Every pre-shipment inspection report looks reassuring. A couple of hundred pieces pulled at random, a form of ticked boxes, a handful of photographs, and an "AQL 2.5 passed" line at the bottom. Then the container arrives, the goods reach a shop or a marketplace, and somewhere in week six a buyer finds the problem the inspection was supposed to catch.
We are on the factory side, so we see both halves of that story: the inspection that passed, and the complaint that followed it. This guide explains what a standard AQL pre-shipment inspection genuinely measures, what it structurally cannot measure on a wooden product, and what to add to an inspection brief so that "pass" means something.
What a pre-shipment inspection actually is
A pre-shipment inspection (PSI) is a visual and functional check of finished goods, carried out at the factory after production is complete and usually after the export cartons are packed. Most inspection companies ask that at least 80% of the order be packed before they attend, so that the check reflects what will actually be loaded.
The inspector does not test the product in the sense a laboratory tests it. They look, they measure, they handle, they count, and they compare what they see against the specification and the approved sample. Almost everything they do is non-destructive and takes seconds per unit, because they have a fixed number of pieces to get through in one day.
That is the whole point of the exercise: a fast, repeatable, statistically framed look at a large lot. It is a good tool. It is just a narrower tool than most buyers assume.
The sampling plan, in plain numbers
Most inspections work to a sampling standard - ISO 2859-1 is the one we see on nearly every order - which converts a lot size into a sample size, and an agreed quality level into an accept and reject number.
The usual defaults on wooden toy orders are: critical 0, major 2.5, minor 4.0. Critical means a hazard, major means the product cannot be sold as it stands, minor is a cosmetic deviation that a customer would notice but tolerate.
| Order quantity | Units inspected (level II, the common default) | Typical accept / reject at AQL 2.5 |
|---|---|---|
| 151 - 280 pieces | 32 | 2 / 3 |
| 281 - 500 pieces | 50 | 3 / 4 |
| 501 - 1,200 pieces | 80 | 5 / 6 |
| 1,201 - 3,200 pieces | 125 | 7 / 8 |
| 3,201 - 10,000 pieces | 200 | 10 / 11 |
Read the table as a buyer and the trade-off is obvious. A single inspection of a 5,000-piece order looks at 200 pieces - four percent. And at AQL 2.5, ten defective pieces in those 200 is still a pass. Whether that is acceptable depends entirely on which defect you are talking about, which is why the classification matters more than the headline number.
What this kind of inspection does well
- Count and identity. The right item, the right quantity, the right size, the right colourway in the right carton.
- Appearance. Scratches, dents, uneven paint coverage, visible glue squeeze-out, dirty surfaces, misaligned printing.
- Assembly and function. Loose screws, parts that do not move as they should, wheels that do not turn, magnets and fittings that are present and seated.
- Packing. Retail box condition, polybag and warning presence, carton dimensions, tape and strapping, the shipping mark.
- Obvious safety items. Sharp edges, protruding fasteners, exposed points, parts that come away with a light pull.
For a factory that has drifted - a paint batch slightly off, a carton undersized, a hinge fitted backwards - a PSI catches it, and that is worth the fee.
What AQL sampling misses on a wooden product
These are the failures we see most often at the factory gate and at the destination, and every one of them can survive an inspection that reports "passed".
1. Moisture content
Wood moves with the humidity around it. A toy assembled and packed in Yunhe in a wet month can leave the factory at a moisture content that is perfectly normal for Zhejiang and much too high for a heated interior in Canada or a dry one in the Gulf. Nothing visible happens for weeks. Then the panel shrinks a little more than the frame around it, a glue line opens, a board cups.
An inspector looking at 125 pieces for four hours will not find this by eye, and a hand-held meter reading on one or two pieces is not normally part of the service. Moisture is a process control, not an inspection item. It has to be specified in the purchase order, measured at the factory on the batch, and recorded - not left to the visual check at the end.
2. Finish adhesion, and what shows up in month three
Paint and lacquer adhesion depends on how the surface was sanded, how much time passed between coats, and how the coating was cured. A finish that is not fully cured feels dry within hours but will still print, scuff or lift when it is packed against another piece in a carton.
The standard checks for this are destructive: a cross-hatch adhesion test, a tape pull, a rub test. They are not part of an AQL visual inspection, and inspectors cannot use them on saleable goods. If finish durability matters for your channel, it has to be evidenced on the sample and on the pre-production approval piece, not on the finished shipment.
3. Small parts, after real play
The mechanical tests that decide whether a part can come loose - torque, tension, drop, tip over - are laboratory tests performed on samples, to EN 71-1 or ASTM F963. An inspector can confirm that a component is present and seated. They cannot tell you whether the glue joint will still hold after a child has dropped the toy forty times onto a tiled floor.
What the factory controls is the process behind it: glue type and quantity, clamping pressure and dwell time, and the age-grade decision that determines which tests the product must pass at all. Ask for the glue specification and the clamping time. Nobody will volunteer it.
4. Batch-to-batch material and colour
Timber is a natural material. Two deliveries of the same species can differ in colour, grain and density, and every dye lot is its own batch. A shipment can pass an inspection that compares each piece against the specification and still be visibly inconsistent when it sits on a shelf next to the previous order.
The fix is a retained standard: a signed physical master sample of the wood tone and the painted colour, kept by both sides, plus a stated tolerance. Inspectors can then compare against a reference rather than against an opinion.
5. Glue lines, joints and internal work
Much of what fails later is inside the product: an under-glued dowel, a screw that bites into end grain, a joint that was clamped for two seconds instead of twenty. None of it is visible in an assembled toy, and an inspector is not permitted to disassemble one.
This is why the in-process checks matter more than the final one. A factory that inspects semi-finished parts - before assembly, and again after painting - catches these at the stage where they can still be fixed. A factory that only checks the packed carton is inspecting its own mistakes.
Which of the three checks you actually need
Buyers often use the words "inspection" and "testing" interchangeably. They are three different jobs, done by three different people, and one does not replace another.
| Who | What they do | What it proves | What it cannot prove |
|---|---|---|---|
| Third-party lab | Destructive and chemical testing to EN 71, ASTM F963, CPSIA | That the sample met the standard on the day it was tested | That the production run matches the tested sample |
| Third-party inspector | Visual and functional AQL sampling of finished goods | That this lot looks and works as specified | Materials, moisture, durability, internal joints |
| Factory QC | In-process checks at cutting, sanding, painting, assembly, packing | That the process is under control | Nothing on its own - it is only as good as its records |
The strongest evidence a buyer can hold is all three, in that order: a lab report for the material and design, an in-process record for the run, and an inspection for the lot that is about to ship.
How to write an inspection brief that finds real problems
Inspectors work to the brief they are given. A brief that says "check quality" produces a generic report. Give them these ten items and the report changes character.
- The specification and the age grade, with the drawing revision number that applies to this order.
- The approved sample, kept by both sides, and stated as the reference for colour, tone and finish.
- The colour standard, with a tolerance - a Pantone reference, or the master piece, not "as before".
- Defect classification, written out. A 2 mm paint run on a handle may be minor on a play kitchen and major on a gift item.
- Moisture content for the batch, with the target range and the meter reading recorded on the report.
- Packing specification: retail box, inner carton, master carton, polybag and warning text, shipping mark, barcode.
- Carton performance to check - drop and stacking tests for the packed carton, or a written statement if it is not required.
- Photographs of anything raised, at the piece and the carton level, plus the opened carton and the packing line.
- Retention: which pieces the inspector should seal and leave behind, so that a later dispute has a physical reference.
- What the inspector may not do - no disassembly, no destructive tests - written down, so nobody promises evidence the service cannot deliver.
What to do when the inspector raises a major
Most first-time buyers react to a failed inspection by asking for rework and a second inspection. That is usually right, but two things are worth insisting on. First, ask what the rework actually was, piece by piece, and ask for photographs of the reworked units rather than of the finished cartons. Second, ask which process step produced the defect. A batch of 5,000 with 30 loose parts is an inspection result; it is also a signal about the assembly line, and if that is not answered, the next order will repeat it.
Ten questions to ask before you book an inspection
- Is the inspection after 100% packing, or at 80%?
- Which sampling standard and which general inspection level are being used?
- What are the critical, major and minor definitions for this product, in words?
- Is a moisture reading included, and on how many pieces?
- Will the inspector measure against our approved sample, or only against a photograph?
- Are carton drop and stacking tests included in the fee?
- How many pieces will be retained after the inspection, and who holds them?
- How are colour deviations classified - minor by default?
- Who signs off a rework, and is the second visit charged?
- Will we receive the raw count data, or only the summary page?
Frequently asked questions
Do I need an inspection if the factory already inspects? They are not alternatives. The factory's own checks control the process; a third-party inspection verifies the result independently. If your order is your first with that factory, or the value is significant, pay for the independent view.
Can an inspection replace lab testing? No. An inspection looks at the goods; a lab looks at the materials and the design against a standard. A passed inspection says nothing about heavy metal migration or mechanical test compliance.
Is AQL 2.5 strict enough? It is a reasonable default for a mature product from a known factory. For a first order, a new design, or a product for young children, tighten the classification rather than the number - make the defects you care about major, so that a small count still rejects the lot.
Why did a shipment pass inspection and still get complaints? Almost always one of four causes: moisture and climate, finish durability, an internal joint, or a difference between the tested sample and the production run. All four sit outside an AQL visual check.
How long should an inspection take? One working day for most wooden toy orders under about 10,000 pieces. If an inspector proposes to finish 200 pieces with functional checks in two hours, ask what is being skipped.
How we handle this at our factory
Every order that leaves Yunhe passes through staged in-process checks - incoming and cut timber, semi-finished parts before painting, finished parts before assembly, and the packed carton - and we record the results. Where a buyer specifies a moisture target, we measure it on the batch and put the reading on the inspection sheet that travels with the order.
We also tell buyers plainly what a pre-shipment inspection will and will not show, before they book one, because an inspection that produces a surprise on arrival is worse than no inspection at all.
Where to go next
- Wooden toy manufacturing process: the ten steps from timber to shipment - where the defects are actually caught.
- Wooden toy sanding: why two to three rounds matter - the step that decides how a finish behaves.
- Wooden toy samples: cost, lead time and approval - what the approved sample should lock down.
- Wooden toy materials compared - how material choice changes how a product ages.
- Wooden toy import requirements by market - the documents each destination expects.
- How to choose a reliable wooden toy manufacturer - twelve checks before you pay for a sample.
What an inspection measures against is the approved sample, and how that sample is signed off is set out on our wooden toy manufacturer page.
What to check before you need an inspection is set out in how to choose a wooden toy manufacturer, and the coating that most failures come from in water-based paint for wooden toys.