Reading inspection reports without guesswork: part 7
An inspection report is a decision tool, not a certificate. Its job is to tell you whether this specific consignment meets the agreed criteria, with enough evidence that you could reach the same conclusion yourself. Most buyer confusion comes from reading reports backwards: starting with the verdict and treating the data as decoration. This guide reverses that habit. It shows how to read the sample, the defects, and the arithmetic first, so the verdict at the end confirms what you already know instead of surprising you.
Key takeaways
- Start every report review with the sample size table and the defect counts by class, because those two numbers carry the verdict and everything else is context.
- Acceptance sampling judges lots by comparing defects found against allowed numbers for a defined sample, so a report must name its standard and level to be checkable[1].
- Critical defects fail a lot on their own, while major and minor defects accumulate against their separate acceptance numbers within the same sample.
- Read the verdict as a factory acceptance decision on this consignment against agreed criteria, not as a judgment on the supplier relationship[2].
- Close the loop in writing: confirm the disposition of failed lots, the rework scope, and the re-inspection sample before the goods move.
Inspection reports fail buyers in two opposite ways: some are too thin to support any decision, and some are so dense that busy buyers read only the verdict line. Both failures have the same root cause. The reader never learned which three numbers carry the decision and which twenty pages are supporting evidence. This article teaches that distinction, so the next report you open answers its question in ten minutes instead of raising five new ones.
The statistical backbone is acceptance sampling: judging a larger lot from a defined sample with explicit accept and reject numbers. Sampling standards formalize this with tables linking lot size, inspection level, sample size, and allowed defects[1]. A report that names its standard, level, and sample size can be checked; a report that omits them can only be believed. Belief is not a quality system.
This seventh part of the report-reading series covers the anatomy of a solid report, defect classes, verdict arithmetic, and the follow-up questions that close the loop. It pairs naturally with the acceptance logic used at dispatch: confirming the finished consignment against agreed criteria before release[2]. Read it once, then keep it beside your inbox as a checklist.
The three numbers that carry the decision
Every report worth its file size contains three numbers: how many units were inspected, how many defects were found in each class, and what the acceptance numbers were for that sample. Find them before reading anything else. The inspected quantity tells you the verdict coverage; the defect counts tell you what was actually seen; the acceptance numbers tell you the bar that was applied. If any of the three is missing, stop reading and request it, because commentary cannot substitute for arithmetic.
Next, check that the inspected quantity matches what the factory actually had ready. Reports sometimes show a sample drawn from a partial lot, with the balance produced after the inspector left. That is legitimate only when the report says so plainly and the buyer accepts the uncovered balance as a separate risk. A verdict that silently covers goods the inspector never saw is the most common way reports mislead without lying.
Finally, read the defect descriptions against the photos, not against your imagination. Each listed defect should map to at least one photograph showing the issue clearly, with the affected quantity noted. Vague defect lines without photos, or dramatic photos without quantities, both break the chain between evidence and verdict. Reports from established providers such as UTS attach photos per defect finding as standard practice, which is one reason their reports survive buyer audits.
Defect classes without the mystique
Critical defects are the non-negotiable class: safety hazards, regulatory failures, or anything that makes the product illegal to sell. One critical defect in the sample fails the lot, full stop. There is no acceptance number for criticals because the allowed number is zero. Reports sometimes bury a critical finding in descriptive text while the verdict line stays cheerful; always search the report for the critical count before accepting any pass.
Major defects impair function, durability, or salability without creating danger: a dead button, a seam that splits under normal handling, a finish that fails within days. Minor defects are cosmetic issues within agreed limits: slight color variation, small surface marks, uneven stitching that does not affect strength. The boundary between major and minor belongs in the product specification, not in the inspector improvisation. If your specification never drew that line, expect every report to draw it somewhere slightly different.
Watch how the report accumulates each class separately. A lot can pass on minors while failing on majors, and the verdict must reflect the worst class outcome, not an average. Buyers who skim for a total defect count miss this entirely. Read the three class lines as three separate verdicts that combine into one, and the combining rule is always the strictest, never the friendliest.
Verdict arithmetic you can redo in two minutes
| Step | Where to find it | What good looks like |
|---|---|---|
| Confirm the sampling basis | Report header or method section | Named standard, level, lot size, sample size |
| Check class counts | Findings table with photos | Critical, major, minor listed separately |
| Compare against acceptance numbers | Verdict section of the table | Each class compared, worst class decides |
| Read the disposition | Verdict and remarks | Accept, reject, or hold with named next step |
| Match photos to lines | Annexed evidence pages | Every defect line has at least one photo |
Questions that close the loop
- Name the sampling basisAsk for the sampling basis in writing: standard, inspection level, lot size, and sample size. File the answer with the report so the next person reviewing the file does not have to guess.
- Disposition failed lots in writingA rejected lot needs a written disposition: rework scope, sorting method, and who pays for re-inspection. Verbal promises to fix it dissolve by the next shipment.
- Define the re-inspection sampleRe-inspection should state its own sample size and whether it covers the full lot or only the reworked portion. A re-check of three cartons is not a new verdict on three thousand.
- File evidence with the verdictPhotographs, tally sheets, and calibration notes belong with the report in one folder. A verdict without retrievable evidence cannot be defended to a customer or a regulator.
Keep one habit across every report you read: reproduce the verdict from the numbers before accepting it. Two minutes with the sample size and the class counts either confirms the inspector arithmetic or starts a useful conversation. Readers who do this stop fearing reports, because the document stops being a certificate and becomes what it should be, a checkable record.
Sources
- Wikipedia: Acceptance sampling —— Overview of acceptance sampling including AQL concepts used in product inspection.
- Wikipedia: Factory acceptance test —— Overview of factory acceptance testing performed before shipment or installation.
Frequently asked questions
What do I check first when a report lands in my inbox?
Check the sample size table first, then the defect counts by class. A verdict without those two is an opinion, not a result. Ask the inspection provider for the sampling standard and level used, and confirm the inspected quantity matches what the factory had ready on the day.
What is the difference between critical, major, and minor defects?
Critical defects threaten safety or regulatory compliance, major defects impair function or salability, and minor defects are cosmetic issues within agreed limits. The report should define its classes or reference the product specification. When classes are undefined, ask for definitions before accepting the verdict.
How do I verify an accept or reject verdict myself?
Compute the defect rate on the actual sample, compare it with the acceptance number for that sample size, and check whether any single critical defect triggered an automatic fail. Sampling plans accept a lot based on defects found in the sample relative to allowed numbers. A report that skips this arithmetic is asking for trust instead of showing work.
Does the report replace laboratory testing?
No. A factory acceptance test checks the finished consignment against agreed criteria before release, while a lab test analyzes samples under controlled conditions. The report tells you whether these goods match the criteria; the lab tells you whether the design and materials meet the underlying requirements.
Ready to design your yard?
Join 312,000+ readers and get our weekly Idea Book — photo-rich plans filtered to your zone, sun, and budget.
Get the Free Idea Book