Most buyers I know do one of two things when they receive an inspection report: check whether it says “Pass” on the cover, or don’t open it at all.
Both are problems.
I’ve been sourcing auto parts out of Guangzhou for years. I’ve seen shipments arrive short, dimensionally off, or with surface finish that didn’t match the approved sample — all after the supplier sent a “passed” inspection report. By the time the container lands at your door, there’s not much you can do.
So here’s what I actually look at when I read one of these reports.
A passed inspection report means: the checked sample met the stated criteria on the day the inspector showed up. That’s it.
Three things determine whether that’s actually meaningful. Who ran it? The supplier’s own QC team, a third-party hired by the supplier, or a third-party you hired yourself. The first two have a conflict of interest. The QC company hired by your supplier has an ongoing commercial relationship with them — they’re not going to fail an order and lose the client. If you’re importing regularly, you should be arranging your own inspections at least occasionally.
When was it done? Pre-production, inline, or pre-shipment? Most suppliers only send pre-shipment reports. That means if something went wrong during production — wrong raw materials, a dimension shift, a coating problem — you’re finding out after the whole batch is made and packed.
What was the sample size? This brings us to AQL.
AQL stands for Acceptable Quality Limit. Every inspection report should list which AQL level was used. If it doesn’t, that’s already a question worth asking.
At AQL 2.5 — the standard for most auto parts — if you’re shipping 1,000 units, the inspector checks 80 pieces. If 6 or more fail, the batch is rejected. This is the baseline minimum for aftermarket parts. AQL 1.0 is stricter. Used for safety-critical parts like brakes or steering components. If your supplier is sending you brake calipers with an AQL 4.0 report, push back. That’s too loose for anything that goes near a wheel. AQL 4.0 is the kind of sampling you’d use for low-risk accessories. Wrong application equals a meaningless report.
The AQL level should reflect what you’re actually importing. If your supplier just sends whatever template they normally use without discussing this with you, that’s a habit worth correcting early.
Now for the content of the report itself. Product identification — part number, OE reference, car model application, carton quantity, total batch size — needs to be cross-checked against your PO line by line. I’ve had orders where the product code on the report didn’t match what was actually shipped. It sounds obvious, but substitution errors happen.
The quantity check matters more than people think. The inspector counts cartons and verifies against the packing list. Watch for discrepancies, even small ones. A 3% short shipment on a 5,000-unit order is 150 pieces. That adds up.
Visual and workmanship inspection covers surface finish, coating uniformity, visible cracks, assembly issues. What matters here isn’t just the defect count — it’s whether there are photos. A report that says “2 minor defects” without close-up images tells you almost nothing. If your supplier’s reports don’t include photos, ask them to start including them.
Dimensional measurements — thread pitch, bolt hole diameter, bracket width, overall dimensions — should be compared against your approved drawing or sample. If the report lists measurements but not your reference specs, you can’t actually verify whether anything passed. Both columns need to be there.
For shock absorbers, filters, sensors, electrical parts, there should be a function test section. What was tested, how, and what were the results. If this section is blank on a report for brake pads, the inspection was incomplete. Don’t accept it as a complete report.
Most suppliers aren’t deliberately dishonest. But I’ve seen enough reports to know where the gaps are.
Cherry-picked samples. Easy to do when the inspector is the supplier’s own employee. They pull the best-looking units before the inspector arrives. Third-party inspectors pull samples randomly from sealed cartons — that’s the main reason they’re worth paying for.
Old reference sample is a common one. The report says goods were compared to “approved sample.” But that sample is from the first order 14 months ago. The current production run uses a slightly different casting or a cheaper coating — no formal change notice was issued, and nobody flagged it. This happens more often than it should.
A real inspection report classifies defects as Critical, Major, or Minor. If the report just says “defects found: 3” with no classification, you don’t know if any of those defects would cause a part to fail in service. This is basic — if a supplier’s QC team doesn’t know this distinction, that’s telling.
One thing buyers don’t account for: at AQL 2.5, a batch can technically pass while containing up to 2.5% defective units. On a 5,000-unit order that’s potentially 125 bad pieces. This isn’t a supplier trick — it’s how statistical sampling works. Factor it into your buffer stock, especially on high-volume SKUs.
If something looks off in a report, ask for the underlying data — photos, measurement sheets, test results. A supplier running a proper QC process has this. If asking for a specific section creates friction, that’s useful information about the supplier.
If you’re consistently getting thin reports — no photos, missing dimensions, no AQL reference — arrange your own inspection on the next order. QIMA, Bureau Veritas, and SGS all operate across Guangdong. On a €15,000 shipment, a pre-shipment inspection runs €200–350. That’s not overhead. That’s the cheapest way to protect your margin.
Most suppliers are using the same template they’ve always used, checking what they’ve always checked. If you want a useful inspection report, send them a checklist: which dimensions to measure, what surface standard to compare against, how to classify defects. Most suppliers will follow it. The ones who push back are the ones worth paying attention to.
