Bag quality inspection applying AQL levels and defect grading to leather bags before shipment

Bag Quality Inspection: AQL Levels, Defects and Grading

AQL inspection does not check every bag. It samples a population to a defined standard, which is why the standard must be agreed before the run.

Bag quality inspection using AQL means checking a defined sample from a production run against an agreed standard, rather than examining every unit in exhaustive detail. Under AQL 2.5 and ISO 2859-1, single sampling at general inspection level II, the inspector draws a number of units calculated from the batch size, examines them for defects, and compares the number of defects found against an acceptance number. If defects are within the limit, the batch passes; if not, it fails or is subject to further action.

This article explains how AQL inspection works, what the inspection levels mean, how defects are classified and graded, how to read a sampling table, and where internal release and third-party inspection differ. It is written for brand owners and sourcing managers who want to understand the report they receive and the standard they are agreeing to before production begins.

The single most important point about AQL is that it is a statistical method, not a guarantee of perfection. It is designed to accept a population that contains a small proportion of defects, and to reject a population that contains too many. That is why the standard, the defect classes and the acceptance limits are agreed in advance: they define what 'too many' means before anyone is tempted to decide it after seeing the goods.

Programme parameters at a glance

ParameterCommitment
Minimum order quantity500 pieces per style / colourway
Sampling lead time6–10 working days (12–15 days for complex constructions)
Bulk production lead time35–50 days after sample approval
Inspection standardAQL 2.5 (ISO 2859-1), single sampling, general level II
IncotermsFOB Xiamen — EXW, CIF and DDP available on request
Payment termsT/T 30% deposit / 70% before shipment; L/C at sight for container loads
Verified production floor space4,950 m² (SGS-verified)
Team137 staff across 7 production lines
Equipment149 machines including computerised saddle-stitch and edge-painting units
Monthly capacity200,000 pieces
EntityFounded 2014; principal has worked in bag development since 2004

What AQL Inspection Is and How It Works

AQL stands for Acceptance Quality Limit, and the method is drawn from ISO 2859-1, the international standard for sampling inspection by attributes. The principle is that inspecting a correctly calculated random sample gives a reliable view of the whole batch. The inspector selects units at random, examines them against specified criteria, records each defect found, and the batch is judged by whether the number of defects exceeds the acceptance number for that sample size.

ElementWhat it definesWhy it is agreed in advance
AQL valueThe quality level the sampling plan acceptsIt sets how strict the inspection is
Sampling levelHow many units are drawn from the batchIt balances inspection cost against assurance
Sample sizeThe number of units inspectedIt follows from batch size and level
Defect classesWhich faults are critical, major or minorIt removes interpretation from the inspection
Acceptance numberThe maximum defects allowed for a passIt makes pass or fail objective
Inspection criteriaThe measurements and standards checkedIt ensures the approved sample is the reference

The method checks attributes rather than measuring every dimension on every unit. An attribute is a pass-or-fail judgement on a specific feature: is the stitch line within tolerance, is the hardware correct, is the edge finish uniform. This is what makes AQL practical at volume, and it is why the technical pack's tolerances matter — an inspector can only apply a standard that has been written down.

AQL Levels and Sampling Plans

AQL defines several inspection levels that determine how much of a batch is examined. General levels I, II and III represent increasing scrutiny, and level II is the default for most consumer goods, including bags. Special levels S-1 to S-4 are used where the cost or time of inspection must be reduced, at the price of less assurance.

LevelScrutinyTypical useTrade-off
General IReduced sample sizeLow-value or low-risk goodsLess assurance for lower inspection cost
General IIStandard sample sizeMost bag programmesThe accepted default balance
General IIIIncreased sample sizeHigh-value or high-risk goodsGreater assurance at higher inspection cost
Special S-1 to S-4Small sample sizesDestructive or expensive testingMuch lower assurance; used for specific checks
Tightened inspectionApplied when a supplier's quality slipsFollowing failures on previous batchesImposed to correct a declining trend
Reduced inspectionApplied when quality is consistently goodTrusted, stable supplyOffered as a reward for consistent results

Level II is the standard because it balances the cost of inspection against the assurance it provides, and it is the level most third-party agencies apply by default. Moving to level III increases the sample and therefore the assurance, which is justified for high-value runs or for goods going to demanding markets. Reducing below level II, or using a special level, saves inspection time but weakens the guarantee.

Quality inspector sampling leather bags for AQL inspection on a production floor
AQL draws a random sample from the batch; the sample size follows from the batch size and the chosen inspection level.

Classifying Defects: Critical, Major and Minor

Before inspection can be objective, every type of fault has to be assigned to a class. The three standard classes are critical, major and minor, and the class determines how a fault counts against the acceptance number. Classifying defects in advance is the work that makes an AQL inspection meaningful, because without it the inspector is applying personal judgement rather than an agreed standard.

ClassDefinitionBag examplesTypical acceptance
CriticalA defect that could harm the user or breach a regulationSharp metal protruding; prohibited substance presentUsually zero tolerance
MajorA defect likely to cause failure or customer rejectionStructural seam failure; wrong or broken hardwareCounted against a strict limit
MinorA defect unlikely to reduce function but noticeableSlight unevenness in edge paint; minor stitch irregularityCounted against a looser limit
Not a defectA natural characteristic of the materialGrain variation and growth marks within toleranceNot counted at all
Cosmetic toleranceA defined acceptable range for appearanceShade variation within an agreed rangeAccepted within the stated tolerance

The 'not a defect' row is the one that causes the most disputes, and it is where leather differs from manufactured material. Natural grain variation, growth marks and slight shade differences are characteristics of full-grain leather, not faults, provided they fall within an agreed tolerance. Agreeing that tolerance before the run is what prevents a genuine characteristic from being treated as a defect at inspection, or a genuine defect from being excused as natural.

Reading an AQL Sample Table

An AQL sampling table converts a batch size and an AQL value into a sample size and an acceptance number. For a given lot, the table states how many units to inspect and the maximum number of each defect class that may be found for the batch to pass. Larger lots require larger samples, and stricter AQL values reduce the number of allowed defects.

Lot size (units)Sample size (level II)AQL 2.5 major — acceptAQL 2.5 major — reject
151–2803223
281–5005034
501–1,2008056
1,201–3,20012578
3,201–10,0002001011
10,001–35,0003151415
35,001–150,0005002122

The table shows why the standard has to be agreed rather than assumed. A batch of 500 units calls for 50 units to be inspected, and would pass at AQL 2.5 with up to 3 major defects but fail at 4. That margin is narrow, and it is exactly the margin a brand is accepting when it agrees the standard. Reading the table before the run shows what is actually being agreed, rather than discovering it in the report.

Where and When Inspection Happens

AQL inspection is the last gate, but it works best when it is not the only one. Inspection typically runs at three points: first-article inspection at the start of the run, in-process checks at intervals during production, and final or pre-shipment inspection on the completed batch. The earlier checks reduce the number of defects that reach the final gate, which is where rework is most expensive.

  • First-article inspection confirms the first units of the run match the full written specification before volume begins.
  • In-process inspection checks stitch density, alignment and finish at intervals, catching drift while it can still be corrected.
  • Final inspection applies AQL 2.5 under ISO 2859-1 to the completed batch, using an agreed random sample.
  • Pre-shipment inspection confirms quantity, specification, packing, labelling and carton configuration before dispatch.
  • During-production inspection can be used where a run is long enough that correcting mid-run is worth an inspector's visit.
  • Container loading supervision confirms the goods loaded match the inspected and released batch.

Where a run is long, inspection during production is often better value than inspection only at the end, because a defect trend caught early can be corrected across the remaining units rather than reworked on completed goods. Final inspection then confirms the correction held. The gates work as a sequence rather than as alternatives.

Grading and the Inspection Report

An inspection produces a report rather than a simple verdict. The report records the sample size, the defects found by class, photographs of each defect and a calculation of whether the batch passes. A batch that fails may be rejected outright or offered for rework, depending on the nature of the defects and what the two parties have agreed in advance.

Batch resultPass, fail or pass with observation, calculated against the acceptance number for the agreed standard
Defect recordEach defect logged by class, with its location on the unit and a photograph
Defect rateThe proportion of defective units in the sample, used to compare runs over time
Measurement recordCritical dimensions checked against the technical pack's tolerances
Reference checkConfirmation that the units match the signed approved sample
Corrective actionWhat will change before the next batch, where a trend is emerging
DispositionWhether the batch ships, is reworked, or is sorted to a revised standard

The report's value over time is as much in its trends as in its verdict. A run that passes with a rising count of a particular minor defect is telling the brand something about the line, and catching that trend across successive reports is easier than waiting for the defect to become major. Grading a single batch is a snapshot; the sequence of reports is a picture of how the programme is behaving.

Bag quality inspection report grading defects on inspected leather bags by class
An inspection report records defects by class and photographs each one; over successive runs its trends matter as much as its verdict.

Common Failures and How They Are Handled

When a batch fails inspection, the response depends on the defects found. Some failures are resolved by reworking the affected units; others require full sorting or, in the worst case, production of replacement goods. Agreeing in advance how each type of failure will be handled prevents a difficult conversation at the moment when goods are waiting to ship.

Failure typeIllustrationUsual response
Critical defect foundA sharp edge or prohibited substance detectedBatch held; root cause corrected before any shipment
Major defect over limitToo many structural or hardware faults in the sampleRework affected units, re-inspect, or reject the batch
Minor defect over limitCosmetic faults exceeding the agreed limitSorting, touching up or rework under an agreed plan
Wrong specificationA component or material differs from the approved sampleCorrected and re-inspected before release
Measurement out of toleranceA critical dimension outside the technical pack rangePattern or process corrected; units reworked or remade
Packing faultCarton configuration or protection not to specificationRepacked to the agreed specification before dispatch

The wrong-specification row is the one most easily prevented and most annoying when it occurs, because it usually traces to an unnamed component or a superseded revision. Naming every component by part number and controlling the technical pack's revision is what keeps the production run aligned with the batch that was inspected and released.

Third-Party Inspection vs Internal Release

Inspection can be performed internally by the production partner or independently by a third-party agency. Both have a role, and they are not alternatives. Internal release is the everyday control that keeps a line producing to standard; third-party inspection is independent verification, most valuable on first runs and on shipments to markets where a brand cannot easily check goods itself.

AspectInternal releaseThird-party inspection
Who performs itThe production partner's quality teamAn independent inspection agency
IndependenceInternal; closely involved in the runIndependent of the production partner
CostPart of productionCharged per inspection visit
Best used forEveryday control and in-process checksFirst runs and high-value or distant shipments
SpeedImmediate; runs continuouslyRequires scheduling; often within days
StrengthDeep familiarity with the line and the styleIndependent verification a brand can rely on at distance
LimitationNot independent of the producerLess familiar with the specific style and its history

Many programmes use both: internal release as the continuous control and third-party inspection as independent confirmation before shipment. Where an inspection standard is agreed with defect classes written down and the approved sample present, internal and third-party inspection should reach similar conclusions — and where they do not, that divergence is itself worth investigating.

Set your inspection standard before production

Send your quality criteria, defect definitions and target markets to service@junyuanbags.com, or message the programme team on WhatsApp at +86 177 5002 0688. We will set out an inspection plan with defect classes, sampling level and acceptance limits for your bags.

Frequently Asked Questions

What is AQL inspection for bags?

AQL inspection means checking a defined random sample from a production batch against an agreed standard rather than examining every unit. Under AQL 2.5 and ISO 2859-1, single sampling at general level II, the inspector draws a sample calculated from the batch size, records any defects found, and compares the count against an acceptance number. Within the limit the batch passes; beyond it, it fails or is reworked.

What does AQL stand for and where does it come from?

AQL stands for Acceptance Quality Limit, and the method is drawn from ISO 2859-1, the international standard for sampling inspection by attributes. AQL values and inspection levels originate in military and industrial quality standards and are now applied widely to consumer goods. The principle is that a correctly drawn random sample gives a reliable view of the whole batch without inspecting every unit.

What is the difference between general level I, II and III inspection?

The general levels determine how many units are drawn from a batch. Level I inspects a reduced sample for lower-value or low-risk goods; level II is the standard default and balances cost against assurance; level III inspects an increased sample for high-value or high-risk goods. Level II is what most third-party agencies apply to consumer goods, including bags.

How many bags are inspected in an AQL 2.5 check?

The sample size follows from the batch size and the inspection level. At general level II, a lot of 151–280 units requires a 32-unit sample, 281–500 requires 50 units, 501–1,200 requires 80, and 1,201–3,200 requires 125. Larger lots require larger samples, and the acceptance number for each defect class rises with the sample size.

What are critical, major and minor defects in bag inspection?

Critical defects could harm the user or breach a regulation, such as sharp metal or a prohibited substance, and usually carry zero tolerance. Major defects are likely to cause failure or customer rejection, such as structural seam failure or wrong hardware. Minor defects are unlikely to reduce function but are noticeable, such as slight unevenness in edge paint, and they count against a looser limit.

Is natural grain variation a defect?

No, provided it falls within an agreed tolerance. Natural grain variation, growth marks and slight shade differences are characteristics of full-grain leather rather than faults. Agreeing the tolerance for these characteristics before the run is what prevents a genuine feature from being treated as a defect at inspection, or a genuine defect from being excused as a natural characteristic.

When should AQL inspection be carried out?

Inspection works best as a sequence of gates rather than a single check. First-article inspection confirms the first units match the specification, in-process checks catch drift during production, final inspection applies AQL 2.5 to the completed batch, and pre-shipment inspection confirms quantity, packing and labelling. Catching a defect trend early allows correction across the remaining units rather than rework on finished goods.

What happens if a bag batch fails inspection?

The response depends on the defects found. A critical defect holds the batch until the root cause is corrected. Major defects over the limit typically mean reworking affected units and re-inspecting, or rejecting the batch. Minor defects over the limit may be resolved by sorting or agreed touch-up. Agreeing how each type of failure will be handled in advance prevents disputes at the point when goods are waiting to ship.

What is in an AQL inspection report?

A report records the sample size, the defects found by class, photographs of each defect and the calculation of whether the batch passes. It also includes a measurement record against the technical pack's tolerances and confirmation that the units match the signed approved sample. Over successive runs, the trends in the reports matter as much as any single verdict.

What is the difference between internal release and third-party inspection?

Internal release is performed by the production partner's quality team and forms the everyday control that keeps a line producing to standard. Third-party inspection is independent verification by an agency with no stake in the run, most valuable on first runs and on shipments to distant or demanding markets. Many programmes use both, with internal release as continuous control and third-party inspection as confirmation before shipment.

Why must the inspection standard be agreed before production?

Because agreeing it afterwards means deciding what counts as a defect in the knowledge of what was found, which is not a standard at all. Setting the AQL value, sampling level, defect classes and acceptance limits in advance means the inspector is applying an agreed rule rather than personal judgement. The technical pack's tolerances and the approved sample provide the objective reference.

Does AQL inspection guarantee perfect quality?

No. AQL is a statistical method that accepts a population containing a small proportion of defects and rejects one containing too many. It is designed to give reliable assurance at reasonable cost, not to eliminate defects entirely. That is precisely why the standard and the defect classes are agreed in advance, so that the acceptable level is defined before anyone is tempted to decide it after seeing the goods.

What is first-article inspection?

First-article inspection checks the first units produced at the start of a run against the full written specification. It confirms that patterns, materials and settings are correct before the run commits to volume, and it is far cheaper than discovering a fault after hundreds of units have been made. It is the first of the gates, with in-process and final inspection following.

Can inspection be done during production rather than at the end?

Yes, and on long runs it is often better value. During-production inspection checks stitch density, alignment and finish at intervals, catching a trend while it can still be corrected across the remaining units. Final inspection then confirms the correction held. Inspecting only at the end discovers problems after their cost has already been incurred across the whole batch.

How is the approved sample used in inspection?

The signed approved sample is the physical reference against which units are compared. It resolves questions that a written specification cannot, such as subtle differences in finish or hand, and it gives both the producer and the inspector the same object to consult. Keeping it on the line during production and present at inspection is what makes the comparison consistent.

What is tightened and reduced inspection?

Under ISO 2859-1, inspection can be tightened or reduced according to a supplier's track record. Tightened inspection applies when quality has slipped, increasing the sample to catch problems and correct a declining trend. Reduced inspection is offered when quality has been consistently good, lowering the inspection burden as a reward for reliable results. Both adjust scrutiny without changing the underlying standard.