The third part of this sampling series covers the booking itself: what importers tell the inspection team before shipment, and how lot size, inspection level, and defect limits turn into a concrete sampling plan. A plan built on accurate lot information makes inspection day smooth; one built on guesses produces arguments about numbers instead of answers about quality. Here is how to book the plan so the sample, the acceptance numbers, and the verdict all rest on solid ground.
Key takeawaysThe sampling plan translates lot size and inspection level into a sample size and acceptance numbers before anyone opens a carton. · Acceptance sampling gives a statistical verdict on a lot from a random sample, not a promise about every unit, which is why random carton selection is the heart of the method [1]. · Technical regulations in the destination market decide which defects count as critical, so the plan must reflect those requirements alongside appearance criteria [2]. · Mixed lots with several SKUs usually need separate plans or explicitly stratified draws rather than one blended sample. · Confirm the lot size and carton access the day before; most failed plans trace back to counts that changed after booking.
The first two parts of this series explained what an AQL sampling plan is and how acceptance numbers work. This third part covers the booking itself: what importers tell the inspection team before shipment, and how those details turn into a concrete plan for the day. A sampling plan is only as good as the lot information behind it.
Importers sometimes treat the plan as a formality, a box to tick while arranging the visit. In practice the plan decides how many cartons get opened, how many units get checked, and where the pass or fail line sits. Booking it carefully is the cheapest quality control an importer can buy.
This article walks through lot size declaration, inspection levels, critical versus major versus minor defect limits, and the special cases of mixed lots and very small shipments bound for final random inspection in China.
Lot size is the foundation
Every sampling table starts from the lot size: the total number of units presented for inspection. Get that number right and the sample size, the acceptance numbers, and the inspection duration all follow. Get it wrong and everything downstream wobbles, because a sample drawn for five thousand units means something different when the lot turns out to hold eight thousand.
Importers should declare the lot by SKU, colour, and size, not as a single headline figure. A lot of six thousand units split across four colours behaves differently from six thousand identical units, and the inspection team needs the breakdown to allocate the draw. The booking form for a final random inspection in China should therefore list each variant quantity separately, with the total as the sum rather than the starting point.
Confirm the figure the day before the visit. Production overruns, split shipments, and cartons held back for accessories all move the total after booking. A short message confirming the packed quantity and the number of cartons ready for access takes two minutes and prevents the most common inspection-day dispute.
Inspection levels and what they change
The inspection level sets how steeply the sample grows with lot size. General level II is the default for most consumer goods: it balances detection power against time on site. Level I suits low-risk, low-value goods where a lighter touch is proportionate. Level III suits high-value, safety-relevant, or first-production shipments where extra detection power justifies the longer day.
Moving up a level does not change the acceptance logic; it changes the sample size and therefore the sensitivity of the verdict. More units checked means scattered defects surface more often and borderline lots resolve more clearly. The trade-off is clock time: bigger samples need more cartons opened, more measurements taken, and more function tests run.
Special levels exist for checks that would destroy the goods or take very long, such as certain physical tests applied to a handful of units. These sit outside the main visual and functional sample and should be agreed explicitly at booking. An importer who assumes every test runs at level II will either overpay in lab time or under-specify the checks that matter.
| Choice at booking | What it sets | Practical effect |
|---|---|---|
| Lot size declaration | Sample size code letter | Determines how many units get checked |
| Inspection level I, II, or III | Steepness of sample growth | Higher level means larger sample and longer day |
| AQL limits by defect class | Accept and reject numbers | Tighter limits fail lots on fewer defects |
| Single versus stratified draw | How sample spreads across SKUs | Stratified draws represent each variant fairly |
| Special test levels | Units for destructive or slow tests | Keeps costly tests proportionate |
Setting AQL limits by defect class
Critical defects threaten safety or regulatory compliance: sharp edges on toys, exposed wiring, missing safety markings. Most buyers set the critical limit at zero tolerance, meaning any confirmed critical defect fails the lot pending review. This is a commercial policy choice informed by the destination market's technical regulations, which define the safety and compliance baseline the goods must meet [2].
Major defects impair function or saleability: a garment with a broken zipper, an electronic item that will not charge, a colour far outside the approved band. Common commercial limits sit in the low single digits, tightened for premium goods and relaxed for low-value promotional items. Minor defects affect appearance without touching function: loose threads, slight print offset, small surface marks. Limits here are the most negotiable part of the plan.
Write the three limits into the booking explicitly rather than relying on defaults. Defaults vary between inspection providers and product categories, and an importer who assumed one set of numbers while the inspector applied another will argue about arithmetic instead of quality. Three numbers in the booking note remove the entire class of dispute. Acceptance sampling then does its proper job: converting the sample result into a lot verdict through published tables rather than negotiation [1].
Mixed lots and small shipments
Mixed lots need deliberate design. When several SKUs share a shipment, a single blended sample can hide a weak variant behind strong ones: the lot passes on average while one SKU fails on its own. The cleaner approach books separate plans per SKU above an agreed quantity threshold, or a stratified draw with per-variant acceptance tracked alongside the headline verdict. Either way, the decision belongs in the booking, not on the warehouse floor.
Very small shipments need a different conversation. Sampling tables assume lots large enough for statistics to bite; below a few hundred units the sample approaches the whole lot and the economics of random inspection shift. Buyers with small lots should ask what plan shape fits: sometimes a full check of critical points plus a reduced sample for appearance, sometimes a full sort of the risky attribute. The worst option is applying a big-lot plan blindly and checking a token handful.
The service page for final random inspection in China describes how the booking captures lot size, variant breakdown, and defect limits, so the plan confirmed before the visit matches the goods presented on the day. Importers who fill in that breakdown carefully get sampling that represents their lot; those who round the numbers get sampling that represents somebody's lot. Precision at booking is the difference.
Sources
- Wikipedia: Acceptance sampling —— Overview of acceptance sampling including AQL concepts used in product inspection.
- WTO: Agreement on Technical Barriers to Trade —— WTO portal for the TBT Agreement covering technical regulations, standards and conformity assessment.
Frequently asked questions
What details must an importer supply to book a sampling plan?
Give the lot size, the breakdown by SKU and colour, and the inspection level you want applied, plus any tightened limits for critical defects. The inspection team then confirms the sample size and the acceptance numbers before the visit. Booking a plan without the lot size is like ordering fabric without metres: nothing can be calculated.
What changes when you move from general inspection level II to level III?
It means more cartons get opened and more units get checked, which raises the chance of finding scattered defects but also lengthens the day. Higher levels suit high-value or high-risk shipments. For routine replenishment lots with stable history, the standard level keeps cost and time proportionate.
Should the sample cover every size and colour proportionally?
It should, within reason. Sizes that dominate the lot deserve proportional representation in the sample, and suspected weak spots can receive extra attention on top of the random draw. But the core sample must stay random: replacing random units with hand-picked ones destroys the statistical meaning of the acceptance decision.
Why do sampling plans fail most often on inspection day?
Counts gaps are the most common cause: the lot size changes after booking, or cartons are not all accessible, so the planned sample no longer matches the goods. The second cause is mixed lots booked as one plan when they needed separate plans. Both are fixed by confirming quantities and carton access the day before.