The frame was MDF, and the warranty said solid wood
Illustrative example This case is written from the patterns we see in furniture export orders, and the figures are representative rather than from a specific client file. It shows how a check works and what it finds. We will replace it with a real, anonymised case as soon as we can publish one with the buyer's permission.
The order, and the one phrase that decided the case
A European homeware distributor ordered 120 three-seat sofas from a factory in Foshan at $340 a unit, an order value of $40,800. Thirty per cent was paid as a deposit. The remaining $28,560 was due against the pre-shipment inspection report.
The frame specification was one sentence long. Solid wood frame, kiln dried. Species was not named, sections were not given, there was no clause saying what solid wood excluded, and no instruction that any member would be opened.
The care and warranty card in every carton was longer and more specific than the specification. Solid wood frame. Ten year structural warranty. It was printed, placed in every carton, and there was no way to check it against the goods without taking the sofa apart.
Everything the buyer's inspection sheet asked for passed. Model, colour mix, unit count, carton count, fabric composition, foam density, overall dimensions, leg and accessory count. All conforming. The report would have gone out as a pass.
This is an illustrative example drawn from a recurring pattern in upholstered furniture orders, where a structural requirement is written as a single adjective and a marketing claim is printed on a card that goes into every carton. The figures are representative of the category rather than taken from a named customer file.
Where to go next
→ our sofa order verification scope, including frame and structure checks
What a frame check on a finished sofa cannot see
The frame of a three-seat sofa is mostly hidden. You see the outside of the arms, the outside of the legs, and the front rail only on a frame-and-panel construction. Everything else is behind the upholstery and under a dust cover.
In this order the visible members were all solid. The side rails were solid beech, the legs were solid beech, and the outside face of the front rail was solid. A rail that is solid on the outside and pressed board on the inside looks the same from every angle a showroom allows.
The approved sample on the buyer's side was the problem in miniature. It was a showroom unit with the dust cover sewn closed, no shipping carton, no legs fitted and no route to the underside. Comparing a delivered sofa against that sample meant comparing upholstery, finish and dimensions, because nothing else about the sample was observable.
So the frame passed, and it passed for a reason nobody had chosen. It passed because the only members the inspection could see were the ones the factory had built correctly.
Where to go next
→ our sofa order verification scope, including frame and structure checks
The moment the bottom cover came off
Legs ship loose. They are bagged in the carton and fixed into the frame through the underside, which means the dust cover has to open. On the third unit of the day the inspector opened the cover to check that the leg fixing points were sound and that all four legs and the hardware pack were present. It was a routine step, and it was not on anyone's list of things to look at.
Under the cover, at the front seat rail, the cut end of the rail was visible. It was a flat uniform brown face with no growth rings, no vessel structure and no grain direction. It looked like pressed wood fibre with a sealed surface. The trade calls that MDF, density board, or in some factories simply board, and it is not a piece of a tree.
The inspector did not cut anything and did not take the sofa apart. The end of that rail was already a cut face. It was the factory's own cut, made when the rail was cut to length, and it was showing.
What happened next is the part of this case that matters, because the same evidence can be gathered in a way that destroys the goods and in a way that does not.
Where to go next
→ our sofa order verification scope, including frame and structure checks
Taking the evidence without taking the goods apart
Four things were done, in this order, and none of them damaged a saleable sofa.
First, the exposed rail end was photographed with a 100 mm scale in frame on three units, and the board-lot stickers on the parts were photographed where they had survived. Second, the factory was asked for the offcut from the board lot the rails were cut from. Rail offcuts are waste, they are stacked beside the cutting machine, and they are the same material as the rail because they came off the same board. One 300 mm length was taken, weighed and cut across. Third, the same offcut was used for the fastener test. Fourth, finished units were weighed on a platform scale without packaging against the approved sample unit weighed the same afternoon.
The offcut is the point of the whole method. It identifies the material of the batch without touching a saleable unit, and it leaves the evidence in the buyer's hands rather than in a factory's promise.
| Check | Specification or approved sample | Delivered | How it was measured |
|---|---|---|---|
| Front seat rail section | 30 × 60 mm solid beech | 16 × 60 mm pressed board | Calipers at the exposed rail end, three positions, three units |
| Mass of a 300 mm length of rail | 378 g, calculated from the specified section | 216 g, weighed | 300 mm offcut from the same board lot, weighed to 1 g |
| Rail mass per linear metre | 1.26 kg/m | 0.72 kg/m | Offcut weight scaled to 1,000 mm; specified figure calculated at 700 kg/m³ for kiln-dried beech |
| Cross-section appearance | Continuous annual rings, visible vessel structure | Uniform fibre structure, no rings, resin-sealed face | Cut across the offcut, photographed with a 100 mm scale |
| Screw pull-out from a rail edge | 260–360 N | 40–70 N | 4 × 40 mm screw driven into the offcut edge, three pulls, mean reported |
| Unit weight excluding packaging | 41.2 kg | 40.0 kg | Platform scale, three units of each, weighed the same afternoon |
Where to go next
→ our sofa order verification scope, including frame and structure checks
What the numbers actually say
Start with the section, because it decides whether this is a material dispute or a dimension dispute. A 30 × 60 mm rail and a 16 × 60 mm rail are not two grades of the same thing. One is a structural member. The other is a sheet good with a veneered face, and it is thinner by 47% in the direction that carries the load.
The 378 g against 216 g pair is the same fact seen from another angle, and it is the measurement that makes the substitution arguable at a desk without anyone having to argue about what fibreboard looks like. 378 g is what a 300 mm length of 30 × 60 mm kiln-dried beech weighs. 216 g is what the delivered offcut weighs. Material and section both sit in that gap, and the gap cannot be explained by moisture or by finish.
The specific gravity route is worth naming because it does not work here. MDF at 16 mm sits around 750 kg/m³ and beech sits around 700 kg/m³. Denser is not better, and a density reading alone cannot separate the two. What separates them is the section, the pull-out strength and the structure of the cut face, which is why the offcut was weighed as a length rather than as a block.
The pull-out numbers are the most useful field result, because they are destructive on a piece of waste instead of on a sofa. A 4 × 40 mm screw into the edge of a 30 mm solid beech rail held between 260 and 360 N. The same screw into the edge of the 16 mm board held between 40 and 70 N, and the edge crumbled visibly as the screw was withdrawn. A joint that depends on a screw biting into end grain and fibre is the whole reason a frame fails later, and on this offcut it failed at one tenth of the strength.
The unit weight is the weakest of the six and it is here on purpose. Each sofa weighed 1.2 kg less than the approved sample unit, which is 2.9% of the weight of the sofa. On its own that number would have justified nothing: furniture is not weighed to a tolerance, upholstery is not uniform, and buyers have signed off lighter chairs without comment. It became evidence only after the offcut had already identified the material, and at that point it did one useful thing, which was to show that the substitution ran through the whole production run rather than being a single unit built wrong.
Where to go next
→ our dining verification scope, including wood species checked against your sample
→ our sofa order verification scope, including frame and structure checks
What the report recorded, and what it did not
The report recorded the exposed cut face and its photographs, the weighed offcut and its section, the screw pull-out results, the unit weights, and the board-lot information the factory supplied. It recorded that the front seat rail and the rear cross rail on the three units examined were pressed board, and that the side rails and legs on those units were solid beech.
It did not establish that every one of the 120 sofas was built the same way. Only three units were opened. On the other 117, the report could say nothing at all about the frame, because the frame was not visible and no other member had been opened. That sentence is the most important line in the document, and it was written in.
It did not establish the moisture content of the delivered rail, because a moisture meter needs a face and a surface, and a rail that is 16 mm and sealed is not the same thing as a rail that is 30 mm and can be measured. It did not load-test the frame: a rail at 16 mm with a 40 to 70 N pull-out strength is a structure that will fail in service, but failure is not something an inspector can time on a factory floor in one day. It did not verify whether the plywood in the crate and partition panels matched the specification, and it did not certify the lot against any standard.
The lot was held as non-conforming, and it was held because the buyer's own defect list classes material substitution in a structural member as critical. What that defect list did not do was define solid wood, and the definition the factory used to decide the rail was acceptable was the one nobody had written. A classification decides what happens after you find something. It does not help you find it.
Where to go next
→ what a report records when a lot passes and a lot fails
Why the factory did it
The factory was not hiding anything. Nobody had asked for solid beech by name, nobody had written a section, and no document had ever said what solid wood excluded. When a frame is described in a single adjective, the person who buys the timber decides what the adjective is worth, and that decision is made against the price of the order, because there is no specification to read.
The arithmetic is small. Two rail lengths per sofa, 2.25 m in total, at 30 × 60 mm solid beech is about 2.84 kg of timber per sofa, roughly $2.70 of material at the factory's buying price. The delivered board was about 1.62 kg, roughly $0.73. The saving was close to two dollars a sofa, so about $236 across the whole order of $40,800.
Two hundred and thirty-six dollars was less than one day of a freight dispute and less than a single day of container demurrage. That is the real reason this happened. It was not worth much to the factory and it was worth nothing at all to the buyer, and the whole gap between those two numbers is a sentence that was missing from the purchase order.
The factory could do the job. It had beech in the building already, for the side rails and the legs, and the same cutting, drilling and dowelling setup ran all four members. This was not a production problem, a tooling problem or a capacity problem. It was a specification problem, and the specification was the buyer's document.
Where to go next
→ our sofa order verification scope, including frame and structure checks
What the specification should have said
Eight lines. Three of them are numbers, one is a list of members, one is a piece of evidence, one is a moment in the production schedule, and one is a sentence about the warranty card. Together they replace an adjective with something a report can defend.
A sofa frame inspection that only sees the outside of the sofa is a carton count with better vocabulary. The specification is what decides whether the next one sees more.
- Name the species and the members. Solid wood: kiln-dried beech (Fagus sylvatica) or ash (Fraxinus mandshurica). This clause covers the front seat rail, the rear cross rail, the side rails, the back uprights, the legs and all joining blocks. Every member in this list is solid wood.
- Say what solid wood excludes, and say it twice. Engineered board is not permitted in any member listed above, including members concealed by upholstery, dust cover or packaging. Not permitted: MDF, particleboard, plywood, OSB or veneered panel. There is no exception for hidden members.
- Give a section and hold it. Front seat rail and rear cross rail: not less than 30 × 60 mm at the smallest dimension after machining. Tolerance on thickness is minus zero. Tolerance on width is plus 2 mm.
- Ask for the evidence while it is free. For each board lot used in the members listed above, the factory shall retain the offcut and mark it with the lot number, and shall release it to the buyer or the buyer's inspector on request, so that material and section can be checked on a piece of waste instead of on a finished sofa.
- Move the check to frame stage. Before upholstery, the factory shall present not less than five randomly selected units with the frame unwrapped and accessible, at a date agreed in the purchase order. A hidden structural member cannot be verified on a finished unit, so the buyer has to be given a moment when it is visible.
- Tie the warranty card to the goods. Each carton shall carry a card stating the frame species and construction. Where that card states solid wood, every member in clause 1 is solid wood, and a mismatch between the card and the goods is a defect in the same order as the goods.
- Set a weight tripwire, and say what it is for. Finished unit weight excluding packaging shall be within 2% of the approved sample unit weight measured on the same day. This is a trigger for a look, not a test: the unit in this case came in 2.9% under, which is outside the band, and 2% is workable because both numbers come from the same model on the same day.
- Classify it before it happens. Material substitution in a structural member is a critical defect in this order. The buyer may reject the lot, require rework at the factory's cost, or withhold the balance until the affected units are rebuilt. Write this into the purchase order, where both the purchasing manager and the production manager will see it, because a critical classification the factory has never seen is not a classification.
The reason the factory could comply with all of this is written into clause 1: it had beech in the building already, for the side rails and the legs. The rail was the only member that moved. A specification that names the members, excludes the materials and asks for the offcut costs the factory nothing extra to comply with, and on this order it would have cost the buyer about two dollars a sofa to be certain.
Where to go next
→ our sofa order verification scope, including frame and structure checks
→ when to book an inspection before you pay
Have an order you need verified? Send us your PI and factory address. Three levels, published prices, $149 per man-day from.