and would not return the product for. Demanding 2.5 on that class means failing lots over untrimmed thread ends, which trains everybody to stop recording them.

There are cases for 4.0 on the major column too, and they are honest trade-offs rather than shortcuts: promotional product where units are given away rather than sold, a first exploratory order where the priority is learning rather than perfection, or a product whose defects are immediately visible to the buyer at receipt and can be sorted at destination.

How Do the Three Levels Apply Within One Inspection?

Simultaneously, on the same sample, one column per class. Standard practice on bag programmes is zero acceptance for critical defects, 2.5 for major defects and 4.0 for minor defects. Each class is counted separately and each can fail the lot independently, so the plan is three verdicts rather than one.

This three-column arrangement is why "we inspect to AQL 2.5" is an incomplete answer to any question. Two vendors writing the same sentence can be running quite different plans: one might hold minors at 4.0 and the other at 2.5, or apply 2.5 to every class including critical. The complete answer names all three.

Does the Product Category Change the Answer?

It changes it more than most buyers expect, because the same column buys different assurance on different products. Duty is the deciding variable. A bag opened several hundred times a day has far more opportunities for a functional fault to surface than one opened twice a year, and a product whose purpose is protection cannot tolerate the faults a decorative product can.

What Role Does Unit Value Play?

A large one, and it cuts in surprising directions. High unit value argues for a tighter column, because the cost of a returned unit is high relative to the cost of a failed lot. But it also argues for more units examined, because a 50-unit sample tolerances of three defects says nothing precise about a 2,000-piece lot of expensive product.

Where both objections are live, the answer is usually not a single column but a layered plan: 2.5 for major defects, a named critical list covering the features whose failure carries the cost, and either a higher inspection level or a supplementary full check on the costly feature. Column choice alone cannot carry that responsibility.

How Does the Level Interact With the Switching Rules?

Directly, because the switching rules key on rejections. A tighter column produces more rejections at identical quality, so the programme reaches tightened inspection sooner and returns to normal later. Once tightened, the accept pair drops again — from an already tight starting point.

The practical consequence is that 1.5 should not be specified casually. If it is specified, agree at the same time which state the programme starts in and how often the state is reviewed, and require it on every report. Otherwise the number in the contract and the number in use quietly diverge.

How Do You Compare Two Suppliers Quoting Different Levels?

Normalise them first, then compare what is left. Ask each for the sample size at your standard lot quantity, the column applied to each class, and the switching state they expect the programme to sit in. Those three answers rebuild the plan and make two quotations comparable; the AQL number alone does not.

Then look past the plan to what actually decides outcomes: the defect classification, who wrote it, and whether re-work after a failure is quoted with dates. A programme with a well-written classification at 2.5 delivers better cartons than one offering 1.5 with a list written by the party whose margin depends on shipping today.

One further consideration settles most borderline cases. Ask what the tighter column would change if the same faults occurred: if the plan would still have accepted the lot, the level is not your problem and the classification is. If it would have failed one or two lots a year that then went on to generate complaints, the tighter column has just paid for itself. That arithmetic is more useful than any argument about numbers.

What Should You Write in the Specification?

Five lines, and the column decision is only one of them. Written once and repeated on every purchase order, they remove the arguments that otherwise consume the afternoon of every inspection.

Where the column choice is still open, it is worth reading the reasoning in full rather than deciding from the table alone: AQL 1.5 vs 2.5 vs 4.0 sets out what each level does and does not buy, and the sampling table lets any combination be checked against your real lot quantity.

Two closing points save the most trouble. First, decide the level before the first lot and then leave it alone for at least three shipments: changing mid-programme makes trend data useless and invites an argument about which lots the change applies to. Second, record it as part of the order rather than as a preference — an instruction nobody disputes later is worth more than a column that is one unit stricter.

Frequently Asked Questions

People Also Ask