Bedding Inspection and AQL Sampling: A Wholesale Buyer's Guide

To inspect a wholesale bedding order you do not check every set — you agree an Acceptance Quality Limit (AQL), pull a statistically sized random sample under ISO 2859-1, sort each flaw into critical, major or minor, and accept or reject the whole lot against pre-set accept/reject numbers. For a typical bedding order most buyers set General Inspection Level II, a major-defect AQL of 2.5 and a minor-defect AQL of 4.0, with zero tolerance for critical defects. The rest of this guide is exactly how that mechanism works, the correct sample-size figures for common lot sizes, and the bedding-specific defects — skipped stitches, colour difference, GSM deviation, seam strength, fitted-sheet elastic — that decide whether your cartons ship.

What AQL actually means for a bedding order

AQL is the quality level that is, in the words of ISO 2859-1, the worst tolerable process average — the maximum percentage of defective units a buyer is willing to accept across a series of lots. It is a threshold for a sampling plan, not a promise about your individual shipment. A major-defect AQL of 2.5 does not mean "2.5% of my sets may be defective and that is fine"; it means the sampling plan is tuned so that a production run running at about 2.5% defective has a high probability of being accepted, and a much worse run has a low probability of slipping through. This is the single most misread point in bedding sourcing: passing AQL is a statistical bet on the whole lot, drawn from a sample, not a guarantee that every set is flawless.

The counterpart concept is the LTPD, or lot tolerance percent defective — the bad quality level a buyer wants a low probability of accepting. Every sampling plan sits on an operating characteristic (OC) curve that plots the probability of accepting the lot against its true percent defective. Two risks live on that curve: the producer's risk of rejecting a lot that is actually at the AQL, and the consumer's risk of accepting a lot that is actually at the LTPD. You cannot drive both to zero without inspecting every piece — the sample size is the lever that trades cost against certainty.

The standards behind AQL: ISO 2859-1, ANSI/ASQ Z1.4 and MIL-STD-105

Three standards describe the same family of attribute-sampling tables, and a bedding supplier or third-party agency will name one of them on the inspection report. The original was MIL-STD-105, a United States defense standard built on the statistical work of Walter Shewhart, Harold Dodge and Harry Romig, whose final revision MIL-STD-105E was officially cancelled in February 1995. Its methodology did not disappear — it lives on in ANSI/ASQ Z1.4, "Sampling Procedures and Tables for Inspection by Attributes," which the US Department of Defense recommended as the civilian replacement, and in the international standard ISO 2859-1, "Sampling procedures for inspection by attributes." For practical purposes the sample sizes and accept/reject numbers you will see quoted are the same across ISO 2859-1 and ANSI/ASQ Z1.4; if a report cites the retired MIL-STD-105 by name, that is a small flag that the paperwork has not been refreshed.

One thing none of these standards do is tell you which AQL to use, or how to split defects into classes — that is a commercial decision you set before production. Getting the supplier relationship right in the first place matters more than any table; our guide at /blog/how-to-verify-chinese-bedding-supplier-factory-vs-trading-company covers separating a real production partner from a pure reseller before you ever book an inspection.

Inspection levels and sample-size code letters

ISO 2859-1 defines seven inspection levels — three general levels (I, II, III) and four special levels (S-1 to S-4) used for cheap or destructive tests where a tiny sample is enough. General Inspection Level II is the default for most consumer-goods inspections worldwide and the right starting point for bedding: Level I pulls a smaller sample (lower cost, less protection) and Level III a larger one (more cost, tighter protection). The workflow is a lookup, not a calculation. First you take your lot size and inspection level to a table that returns a sample-size code letter (A through R). Then a second table maps that code letter, plus your chosen AQL, to the exact sample size and the accept and reject numbers.

A worked example: an order of 2,000 sets of one style, inspected at General Level II, lands in the 1,201–3,200 lot band, which gives code letter K and a sample size of 125 sets. The inspector opens cartons at random until 125 sets are drawn, then judges those 125 against your checklist. Change the style or the order quantity and the code letter — and therefore the sample size — moves with it, which is why AQL is applied per style, not per shipment.

Reading the accept and reject numbers

Under a single sampling plan the rule is blunt: inspect the sample, count the defectives in each class, and if the count is at or below the accept number (Ac) the lot passes; if it reaches the reject number (Re) the lot fails. Ac and Re are always consecutive integers — there is no in-between. The table below gives the real figures for General Inspection Level II, single sampling, normal inspection, at the two AQLs most bedding buyers use for major and minor defects.

Read a single row across and the trade-off is visible: at code K the sample is 125 sets whatever the AQL, but the tolerated defect count rises from 7 majors to 10 minors, because minor flaws matter less. Note too that the sample size scales far more slowly than the lot — doubling a 2,000-set order to nearly 4,000 only moves you from 125 to 200 sets inspected. That is the whole point of acceptance sampling: statistical protection without the cost of a full sort.

Critical, major and minor defects: how to classify a bedding flaw

AQL is applied per defect class, so the classification you agree before production is as important as the AQL number itself. A critical defect is one likely to make the product unsafe or unusable. A major defect is one that reduces usability or would probably cause the end customer to return the product. A minor defect is a deviation from the specification that does not affect usability and is unlikely to trigger a return. Bedding sits at direct-skin-contact class, so anything touching safety or chemical compliance is treated as critical.

A bedding defect checklist: what actually fails inspection

Generic AQL levels are useless without a bedding-specific checklist tied to your tech pack and sealed sample. These are the check points a competent inspector runs on each set in the sample, how each is tested, and the class it usually falls into. Agree this list in writing before production so the inspector is not improvising on the day.

DUPRO vs final random vs 100% inspection: when to inspect

AQL sampling can be run at different moments, and the moment changes what the inspection can save. A during-production (DUPRO) inspection samples an order that is part-finished, so a systemic fault — wrong fabric, off colour, GSM out of tolerance — is caught while there is still time and material to correct it. A final random inspection (FRI) samples the finished, packed lot and reports its as-shipped condition; it is the standard gate before you release balance payment. A 100% inspection sorts every piece and is normally a salvage operation after a lot has already failed.

The bedding inspection workflow, step by step

Put the pieces together and every AQL inspection follows the same five steps, whether your own QC or a third-party agency runs it.

Bedding-specific checks worth spelling out

Colour difference and the sealed sample

Colour is the most disputed major defect in bedding because "white" and a specific print shade drift between dye lots. The only defensible reference is a signed, sealed sample judged under standard D65 daylight against a grey scale — an eyeball comparison in a warehouse doorway is worthless. Approve that sealed sample during sampling, before bulk; our guide at /blog/how-to-order-bedding-samples-from-china-before-bulk covers locking colour, hand-feel and construction so the inspection has a fixed target.

GSM, thread count and dimensions

GSM (grams per square metre) is the metric weight of the fabric and a direct proxy for how much cotton is in your sheet; it is verified by cutting a fixed-area swatch, weighing it and comparing to the agreed tolerance. Thread count — the warp and weft threads counted in a square inch or square centimetre — is checked with a pick glass, useful because a headline thread count is easy to overstate. Finished dimensions of the flat sheet, fitted sheet, duvet cover and pillowcases are tape-measured against the tech pack with a stated tolerance, since an oversize duvet cover or a shallow fitted sheet is a guaranteed return.

Seam strength and fitted-sheet elastic

Seam strength and elastic are the defects that survive the showroom and fail in the laundry — which is exactly where hotel programmes get hurt. A skipped stitch or a weak corner seam opens after a few wash cycles, and fitted-sheet elastic that will not recover stops the sheet gripping the mattress. If you are buying for hospitality, pair AQL sampling with wash-cycle durability testing; our note at /blog/hotel-linen-durability-200-wash-cycles explains why a sheet that passes a visual AQL check can still fail at 200 washes.

How we build inspection into your order

BeddingTextilePro runs a source-factory direct supply model: we hold locked, dedicated production lines at large-scale Nantong mills, and your goods ship direct from the mill. Because our own QC team sits on those lines rather than at a distant office, we schedule three in-line inspections that we assign to the mill lines — a first-off-the-line check as the opening cartons are sewn, a during-production (DUPRO) check near the mid-point to catch any systemic fault early, and a final random inspection to General Inspection Level II before the balance is invoiced. Colour is judged against your sealed sample, GSM and dimensions are measured against the tech pack, and the accept/reject decision is recorded per defect class.

You are welcome to appoint your own third-party agency for the final random inspection; we book the date and open the cartons for them. Because you buy on mill pricing with no middleman markup, inspection is the one place we would rather over-invest than cut — a rejected lot is far more expensive than the sample check that caught it. Buyers building a /private-label-bedding programme or sourcing /hotel-linen can have the AQL, level and full checklist written into the order confirmation so there is nothing to argue about on inspection day.

Setting the right AQL before you place the order

The mistake that costs money is treating AQL as a formality quoted after production. Fix four things in writing before the deposit: the inspection level (General II for most bedding), the AQL for each defect class (commonly major 2.5, minor 4.0, critical 0), the sealed sample the inspector will judge against, and a bedding-specific checklist covering stitching, seams, colour, GSM, dimensions, thread count and elastic. Do that and the accept/reject numbers become a neutral, quotable rule both sides agreed to — not a negotiation held over your cartons at the port.

Questions fréquentes

What AQL should I use for a wholesale bedding order?

For most bedding, use General Inspection Level II with a major-defect AQL of 2.5, a minor-defect AQL of 4.0, and zero tolerance for critical (safety or legal) defects. This is an industry convention, not a rule fixed by ISO 2859-1, so you can tighten it — Level III or major AQL 1.5 — for premium hotel programmes or private-label ranges where returns are costly.

How many bedding sets does the inspector actually check?

The sample size comes from your lot size and inspection level, not a fixed percentage. At General Level II, a 2,000-set lot of one style is code letter K and 125 sets are inspected; a 5,000-set lot is code L and 200 sets. The sample scales far slower than the order, which is the whole point of acceptance sampling — statistical protection without inspecting every piece.

Does passing AQL inspection mean the order has zero defects?

No. AQL is a statistical acceptance rule based on a random sample, not a guarantee that every set is perfect. A lot that passes at major AQL 2.5 can still contain a small number of defective sets within the accepted count. AQL controls the probability of shipping a bad lot; for zero defects you would need a 100% inspection, which is slow and costly.

Do I need a DUPRO inspection or just a final inspection?

A final random inspection on the finished, packed lot is the standard gate before balance payment and is enough for repeat orders of proven styles. Add a during-production (DUPRO) check on first orders, new styles or new suppliers, because it catches systemic faults — wrong fabric, off colour, GSM out of tolerance — while there is still time and material to correct them cheaply.

Sources et références

1. 1.Wikipedia — Acceptance sampling (AQL, single vs multiple sampling, OC curve)
2. 2.Wikipedia — MIL-STD-105 (history, 1995 cancellation, ANSI/ASQ Z1.4 and ISO 2859-1 succession)
3. 3.NIST/SEMATECH e-Handbook of Statistical Methods — Single sampling plans for attributes (AQL, LTPD, producer's/consumer's risk)
4. 4.doEEEt — What does Acceptance Quality Limit (AQL) mean? (inspection levels, code letters, ISO 2859-1)
5. 5.OEKO-TEX — STANDARD 100 (harmful-substance testing; bedding is direct-skin-contact Product Class II)
6. 6.Wikipedia — Units of textile measurement (thread count and grammage / GSM)