Short answer: Start with General Inspection Level II unless your specification requires another level. Consider Level I for stable, low-risk products when the smaller sample is justified; consider Level III when greater discrimination is needed, such as with a new supplier or an unstable quality history. Use Special Levels S-1 to S-4 for agreed checks that need small samples, not as an automatic replacement for the general inspection scope. Confirm the lot size, AQLs, inspection severity, and standard edition before selecting the sampling plan.
We are UTS, a third-party provider of product inspection, factory audit, supplier evaluation, and laboratory testing coordination. Our reports present actual findings within the agreed scope, not corrective-action plans, improvement recommendations, or shipment-release advice. The guidance below helps you define an inspection scope; it is separate from report content.
What the Inspection Level Controls
An inspection level sets the relationship between lot size and sample size. It is not an AQL value or a defect category. ISO 2859-1:2026, Edition 3, provides AQL-indexed sampling schemes for lot-by-lot inspection by attributes. Use the standard and edition specified in your agreement rather than assuming every inspection uses the latest edition [1].
For a single-sampling plan, n is the sample size and c is the acceptance number. In the defective-unit model described by NIST, the lot is rejected when the number of defective units exceeds c. Your procedure must specify whether you count nonconforming units or individual nonconformities; those are not interchangeable [2].
Lot size and inspection level lead to a sample-size code letter. The applicable sampling table, AQL, and inspection severity then determine the plan. General Levels I, II, and III are separate from normal, tightened, and reduced inspection: choosing Level III is not the same as switching to tightened inspection. NIST explains these distinctions in its historical MIL-STD sampling overview; use your agreed current standard for the actual tables and switching rules [3].
Compare General Levels
For a given lot size, the general levels provide different sampling effort and discrimination. Exact samples depend on the applicable tables; different levels do not necessarily produce different sample sizes for every lot-size range.
| Level | Relative sampling effort | When to consider it |
|---|---|---|
| General Level I | Generally lower | Stable specifications and quality history, low consequences of defects, and an agreement that accepts less discrimination. |
| General Level II | Usual starting point | Routine lot inspections where the specification does not require another level. |
| General Level III | Generally higher | Greater discrimination is justified by product risk, a new supplier or design, or inconsistent previous results. |
When Level I Is Justified
Consider Level I only when you can explain why less sampling evidence is acceptable. Useful supporting information includes unchanged specifications, relevant previous inspection results, and low consequences of the defects being assessed. A repeat order alone is not enough if the material, process, or product has changed.
When to Consider Level III
Level III may suit a new supplier, a changed design, or a history of inconsistent workmanship when you need greater discrimination. Confirm that the site can present the agreed lot, provide access to the cartons, and support the required inspection time. If the planned sample cannot be completed, agree how to handle the limitation rather than silently reducing the sample.
A larger general sample does not replace product-specific safety testing or mandatory requirements. Those checks may need their own methods and sample sizes regardless of the general level.
Use Special Levels Deliberately
Special Levels S-1, S-2, S-3, and S-4 provide options where small samples are necessary and the associated sampling risk is acceptable. Consider them for a defined characteristic, such as a destructive test or a time-intensive performance check. The reason for the smaller sample should be explicit in the inspection agreement.
Do not use a special-level result for one characteristic to imply that appearance, packing, and every other characteristic passed. Define which checks use the general level and which use a special level. Our laboratory testing coordination follows the applicable test method; a special inspection level does not override that method's sample requirements.
Choose the Level in Four Steps
- Define the lot and its risks. Identify the purchase order, product, quantity, and included SKUs. Agree whether different models, materials, colors, or sizes belong in one lot or need separate assessment. Review what changed since the previous inspection and the consequences of the defects you want to detect.
- Agree on the acceptance criteria. Define critical, major, and minor defects and how they are counted. Our published final random inspection examples include Major AQL 2.5 and Minor AQL 4.0, with custom criteria subject to agreement. Specify critical-defect handling separately: a zero-acceptance requirement means no critical defects are accepted in the inspected sample, not proof that none exist in the lot. Do not treat a website's shorthand “Critical 0” as a standard AQL-table column.
- Select the level and applicable plan. Start with Level II; justify Level I or III using the risk and history identified above. Record normal, tightened, or reduced inspection separately and follow the agreed standard's switching rules. Use the lot size, code letter, and applicable AQL table to obtain the actual sample size and acceptance/rejection numbers. Follow table instructions rather than calculating them from the AQL percentage.
- Confirm readiness and document the scope. Check access to the goods, the approved reference sample, specifications, packing requirements, equipment, and any separate tests. Send the agreed lot definition, level, AQLs, counting basis, severity, standard edition, and checklist to our team before the visit.
The sample must be selected at random from the defined lot; a convenient set of hand-picked pieces does not provide the same basis for a lot conclusion. NIST describes acceptance sampling as using a random sample to decide whether to accept or reject a lot, rather than estimating its exact quality [4].
How Our Published Workflow Applies
Our final random inspection workflow starts with your purchase order details, specifications, and factory address. We prepare the checklist around the product requirements, AQL criteria, and special testing needs before random sampling and on-site checks. The report records actual inspection findings, including measurements and observed defects; you make the subsequent shipment decision. For choosing a level, this means the lot definition and test scope must be ready before sampling begins.
The inspection stage also matters. Our published readiness points are about 5–10% production completion for initial production inspection, about 30–50% of goods finished for during-production inspection, and 100% production completion with at least 80% packed for final random inspection. These are service-stage conditions, not alternative AQL levels. An early-production visit is not evidence about goods that have not yet been produced.
What a Higher Level Cannot Establish
- It does not change the AQL by itself. A higher level and a lower AQL are different choices. Specify both rather than describing Level III as a “stricter AQL.”
- It does not prove the lot is defect-free. Sampling carries producer's and consumer's risks. Their probabilities depend on the plan and quality level; there is no universal risk percentage attached to Level I, II, or III.
- It does not assess the entire supplier. Our factory evaluation addresses supplier systems and capability; a lot inspection addresses the defined goods.
- It is not a full inspection or compliance certificate. If every unit needs checks, define those checks as a separate full inspection scope. Neither a sample result nor a full inspection establishes compliance with requirements outside the agreed checks.
Our reports state the actual scope, measurements, observations, test results, defects, and any checks that could not be completed. They do not contain corrective-action recommendations, improvement plans, or shipment-release advice. Inaccessible goods and unfinished tests are limitations to record, not results to assume.
Questions About Choosing
Is Level II Mandatory?
Not in every inspection. Start with the specification and agreement. If another level is required, use it or resolve the requirement before sampling. Do not change the level after seeing the result to obtain a different outcome.
Can General and Special Levels Be Combined?
Yes, when the scope identifies the characteristics covered by each plan. For example, an agreed general level may cover workmanship while a special level covers one time-intensive check. Keep each sample and acceptance rule distinct, and do not extend a limited test result to unrelated characteristics.
Does a Larger Lot Require Level III?
No. Lot size is already an input to the sample-size code-letter table. Review the plan when lot size changes, but choose the inspection level for the discrimination needed. Sample sizes change in table-defined steps, not in direct proportion to every increase in lot size.
What If the Whole Lot Is Not Available?
Do not automatically redefine the lot as whatever happens to be accessible. Confirm with our team whether the visit should be rescheduled or whether a clearly identified, agreed partial scope is appropriate. Any report must identify what was actually inspected and the limitations; its findings cannot establish the condition of uninspected goods.
Confirm Your Scope with UTS
Send our team the purchase order quantities, SKU breakdown, specifications, relevant previous inspection results, proposed AQLs and level, production and packing status, and any separate test requirements. These materials help establish whether the proposed sampling plan can be carried out as agreed.
Not sure which level your lot needs?
Share your lot details, specifications, and proposed acceptance criteria. Our team can help clarify the inspection scope before sampling starts.
Discuss Your Inspection ScopeBottom line: Use General Level II as a starting point, justify any move to Level I or III, and assign special levels only to clearly defined checks with acceptable sampling risk. Agree on the lot, AQLs, inspection severity, and standard edition before using the tables to finalize the plan.
References
- [1] ISO 2859-1:2026, Edition 3 — Sampling schemes indexed by acceptance quality limit (AQL) for lot-by-lot inspection.
- [2] NIST/SEMATECH e-Handbook — Types of lot acceptance sampling plans, single-sampling notation, and sampling risks.
- [3] NIST/SEMATECH e-Handbook — Historical MIL-STD sampling overview: inspection levels, code letters, and normal/tightened/reduced schemes. This is conceptual background, not the current ISO tables.
- [4] NIST/SEMATECH e-Handbook — Acceptance sampling, random selection, and the purpose of a lot decision.
