AQL 2.5 does not mean that 2.5 per cent of a bag order may be defective. It means that the order is inspected by sampling under ISO 2859-1 and that the number of defects found in the sample is compared against an acceptance threshold set at 2.5 for the relevant defect class. The 2.5 is a point on a sampling plan, not a defect allowance, and understanding that distinction is what separates a well-run inspection from an argument.
This guide explains what the acceptable quality limit actually controls, where AQL 2.5 sits within the ISO 2859-1 framework, how sampling plans and inspection levels determine how much is checked, how defect classes combine to produce an accept or reject decision, and the misreadings that cause buyers and partners to talk past each other. It is written for brands specifying an inspection standard and for anyone who has been told that AQL 2.5 "allows" a certain number of faults.
The programme applies AQL 2.5, single sampling, general level II, as its standard. That combination is a complete instruction: it fixes the sampling scheme, the acceptance criteria and the defect classifications, so an inspection produces a decision either party can trace back to the rule rather than to an opinion about a photograph.
Programme parameters at a glance
| Parameter | Commitment |
|---|---|
| Minimum order quantity | 500 pieces per style / colourway |
| Sampling lead time | 6–10 working days (12–15 days for complex constructions) |
| Bulk production lead time | 35–50 days after sample approval |
| Inspection standard | AQL 2.5 (ISO 2859-1), single sampling, general level II |
| Incoterms | FOB Xiamen — EXW, CIF and DDP available on request |
| Payment terms | T/T 30% deposit / 70% before shipment; L/C at sight for container loads |
| Verified production floor space | 4,950 m² (SGS-verified) |
| Team | 137 staff across 7 production lines |
| Equipment | 149 machines including computerised saddle-stitch and edge-painting units |
| Monthly capacity | 200,000 pieces |
| Entity | Founded 2014; principal has worked in bag development since 2004 |
What AQL 2.5 Does and Does Not Mean
The single most persistent misreading of AQL is the belief that the number is a defect allowance. It is not. AQL stands for acceptable quality limit, and it sets a threshold against which a sample result is judged. The figure belongs to a sampling plan that also specifies how many pieces to inspect and how to classify what is found.
The practical consequence is that a buyer should never accept the sentence "AQL 2.5 means 2.5 per cent defects are allowed". What it means is that the inspection will draw a specified sample, count defects by class, and compare the counts against the acceptance numbers for those classes at the 2.5 threshold. Everything that follows in an inspection report depends on those three elements being stated.
Where AQL 2.5 Comes From
AQL sits inside a family of international standards for inspection by attributes, most commonly ISO 2859-1, which is technically aligned with the long-established MIL-STD-105 lineage. The standard does not prescribe a single number; it provides a system of tables from which a sampling plan is selected according to the order size, the inspection level and the chosen AQL.
| Element | What it fixes | Why it matters |
|---|---|---|
| The standard (ISO 2859-1) | The sampling schemes, tables and switching rules | Gives the method a defined basis rather than a vendor's habit |
| The AQL value (2.5) | The acceptance threshold for a defect class | Determines how strict the test is, not how many faults are permitted |
| The inspection level (general level II) | How much is sampled relative to the order size | A higher level samples more; level II is the customary general default |
| The sampling scheme (single sampling) | Whether one, two or multiple samples are drawn | Single sampling reaches a decision from one sample, which is simpler to administer |
| The defect classes | How findings are categorised and counted | Prevents a cosmetic mark being treated as equivalent to a structural failure |
Because the standard is a system, naming only the AQL leaves the plan incomplete. "AQL 2.5" on its own does not tell an inspector how much to sample or how to classify a defect. The full instruction is AQL 2.5, single sampling, general level II, and it is that complete phrase that should appear in a purchase order or an inspection brief.
Sampling Plans and Inspection Levels
The relationship between order size and sample size is the heart of AQL. The standard provides three general inspection levels, and the sample drawn scales with the lot size within each level. A larger order does not mean a proportionally larger sample; the sample grows more slowly than the lot, which is why AQL is efficient at scale.
| Inspection level | Relative sample size | When it is used |
|---|---|---|
| General level I | Smallest sample | Used when a lower confidence is acceptable, such as for a well-established, low-risk product |
| General level II | Standard sample | The customary default for general inspection, and the level this programme applies |
| General level III | Largest general sample | Used when higher confidence is required, such as for a critical or high-value order |
| Special levels S-1 to S-4 | Smallest samples overall | Used for destructive or very costly tests, trading confidence for feasibility |
Two consequences follow. First, the sample size is a function of both the lot size and the level, so quoting an AQL without the level is incomplete. Second, raising the inspection level is a way to increase confidence without changing the AQL threshold, which is often a more sensible adjustment than tightening the AQL itself when a buyer wants more assurance.
Single sampling, which the programme uses, draws one sample and reaches a decision from it. Double and multiple sampling schemes exist and can reduce the average amount inspected, but they add steps and complexity. For bag inspection, single sampling is the clear default because the decision is easier to explain and the report easier to read.
Defect Classes and How They Combine
AQL decisions rest on how defects are classified, because the classes are counted separately against their own thresholds. Classifying a defect wrongly therefore changes the outcome of the test, which is why the classes should be agreed before production rather than argued about after a report is issued.
| Class | Definition | Typical examples on a bag | Effect on the decision |
|---|---|---|---|
| Critical | Unsafe or unusable; a fundamental failure | A closure that will not secure, a strap that detaches under normal load | Any critical defect normally fails the inspection outright |
| Major | Materially affects appearance or function | A structural stitch failure, a wrong hardware component, a significant dimensional deviation | Counted against a strict major threshold; a few can fail the lot |
| Minor | Small cosmetic imperfection without functional impact | A light surface mark, an uneven edge on an internal panel | Counted against a more permissive threshold than majors |
The combination rule matters. An inspection is judged on all three counts together, and a lot can fail on majors even when the minor count is negligible, or on criticals regardless of the other counts. This is why a report that gives only a total defect figure is nearly useless: it conceals the classification that actually decided the result.
Two bags can carry identical defect counts and produce opposite decisions if the defects fall into different classes. That is the point of classification. It ensures that a cosmetic blemish, which a customer may never notice, is not weighed the same as a structural fault that will bring a bag back as a return. Agreeing the classes in advance is the single most useful preparation a buyer can make.
Reading an AQL Table: A Worked Example
An AQL table converts a lot size and an inspection level into a sample size and a set of acceptance numbers. Reading it is straightforward once the structure is familiar, and the example below shows the shape of the reasoning without pretending to reproduce the full standard.
| Step | What you look up | What you get | How it is used |
|---|---|---|---|
| 1. Lot size | The number of units in the order | A lot-size band | Places the order in the correct row of the table |
| 2. Inspection level | General level II | The code letter for that band | Selects the sample size for the lot |
| 3. Sample size | The code letter | The number of pieces to inspect | Defines how much of the order is examined |
| 4. Acceptance numbers | The AQL value for each defect class | An accept and a reject number per class | Compare the defect counts to these numbers |
| 5. Decision | The counts from the inspection | Accept or reject | Met if counts are within the acceptance number for each class |
The critical insight is at step four: the acceptance number is not a percentage applied to the order. It is a whole number that belongs to the intersection of the sample size and the AQL. A larger sample carries a larger acceptance number and a smaller one carries a smaller number, which is why the same AQL can produce different tolerances on different orders. Any discussion of tolerance that omits the sample size is incomplete.
Common Misreadings of AQL 2.5
Most disagreements about AQL come from a small set of misreadings, each of which can be avoided by stating the full plan and agreeing the classes before production. Naming the misreadings is the quickest way to prevent them.
- Treating 2.5 as a defect allowance rather than a threshold on a sampling plan.
- Quoting an AQL without the inspection level or the sampling scheme, leaving the plan incomplete.
- Reading a total defect count without separating critical, major and minor classes.
- Assuming a larger order permits proportionally more defects, when the sample grows more slowly than the lot.
- Assuming a passed lot means every unit conforms, when the decision rests on a sample.
- Applying a general-purpose AQL to a component where a stricter class is warranted, such as a load-bearing strap.
- Disagreeing about a result without checking whether the classification of a defect matched the agreed classes.
Each misreading has the same remedy: write the full inspection instruction into the order, agree the defect classes in advance, and read the report against the plan rather than against an impression. A buyer and a partner who share that discipline can disagree about a result and still resolve it, because both are working from the same rule.
AQL in Practice on a Bag Line
On a bag programme, AQL 2.5, single sampling, general level II is applied at final inspection, with the approved sample present as the reference and functional checks covering the elements a visual pass would miss. The standard defines the statistics; the practical work is in making the inspection thorough enough that the statistics mean something.
It is worth restating where this sits in the wider process. Bulk production runs 35 to 50 days after sample approval, and the final inspection falls at the end of that window, before shipment. Because the balance is due before shipment, the inspection is the buyer's last and strongest point of verification, and an inspection standard that is stated completely is what makes that verification reliable.
Beyond AQL: When Sampling Is Not Enough
AQL is a powerful default, but it is a sampling method, and sampling is not always the right tool. Where a defect would be especially damaging, or where the order is small enough that inspecting everything is practical, a stricter approach than AQL 2.5 may be warranted, and recognising those cases is part of using the standard well.
- For a high-value or high-visibility order, consider raising the inspection level to general level III for greater confidence at the same AQL.
- For a component where failure is unacceptable, such as a load-bearing attachment, apply 100 per cent inspection rather than relying on a sample.
- For a small order where the whole lot can be checked economically, full inspection can replace sampling entirely.
- For a first article after any change to materials or components, inspect the first units against the full specification before the run proceeds.
- For a repeat order, use the previous report's findings to direct attention at the defect classes that caused trouble last time.
None of this weakens AQL 2.5 as a default; it simply places it where it belongs. The standard is the right control for a general bulk order, and the cases above are the exceptions in which a stricter or wider control pays. Used together, sampling and targeted full inspection between them cover the practical range of risk in a bag programme, and both depend on the same discipline of stating the standard clearly and writing down what was found.
Set the right inspection standard for your order
Send your order details to service@junyuanbags.com or message the programme team on WhatsApp at +86 177 5002 0688. We will confirm the inspection plan as AQL 2.5, single sampling, general level II, agree the defect classes in advance and document the result so both parties can trace the decision to the standard.
Frequently Asked Questions
Does AQL 2.5 mean 2.5 per cent of bags may be defective?
No. This is the most common misreading. AQL stands for acceptable quality limit, and 2.5 is a threshold on a sampling plan, not a defect allowance. The inspection draws a defined sample under ISO 2859-1, counts defects by class, and compares those counts against acceptance numbers set at 2.5 for the relevant class. The 2.5 is a point on a table, not a permitted percentage of faulty goods.
What is the full instruction for an AQL inspection?
The complete instruction is AQL 2.5, single sampling, general level II. Naming the AQL value alone is incomplete, because it does not fix how much is sampled or how defects are classified. The sampling scheme, the inspection level and the defect classes together with the AQL define the plan, and it is the full phrase that should appear on a purchase order or an inspection brief.
Where does AQL come from?
AQL sits inside ISO 2859-1, a family of international standards for inspection by attributes, technically aligned with the long-established MIL-STD-105 lineage. The standard does not prescribe a single number. It provides tables from which a sampling plan is selected according to the order size, the inspection level and the chosen AQL, together with switching rules that govern continued application.
What is the difference between general level I, II and III?
The three general inspection levels determine how much is sampled relative to the order size. Level I draws the smallest sample, level II the standard sample and level III the largest general sample. Level II is the customary default and the level this programme applies. Raising the level to III increases confidence at the same AQL, which is often a better adjustment than tightening the AQL itself.
What are critical, major and minor defects on a bag?
Critical, major and minor are the three classes against which an inspection is judged. A critical defect is severe enough to make the bag unsafe or unfit for use, and its presence normally fails the lot outright. A major defect impairs appearance or function in a way a customer would notice, such as a failed stitch line or an incorrect fitting. A minor defect is a cosmetic flaw that leaves the bag fully usable. The classes are counted separately because each carries its own threshold.
Why are defects classified separately under AQL?
Because the classes are counted against their own thresholds, and a lot can fail on majors even when the minor count is negligible, or on criticals regardless of other counts. Classification ensures a cosmetic blemish is not weighed the same as a structural fault. Two bags can carry identical counts and produce opposite decisions if the defects fall into different classes.
Does a larger bag order allow more defects?
Not proportionally. The sample size grows more slowly than the lot, so a larger order does not permit a matching increase in defects. The acceptance number is a whole number tied to the sample size and the AQL, so it changes with the sample size rather than scaling with the order. Discussing tolerance without stating the sample size is therefore incomplete.
What is single sampling and why is it used?
Single sampling draws one sample and reaches a decision from it. Double and multiple sampling schemes draw more than one sample and can reduce the average amount inspected, but they add steps and complexity. For bag inspection, single sampling is the clear default because the decision is easier to explain and the report easier to read, which matters when both parties need to rely on the outcome.
Can an order pass inspection and still contain defects?
Yes. Because the decision rests on a sample, a passed lot means the sample met the acceptance criteria, not that every unit conforms. This is a property of any sampling method rather than a flaw in AQL. Where a defect would be especially damaging, a higher inspection level or 100 per cent inspection on the critical component is a more appropriate control.
When should a buyer use a stricter standard than AQL 2.5?
Use stricter or wider controls where a defect would be especially damaging or where full inspection is practical. Options include raising the inspection level to general level III for more confidence at the same AQL, applying 100 per cent inspection to a critical component such as a load-bearing attachment, or inspecting the whole lot where the order is small. None of these replaces AQL 2.5 as the general default.
How should the approved sample be used in an AQL inspection?
The signed approved sample is the reference against which every judgement is made, so the inspector is comparing goods to a fixed standard rather than to a fresh interpretation of the design. Where the physical sample cannot be present, a documented specification with a dimensional record and detailed photographs should stand in, and any deviation from the reference is the most important finding to record.
What should an AQL inspection report contain?
Read a report for four things: the sample size and the inspection level used, the defect counts broken out by class, the measured figures against the stated tolerance, and evidence of any deviation from the approved reference. A single overall verdict without those details cannot be traced back to the plan, and it is that traceability which makes the outcome defensible if a question is raised later.
How does AQL 2.5 interact with the payment terms?
The balance is due before shipment, and the final inspection at AQL 2.5 is what confirms the goods before that payment is released. The inspection is therefore the buyer's last and strongest point of verification, and an inspection standard stated completely is what makes the release reliable. Paying the balance before inspection gives up that protection.
Are there alternatives to AQL inspection?
Yes, and they are used alongside AQL rather than instead of it. Full or 100 per cent inspection examines every unit and suits small orders or critical components. First-article inspection checks the first units after any change to materials or components before the run proceeds. Periodic independent inspection maintains a baseline on a stable repeat programme without examining every shipment.
What causes most disagreements about an AQL result?
Most disagreements come from a small set of misreadings: treating 2.5 as a defect allowance, quoting an AQL without the inspection level or sampling scheme, reading a total defect count without separating the classes, or applying a general AQL to a component that warrants a stricter class. Writing the full instruction into the order and agreeing the defect classes in advance removes almost all of them.
How does AQL fit into the overall bag production timeline?
Bulk production runs 35 to 50 days after sample approval, and final inspection at AQL 2.5 falls at the end of that window, before shipment and the release of the balance. Because the inspection is scheduled ahead of shipment, it must be booked in advance rather than treated as a last-minute step. Its position in the calendar is what makes it a control rather than a formality.