Sampling levels used in amazon fba inspection
Every inspection verdict rests on a sample, and the sampling level decides how large that sample is. Too small a sample lets real defect rates hide; too large a sample wastes money confirming what a smaller one would already show. Amazon FBA sellers face this tradeoff on every shipment, compounded by compliance markings that must be checked differently from ordinary defects. This guide explains the standard levels, when each fits FBA shipments, and how regulatory checks sit alongside them.
Key takeaways
- General Level II is the default for FBA pre-shipment inspection, balancing defect detection against inspection cost for typical consumer goods lots.
- Drop to Level I only for mature products with proven suppliers, and rise to Level III for high-risk launches where a missed defect is expensive.
- Reserve special levels S-1 to S-4 for checks where full sampling is impractical, such as heavy functional tests or dimensional audits on large lots.
- Verify CE marking presence, format, and accompanying information on every variant, since marking rules attach to the product model rather than the sample draw.[1]
- Remember that destination technical regulations can hold non-compliant shipments regardless of AQL results, which is why sampling never replaces compliance verification under the WTO TBT framework.[2]
Sellers often treat the inspection sample as a technicality: the inspector picks some cartons, looks at some units, and the lot passes. In reality the sampling plan is the entire statistical foundation of the verdict. A lot accepted on thirteen units carries a very different meaning from the same lot accepted on two hundred, and buyers who do not understand the difference routinely misread clean reports as proof of near-perfect quality.
The standard system behind these numbers is decades old and deliberately simple. Given the lot size and the chosen level, a table yields a code letter, and the code letter yields a sample size plus acceptance and rejection thresholds for each defect class. Every professional inspection follows this chain, which means a seller who learns it can predict sample sizes, budget inspection days, and argue from facts when a supplier disputes a failure.
Regulatory compliance adds a second track. Markings, warnings, and technical documentation do not behave like random defects, so they are verified by different logic than workmanship sampling. The sections below cover the standard levels first, then show how compliance checks fit beside them without confusion.
How the level system works
The chain runs lot size to code letter to sample size. A lot of five hundred units inspected under General Level II maps to code letter H, which means fifty units examined; the same lot under Level I maps to code letter E, meaning only twenty units. Acceptance numbers then depend on the AQL limits agreed for critical, major, and minor defects: with a common major AQL, fifty units might tolerate one major defect while twenty units tolerate none, so the level changes both the evidence and the verdict.
This is why the level must be agreed before the booking, not improvised on arrival. Changing levels mid-visit invalidates the preparation: the inspector budgeted time for one sample size, the factory staged a different quantity of open cartons, and any verdict produced under a renegotiated plan looks arbitrary to whichever side loses. Professional practice writes the level into the inspection scope alongside the AQL values, defect list, and reference samples.
Code letters also govern re-inspections after failures. A failed lot that returns for re-check is normally re-sampled at the same level or stricter, never relaxed to make passing easier. Sellers should state this rule in the supplier agreement, because factories under delivery pressure will otherwise lobby for the smallest sample that could pass.
Which general level fits your shipment
General Level II suits most FBA pre-shipment inspections: new suppliers, new products, seasonal rush production, and any lot whose failure would trigger listing-damaging returns. Its sample sizes detect defect rates in the low single digits with reasonable confidence, which matches the quality bar that keeps Amazon metrics healthy. Unless there is a specific reason to deviate, Level II is the defensible default and the easiest to justify to a supplier.
Level I fits mature products from proven factories with several consecutive clean shipments, where the inspection functions as surveillance rather than discovery. The smaller sample costs less inspector time and disturbs the warehouse less, which matters for frequent small-lot programs. The risk is explicit: defect rates that Level II would catch can pass Level I untouched, so downgrading is a privilege the supplier earns with data, not a discount granted at the first booking.
Level III belongs to high-stakes situations: product launches with large first orders, categories with safety exposure, lots produced under known disruption, and any shipment where a single prior failure already occurred. The larger sample raises inspection cost and time, sometimes requiring two inspectors or two days, but it converts the verdict from a rough signal into evidence strong enough to support hard commercial decisions.
Levels at a glance and where each is used
| Level | Sample size character | Typical FBA use |
|---|---|---|
| General Level I | Smallest of the general levels, roughly forty percent of Level II | Mature SKUs from proven suppliers, routine surveillance of repeat lots |
| General Level II | The industry default, balanced cost and detection | Standard pre-shipment inspection for most FBA lots and new suppliers |
| General Level III | Largest general sample, roughly one and a half times Level II | Launches, safety-related categories, post-failure re-inspections |
| Special S-1 to S-4 | Very small samples for expensive or destructive checks | Lab-style functional tests, dimensional audits, drop and stress checks |
| Full verification | Every unit or every carton checked for the attribute | Barcodes, FBA labels, variant marking, and regulated warnings |
Choose the level in five steps
- Rate the cost of escapeEstimate what happens if a three percent defect rate reaches customers: review damage, return shipping, disposal fees, and possible listing suspension. High escape cost pushes the level up regardless of lot size.
- Check supplier historyReview the last three inspection reports for this factory and product family. Consecutive clean lots at Level II justify trying Level I; any failure in the window keeps the level up or raises it.
- Separate destructive checksAssign special levels to tests that damage units or consume expensive time, such as pull tests and internal examinations. Do not let these shrink the general workmanship sample, which must stay at its full size.
- Confirm the day the quantity allows itVerify that the finished and packed quantity on the inspection date supports the planned code letter. If production is incomplete, either reschedule or record explicitly that the verdict covers only the finished portion.
- Write it all downRecord the level, AQL values, defect classification, and any full-verification attributes in the booking scope. A level that exists only in conversation will be renegotiated the moment it produces an inconvenient result.
Where regulatory checks sit beside sampling
Markings and documentation follow different logic from random defects. A missing or malformed conformity marking is a systematic error: if the artwork file is wrong, every unit carries the same error, so finding it once is enough and sampling theory adds nothing. Professional scopes therefore verify CE marking elements, warnings, and traceability information across variants and packaging levels as full checks, independent of the AQL sample drawn for workmanship.[1]
The stakes justify the extra attention. Destination markets enforce technical regulations at the border, and consignments that fail marking, labeling, or documentation requirements can be held or refused no matter how clean the workmanship sample was. These enforcement regimes operate under international disciplines on technical barriers to trade that require measures to be transparent and non-discriminatory, but they remain fully binding on the seller.[2] Sampling confidence never substitutes for compliance evidence.
Confirm the level with UTS when booking so the inspector arrives with the right tables and the day stays on schedule. Stating the general level, the special-level tests, and the full-verification attributes in one scope prevents the two most common failures: a visit too short for the sample drawn, and a clean AQL verdict on a shipment that compliance checks would have stopped.
Sources
- Wikipedia: CE marking —— Overview of CE marking as conformity marking for products sold in the European Economic Area.
- WTO: Agreement on Technical Barriers to Trade —— WTO portal for the TBT Agreement covering technical regulations, standards and conformity assessment.
Frequently asked questions
Is General Level II always the right choice for FBA?
It is the right default for most FBA pre-shipment inspections because it balances detection and cost for typical lot sizes and consumer goods risk. Deviate downward to Level I only with proven supplier history, and upward to Level III for launches, safety exposure, or post-failure re-inspections. Document the reason for any deviation so the choice survives later scrutiny from the supplier or your own team.
What are special levels S-1 to S-4 used for?
Special levels produce very small samples for checks where examining hundreds of units is impractical: destructive testing, time-consuming functional tests, and detailed dimensional audits. They answer whether a specific capability exists rather than estimating a defect rate across the lot. Never use a special level as the general workmanship sample, because its tiny size cannot support any verdict about overall lot quality.
Should CE marking be sampled or fully checked?
Check marking elements across every variant and packaging level rather than relying on the AQL sample alone. Marking errors are systematic, flowing from artwork and printing setups that affect all units identically, so the standard statistical reasoning behind sampling does not apply. Verify the mark's presence, proportions, accompanying information, and consistency between product, packaging, and documents.
Can a lot pass AQL sampling but still be rejected at the border?
Yes. AQL sampling estimates workmanship defect rates; border authorities enforce technical regulations covering marking, restricted substances, documentation, and declared standards. A lot with excellent finish can still be held for a missing marking or a non-compliant label. This is why professional scopes run compliance verification as a separate track alongside sampling instead of assuming a clean AQL result covers everything.
See your numbers 23 days earlier.
Book a 30-minute walkthrough with a solutions engineer. Median time-to-first-insight: 14 minutes from signup.
▶ Get a live demo →