Two teams can look at the same pallet of finished goods & reach opposite conclusions. One sees a carton with a scuffed lid & calls the shipment acceptable. The other sees the same carton & holds the whole lot. Neither side is acting in bad faith. They simply have no shared definition of what counts as a fault, or how many faults are too many.
Acceptable quality limit, or AQL, exists to remove that argument. It gives a brand & a factory a common sampling method & a common threshold, agreed before production starts, so the decision at the dock is a calculation rather than a negotiation. This piece explains how AQL works for cosmetic goods, which defect classes matter in this category, & what to write into the quality agreement so the inspection actually protects the launch.
What AQL actually sets
AQL is not a promise that every unit is perfect, & it is not a target defect rate. It is the worst tolerable process average when a continuing series of lots is submitted for acceptance sampling. In plain terms, it fixes two things: how many units the inspector opens, & how many faults inside that sample will still let the lot pass.
Those two numbers come from a standard table indexed by lot size & inspection level. A shipment of 5,000 units at general inspection level II draws a specific sample size, & that sample carries a fixed accept & reject count for each chosen AQL value. The inspector follows the table. Nobody argues about whether the sample felt large enough.
The useful consequence for a brand is that quality expectations become a written number. "We want good quality" cannot be audited. "Critical defects at AQL 0, major at 1.0, minor at 2.5, general level II" can be checked line by line against the inspection report.
Sampling levels & lot sizes
Inspection levels run from a light sample to a heavy one. Level I draws fewer units & is cheaper per inspection; level III draws more & catches more. General level II is the common default for finished cosmetic goods, with level III used when a line is new, when the previous lot failed, or when the launch carries a retailer deadline.
Lot size drives the sample more than brand preference does. A 500-unit pilot run & a 50,000-unit reorder will not draw the same sample even at the same level, & small lots can end up with a sample that feels disproportionately large. That is expected. With small runs the sampling plan asks for a bigger fraction because the confidence requirement does not shrink with the order.
Switching rules matter as much as the level. Standard plans include tightened, normal & reduced inspection. A factory with a clean record can move to reduced inspection & pay for fewer inspector hours; a factory that fails two lots out of five moves to tightened inspection & pays for more. Writing the switching rules into the agreement stops the factory from quietly staying on the easiest setting.
Critical, major & minor defects
Cosmetic inspection works with three classes. Critical defects are those that make the unit unsafe or illegal to sell: wrong content, broken seal on a product that relies on the seal, microbial failure, or a label that omits a required detail. A single critical defect fails the lot regardless of the agreed number.
Major defects are those a customer would notice & return: a pump that does not prime, a cap that will not stay closed, a fill level well below the stated volume, a print area that is illegible. These carry the tightest numeric AQL because they drive returns.
Minor defects are cosmetic blemishes that most buyers would accept without comment: a light scratch on a carton, a slightly off-centre label within tolerance, minor variation in shade of the outer box. They carry the loosest AQL because rejecting a lot over them costs more than the blemish does.
Classification disputes are the most common source of friction, & they are almost always caused by a defect list that was never written down. Fix the list before the run, not after the report.
Cosmetic-specific defect examples
Cosmetics differ from hard goods in ways that change the defect list. The product is a formulation, not a component, so faults can be invisible at the moment of packing & obvious three weeks later. Odour drift, colour shift, separation & viscosity change are all faults that appear after the inspector has signed.
That is why the agreement should name a retention sample & a post-fill observation window. Units held at controlled temperature from the same batch give you a reference if the market reports a problem. Without a retention set, a complaint turns into a debate about who caused it.
Why cosmetics complicate inspection
Packaging variety adds complexity. A single order can contain glass jars, aluminium tubes, airless pumps & printed cartons, each with its own failure modes. Glass shows chips & stress cracks; tubes show weak crimps & pinholes; pumps show inconsistent dosing; cartons show registration drift & glue squeeze.
The formulation adds another layer. Two jars from the same batch can differ slightly in colour without either being defective, because natural raw materials carry inherent variation. A defect list that does not allow for this will generate false failures & slow the line down.
Fill weight & dosing variance
Fill weight is the fault most likely to reach a regulator rather than a customer. Nominal volume on pack has to match what is in the container, & most markets apply an average-content rule rather than a per-unit minimum. That means the inspection should record individual weights & compute the average, not simply pass or fail each unit.
Ask the inspector to weigh a defined number of units from the sample, including the first units off the line & units from the end of the run. Start-up & end-of-run are where dosing drift shows up, because the filler is stabilising or emptying.
Print, label & artwork faults
Artwork faults are usually classified as major when they affect required information & minor when they affect appearance only. Illegible ingredient listing, missing net content, or a wrong shade name on the carton are major. A hairline scuff on a decorative element is minor.
Colour matching deserves its own line in the agreement. Pantone targets on coated cartons rarely land exactly, & the practical fix is a signed drawdown plus an agreed tolerance band, rather than a subjective judgement at the dock.
Closure, torque & leak checks
Closures are a frequent cause of returns & transit damage. Torque values should be recorded for a sample of units, & the acceptance band should be written into the specification rather than left to the inspector's judgement. Too little torque leaks; too much torque cracks the neck or makes the pack hard to open.
Leak testing deserves a defined method. A simple inverted hold at a stated temperature for a stated period is easy to replicate & produces evidence. A vague "check for leaks" instruction produces inconsistent results & no record.
Odour, colour & texture drift
Sensory drift sits between quality & formulation. Batch-to-batch variation in odour intensity or colour depth may be within the natural range of botanical raw materials & still be noticeable to a customer who bought the previous batch.
The practical control is a reference standard kept by both sides: a retained sample from the approved batch, plus a written description of the acceptable range. Inspectors compare against that reference instead of their own memory.
Who pays for rework & reinspection
A failed lot raises immediate commercial questions. Who pays for sorting? Who pays for the inspector's second visit? Who carries the cost of a delayed launch? If the agreement is silent, these questions are settled by whoever has more leverage that week.
A workable clause sets out that the factory sorts & rectifies at its own cost when the failure is attributable to production, that the brand pays for reinspection only when the first inspection passed & a later handling issue caused the problem, & that both sides accept the third-party report as the basis for the decision.
Writing AQL into the QC agreement
The agreement should name the standard being used, the inspection level, the AQL values per defect class, the defect list with classifications, the sample method for fill weight, the torque & leak methods, & the retention sample arrangement. Anything left open will be interpreted in the factory's favour at the moment it matters.
It should also state who appoints the inspector & who pays. Third-party inspection is inexpensive relative to a failed launch, & an independent report removes the argument about whether the factory's own QC data was complete.
When to step up to 100% inspection
Sampling accepts that some defective units will pass. For most cosmetic lines that is an acceptable commercial trade-off. It stops being acceptable when the defect is critical, when the launch is a retailer own-brand programme with a chargeback clause, or when the previous lot from the same line failed.
In those cases, full inspection or a documented sorting pass is the safer route. It costs more per unit & it is worth it when the downside of a single escaped unit is a delisting.
What AQL does not cover
AQL is an acceptance procedure for a finished lot. It says nothing about whether the formulation is stable over its shelf life, whether the preservative system performs, or whether the dossier behind the product is complete. Those are controlled by stability work, challenge testing & documentation, all of which happen before the inspection date.
It also does not replace process control. A lot that passes inspection because the sampling plan allowed two minor faults still tells you the line is producing minor faults. Trend the reports across lots & the sampling data becomes a process signal, not just a gate.
| Defect class | Typical cosmetic example | Usual AQL band | Action on failure |
|---|---|---|---|
| Critical | Broken seal, wrong content, microbial failure | 0 (reject on any) | Hold lot, quarantine, investigate root cause |
| Major | Pump will not prime, illegible ingredient print | 1.0 or tighter, as agreed | Reject or 100% sort at factory cost |
| Major | Fill weight below average-content rule | 1.0 or tighter, as agreed | Reject or top-up under supervision |
| Minor | Light carton scuff, glue squeeze within tolerance | 2.5 or looser, as agreed | Accept with note, watch trend |
| Minor | Label off-centre within tolerance band | 2.5 or looser, as agreed | Accept with note, watch trend |
| Borderline | Shade drift inside agreed reference range | Defined by reference sample | Compare to retained standard |
| Post-fill | Odour or viscosity change in retention set | Observed, not sampled | Escalate to formulation review |
What to put in the brief
- Inspection standard, inspection level & the AQL value for each defect class, written as numbers.
- A classified defect list covering fill weight, closure torque, leak, print, seal integrity & sensory drift.
- Sample method & count for fill weight, including start-up & end-of-run units.
- Closure torque band & the leak test method with temperature & duration.
- Retention sample quantity, storage condition & the observation window after filling.
- Who appoints & pays for inspection, & who pays for sorting & reinspection after a failure.
- Switching rules: what moves the line to tightened or reduced inspection.
- Reference standard arrangement for colour & odour comparison across batches.
Where our capability fits
We build the inspection plan with the brand before the first production slot is booked, because the defect list is far easier to agree on paper than at the dock. Our team writes the classified defect list into the specification, keeps retention samples from every production batch, & supplies the fill weight, torque & leak records with the shipment rather than after a query. Where a brand wants independent verification, we work to the third-party inspector's plan & regard the report as the acceptance record.
That service covers private label (your brand on a proven base), OEM/ODM contract manufacturing for brands bringing their own specification, & full custom formula development when the product needs to be built from the brief. Related programmes you may want to review include softgel capsule blister sample packs, argan oil shampoo sample packs & time-release serum capsule sets. For the wider quality picture, see our notes on factory QC & consistency & the supplier audit checklist.
Ready to put an inspection plan behind your next run? Send the specification through our inquiry page, or write directly to capsulebiotech@163.com. If a call suits better, reach us on WhatsApp at +86 158-7655-5319 (Capsule Series), +86 158-7655-5319 (Cosmetics Series).
This article gives general sourcing guidance for brand owners & is not legal or regulatory advice. Inspection standards, sampling plans & market requirements vary by destination & change over time; please confirm current requirements against official texts & with your own professional advisers before relying on them.