Short answer first: a good inspection report gives you the raw observations that let you make the call yourself — quantities counted, cartons selected, defects found, photos with context. A bad report spares you the effort by making the call for you: everything green, nothing opened, few photos, no numbers that tie back to a specific carton. A pass you can audit is worth more than a pass you have to trust, and you can tell the difference in two minutes using the five checks at the end of this article.
Most buyers only learn this after a container arrives wrong behind a report that read “passed.” The report was not false in any single line — it was empty in the places that mattered. This guide walks through what each section of a real pre-shipment inspection (PSI) report contains, and where the rubber-stamp version quietly omits.
Anatomy of a Real Report: The Seven Parts
A competent third-party PSI report is not a certificate. It is an evidence package. Seven parts, always:
1. Identification block — the boring part that is load-bearing
Order number, PO number, product name per the approved spec, factory name and exact inspection address, date, inspector ID, and the quantities ordered vs inspected vs packed. If a report cannot be tied to your PO revision and that factory’s floor, it is a piece of paper describing someone else’s shipment. Check the address matches the production site you approved — factories with multiple plants have been known to invite inspectors to the tidy one.
2. Sampling record — how many, and who chose
The quantity inspected should follow the sampling plan you agreed (for most consumer goods, ISO 2859-1 Level II — the logic is covered in our AQL guide). What matters beyond the number: the report should say which carton numbers were opened, and that the inspector selected them. “Factory presented 5 cartons for inspection” is a sentence to circle in red. Cartons chosen by the party whose goods are being judged are not a random sample, and every AQL statistic in the report rests on that word.
3. Quantity verification — counted, not estimated
Real: carton counts verified against packing list, units physically counted from a random selection of cartons, assortment checked SKU by SKU against the PO. Rubber-stamp: a carton count from a distance and an “as declared.” The gap between “assortment verified” and “assortment as per packing list” is exactly the gap where wrong-color units live.
4. Workmanship results — the defect table
Every unit of the sample gets checked and every defect recorded by class — critical, major, minor — with a running total that you can read against your AQL thresholds. A real table tells you: sample 125 units, 4 majors found, 9 minors found → 4 ≤ 7 (accept majors), 9 ≤ 10 (accept minors), 0 criticals → pass. A bad table tells you: “workmanship: within acceptable limits.” One of those numbers can be re-verified by you; only one of them is a claim.
Watch specifically for zero-defect results. They happen — genuinely good production lines with stable suppliers do exist. But on a first order from a new factory, a 125-unit sample with literally zero minors is statistically unlikely. More often it means the inspection window was used as a guided tour.
5. Measurements and functional tests — numbers, not adjectives
A proper report contains actual measurements: dimensions against your spec with tolerances, weight per unit, thickness where it matters, color reading where specified. And for functional products: the tests that were run, their pass criteria from the spec, and results. “Zipper functions smoothly” is an impression; “20 open/close cycles, 0 failures, zipper brand matches BOM” is a test.
6. On-site checks beyond the product
Packaging and shipping-mark verification against your PO, carton drop tests / stacking checks where agreed, barcode/label scans that actually resolve on a reader, and — a good inspector’s most underrated line — packing progress confirmation: what percentage of the order is finished and sealed at inspection time. An inspection at 60% packed with the rest “in progress next week” is not a pre-shipment inspection at all; it is a mid-production one with the wrong header.
7. Photos — the difference between evidence and decoration
This is where reports are won or lost, and where “timestamped inspection photos” (a phrase every QC firm uses) does actual work. Real photo sets look like this:
Every defect photographed with a reference — a tape measure against it, the defect circled or pointed at, the unit next to its carton label. A photo of “a scuff” at 10 cm with no scale and no context proves nothing.
Carton labels of opened cartons, photographed in the stack sequence, so you can trace which boxes the sample came from.
The line or packing area as found on arrival, sealed-carton stacks before and after, and the loading of a container if that is in scope.
Timestamps embedded in the images themselves, and images that visibly cannot have been taken on a different day or at a different site — the inspector’s hand card with the date in frame is an old trick that still works.
Five Checks You Can Run in Two Minutes
Before you accept a pass, run these against the report. Every one targets a specific way a bad report stays technically true while hiding the shipment.
Do the defect counts tie to the sampling plan? Find: sample size, defects per class, your AQL thresholds, accept/reject numbers. If the report shows “PASS” but never shows the arithmetic — ask for it. Refusal to show arithmetic is your answer.
Which cartons were opened, and who picked them? Carton numbers must appear. If they don’t, or if selection is attributed to the factory, treat the sample as unverified.
Can you see the date in the photos, in a way that can’t be faked? Embedded timestamps, the daily hand card, the same carton labels as the sampling record. A photo folder of fifty pretty product shots taken at factory PR lighting — beautiful, useless.
What percentage of the order was packed at inspection time? Anything under ~90% is a mid-production report. If the balance packed next week, the next week’s output is uninspected — and substitution (see sample vs. bulk mismatch) historically happens exactly then, after the inspectors leave and before the containers close.
Does the report answer your spec’s critical points by name? If your PO calls out five critical dimensions, two safety features and a color standard, the report must contain measurements for all nine, or admit it didn’t check them. A report that checks the easy 90% and is silent on the hard 10% has inspected a different order than yours.
The Two Ways a Bad Report Hides Behind Silence
Most buyers search reports for false statements. The real risk is absent ones.
Silent scope reduction. “Not in scope today: drop test (no lab access), metal detection (not available at site), inner-box print quality (cartons sealed).” Each line is honest and each shrinks the guarantee by another slice. The correct response is not anger — it is specifying, in the PO and booking, exactly which checks are mandatory and refusing to pay for a report that treats them as optional.
Pre-sort contamination. The report may be perfectly honest about what the inspector saw because the inspector saw a curated reality: the pre-sorting and re-packing tactics mean the sample you paid to audit was cleaned for the audit. Countermeasures live in the process, not the paperwork — unannounced windows, inspector-selected cartons, re-counts on sealed boxes before container doors close. (All of it is in the AQL guide — the report tells you whether those countermeasures were used.)
Who Wrote the Report, and to Whom
One structural question most buyers forget to ask: does the inspector work for you, or for the party who scheduled the visit?
A report is only as adversarial as the person who wrote it. An inspection company paid by and dispatched by the buyer verifies for the buyer. An inspection arranged by the supplier, “included free with your order,” verifies something else. When the line “inspector was escorted by factory QC manager throughout” appears — note it; escorting is standard, but who opened the boxes matters and the report should say.
This is the same principle EzySourcing runs the whole engagement on, in writing: we represent the buyer, the factory pays us nothing, and the original records behind our inspection reports are yours on request — the defect lists and photo sets, not just the summary verdict. That is what we mean when we say commitments are put in writing on the Why Us page: if the report cannot be audited line by line by you, the pass is a favor, not a finding.
A Pass Reads Like This: A Skeleton Example
A compressed version of what a real, auditable pass looks like:
Order EZ-2417 (spec rev C) · 4,800 pcs · Factory , Building 3, , verified against business license on site · Lot 1 of 2: 3,200 pcs packed (66.7%); Lot 2 in sewing, scheduled packing 14 Oct — re-inspection required before release · Sample per ISO 2859-1 Level II, code K: 125 units, cartons 214–224 selected by inspector from finished stack · Critical: 0/125 · Major: 3/125 (stitch skip A3, A19, B7 — photos 12–14; acceptance 7, reject 8 → pass) · Minor: 12/125 (loose threads, photos 15–21; acceptance 10, reject 11 → fail at 4.0, escalated: factory offered 100% thread trim; buyer approved rework) · Measurements: 12/12 within ±2 mm (table); color: ΔE 0.8 vs golden sample D65 (pass) · Barcode scan: 6/6 read first pass · Drop test ISTA 1A: carton intact, contents OK · Photos: 47, timestamped, opened-carton labels included · Verdict: HOLD pending Lot 2 inspection and thread-rework verification.
Read the ending twice: the report that protects you is allowed to say hold. A verdict field that only ever accepts “pass” describes a payment receipt, not an inspection.
What This Means in Practice
You do not need to become a QC professional to benefit from this article — you need the two-minute checklist and one habit: treat a report as an evidence package to be interrogated, not a certificate to be filed. Counts tie to plans, cartons trace to numbers, defects tie to photos, scope matches spec, and “hold” is a word you expect to see sometimes.
If you would rather someone else run that discipline on every order — inspector-selected cartons, timestamped evidence, and original records handed to you rather than summarized for you — that is the definition of our quality control service. And if you are waiting on a container right now with a report that reads too green to question: send it over. A second opinion costs nothing, and suspicion about a shipment is exactly the thing worth spending ten minutes on before the balance transfers.
WhatsApp: +852 46401579 · Email: info@ezysourcing.com
Written by the founder of EzySourcing, a Hong Kong-based sourcing firm working exclusively for buyers — after more than a decade advising foreign importers on sourcing and doing business in China.