The fire certificate was genuine, but it belonged to an earlier batch of foam, and 70 of the 240 lounge chairs were upholstered from a second mill that had no test evidence at all
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 line nobody could check on site
A European hotel group refurnishing its lobby lounge areas ordered 240 upholstered lounge chairs for delivery across three properties. The chairs were specified at $85 each, so the order was $20,400. 30 per cent was paid on order, $6,120, and the remaining $14,280 was held against a pre-shipment inspection report.
Most of the specification was checkable on site with a tape measure and a bathroom scale. Frame sections, seat height, arm height, the cover colour against the approved swatch, the carton count against the packing list: each of those has a number and a way to measure it, and an inspector can decide it on the floor of the factory.
One line could not be checked that way, and it was two lines in the material schedule. Fire performance of the completed upholstered chair to EN 1021-1 and EN 1021-2, current test report to be provided before shipment. The requirement is specific, it is the requirement a hotel buyer is right to write, and there is no way to settle it in a warehouse with the goods still wrapped. It can only be settled by a test report, and the report has to belong to the material that is actually inside the chair.
The factory did have a report. The laboratory was real, the report number was real, and it did not cover the goods in the container.
This is an illustrative example drawn from the pattern behind most fire-certification complaints on upholstered furniture: the paperwork is genuine, the material is real, and the link between the two is missing. The figures are representative of the category rather than taken from a named customer file.
Where to go next
→ our seating verification scope, covering foam, covers and frame checks against your specification
→ what a report records when a requirement could not be verified
Three documents, and only one of them had a test result on it
The factory's documentation arrived as three sheets. The first was a test report from a third-party laboratory, dated four months before the order, describing flexible polyurethane foam of 25 kg per cubic metre, moulded, with a flame retardant additive, tested to FMVSS 302 with a horizontal burn rate of 65 mm per minute.
The second was a safety data sheet for the same foam. The third was a supplier's declaration that the foam was self-extinguishing and conformed to EN 1021-1 and -2. All three are the kind of paperwork a well-run factory sends without being asked, and all three are easy to mistake for evidence.
A safety data sheet is a chemical handling document. It is written for storage, transport and first aid, and it says nothing about how a material behaves in a fire in a hotel room. A foam with a perfectly complete safety data sheet has told you what is in it and not what it will do. A supplier's declaration is a statement by the party selling the foam, made in the seller's own interest, with no result behind it.
The report was the only document carrying a laboratory's name, which is exactly why it was the one everybody looked at. It was also the only document that had been tested against the wrong standard, and the two pieces of paperwork that could have been dismissed as soft evidence were the only ones that mentioned the right standard, in words, without a result attached to it.
Where to go next
→ what a report records when the count and the documents disagree
Why a car interior standard is not a furniture certificate
FMVSS 302 is the flammability requirement for materials used in the interior of passenger vehicles, and its horizontal burn rate limit of 100 mm per minute was set for the interior of a car. A hotel lounge chair is not a car seat. An automotive interior material does not become a furniture flammability certificate by being printed on a laboratory letterhead with a real report number on it.
The order had asked for EN 1021-1 and EN 1021-2 on the completed chair, which are the two methods used for upholstered seating in Europe: one for smouldering behaviour, one for a small flame ignition source. They are performed on the seating as a construction, cover, filling, barrier and seam together, because the assembly is what burns and the components can hide each other's weaknesses. A moulded foam slab is not that assembly.
The gap was not academic. The report described a foam of 25 kg per cubic metre. The order specified the seat cushion foam at a minimum of 33 kg per cubic metre, and the two delivered batches weighed more than either. So the one document with a laboratory's name on it described a material lighter than the material ordered, lighter than the material delivered, and tested to a standard that does not apply to the product at all.
That is the shape this failure takes more often than not. The factory is not hiding anything, the paperwork is not forged, and nobody has misread a document. A laboratory report exists, it is for a real product from a real supplier, and the connection between it and the container was simply never made because making it was nobody's job.
| What the order required | What the report on file was | Where that leaves it |
|---|---|---|
| Completed upholstered chair tested to EN 1021-1 and EN 1021-2 | A moulded foam slab tested to FMVSS 302, an automotive interior standard | Does not meet the requirement. A component test is not a seating test, and the cover, filling and seam are not in it |
| Foam at a minimum of 33 kg/m3 for the seat cushion | Foam described at 25 kg/m3 | The report describes a lighter material than the one ordered |
| Current report for the material supplied on this order | Dated four months before the order was placed | Correctly dated for nothing in this shipment |
| Report to identify the batch of foam supplied | A sample description carrying the mill's earlier production code | The line that decides the case. This is the one that cannot be argued around |
Where to go next
The batch number was on the offcuts, which is where it always is
Moulded foam arrives at an upholstery factory as blocks, and every block is marked by the mill with a batch code and a date. The cutting table cuts seat cushions and back cushions out of those blocks, and the offcuts stay on the table, still carrying the mill's marking, because the offcut is useful for small parts and nobody throws a marked board away.
That marking is the whole case. It takes a request to the cutting supervisor, five minutes at the offcut bin and one photograph of a marked offcut against a delivery note to establish which batch the cushions in front of the inspector came from.
The delivery note for this order named two batches. Mill A, code A-2455, 170 chairs' worth of foam. Mill B, code B-8890, 70 chairs' worth. The test report named mill A, batch A-2241, produced some months earlier. A-2455 and A-2241 are the same mill and the same product grade, which is why the factory treated the report as covering the shipment, and they are not the same material, because a test report belongs to a defined sample from a defined batch.
So the reconciliation at the end of the day was short. 170 chairs carried a batch the factory could point at and no report that mentioned it. 70 chairs carried a batch from a second mill that had no test report of any kind, only a safety data sheet and a declaration, and a supplier declaration does not get better by being the only document in the file.
| Document in the file | What it names | What it actually covers |
|---|---|---|
| Test report from the third-party laboratory | Mill A, batch A-2241, foam at 25 kg/m3 | A moulded foam slab, tested to FMVSS 302, four months before the order |
| Safety data sheet | No batch identified | Composition and handling data for the mill's standard grade |
| Supplier's declaration | No batch identified | A written statement that the foam is self-extinguishing, with no result attached |
| Foam delivery note for this order | Mill A batch A-2455 and mill B batch B-8890 | The material actually delivered, 170 chairs and 70 chairs |
| Marked offcuts from the cutting table | A-2455 and B-8890 | The material in the cutting room on the day of inspection |
Where to go next
A hundred millimetre cube and a bathroom scale
There is one measurement that can be taken in a warehouse and is worth more than any document: the density of the foam, taken from the actual cushion core. A 100 mm cube cut from the block gives exactly one litre, so the mass in grams on the scale is the density in kilograms per cubic metre, with no conversion and no reference to a table. Two inspectors can cut a cube and get the same number, and the number cannot be argued with.
Cubes were cut from the offcuts of each batch, where the mill's batch marking was still visible beside them. Batch A-2455 came back at 34 grams, which is 34 kg per cubic metre and comfortably over the 33 kg minimum the order set. Batch B-8890 came back at 30 grams, which is 30 kg per cubic metre and 3 kg under the minimum, a 9 per cent shortfall on the material the order was paying for.
That single number did more work than the whole file of paperwork. It proved the delivered foam was not the foam in the report, and it separated the order into two populations, because the 30-gram material was not spread evenly across the 240 chairs. It had arrived on a second delivery late in the production run and was used to finish the order, and it was in the cushions of 9 of the 32 chairs in the sample.
Density is not a flammability test, and it should not be presented as one. A 34 kg per cubic metre foam is not fire rated because it is heavy, and a compliant high-density foam can fail EN 1021 if the cover is wrong. What density gives you is a fingerprint, and a fingerprint is what was missing: it told the inspector that two different materials were in one order, and it told the buyer which chairs those were.
| Foam batch | Chairs in the lot | Cube mass per litre | Against the 33 kg/m3 minimum |
|---|---|---|---|
| A-2455, mill A | 170 of 240 | 34 g on the scale, 34 kg/m3 | Over the minimum, but no report names this batch |
| B-8890, mill B | 70 of 240 | 30 g on the scale, 30 kg/m3 | 3 kg/m3 under the minimum, and no report of any kind |
| Batch named on the test report | None of them | A-2241, described at 25 kg/m3 | Not the material in this container |
Where to go next
What on-site verification can settle, and what it cannot
This is the part buyers most often get wrong in both directions. A verification visit cannot certify fire performance, because EN 1021-1 and -2 are laboratory methods performed on defined specimens under controlled conditions, and nobody settles them with a lighter and a chair in a warehouse. What on-site work can do is decide whether the evidence exists, whether it is the right kind of evidence, and whether it belongs to the material in front of you.
That is more than it sounds like, because those three questions are where the failures actually live. An inspector can confirm that a report exists, read what it tested and to which standard, and confirm that the sample description is consistent with the material specified. An inspector can confirm the batch by the mill's marking on the block or the offcut and reconcile it to the delivery note. An inspector can take the density of the foam, the thickness of the barrier, the fibre content of the cover and the construction of the seam. None of that needs a laboratory, and all of it is the difference between a report that belongs to the goods and one that does not.
What the visit cannot do is create a test result. If the evidence for the delivered batch does not exist, the honest finding is that it does not exist, and the remedy is a test, not an opinion. An improvised burn test on a chair in the factory yard settles nothing: it is not EN 1021, it is not repeatable, it is not admissible, and it damages saleable goods to produce a number nobody can rely on.
The right answer for the buyer is to arrange the test rather than to argue about the paperwork. Sealed witness samples cut from each batch at the factory, labelled with the batch code and the date, give the buyer's own laboratory material that provably came out of the container. That is the point of witness sampling, and it is the one clause in the specification that would have made this case cost a test instead of a re-order.
- Whether a report exists, what standard it was tested to, and whether that standard is the one the order requires
- Whether the material named in the report is the material specified, on density and construction
- Whether the batch on the report is the batch in the goods, read off the mill's marking on the block or the offcut and reconciled to the delivery note
- The density of the foam in the cushion core, measured on a 100 mm cube, and the thickness and position of any barrier
- The fibre content of the cover and whether the cover itself carries the flammability evidence the market requires
- Sealed, labelled witness samples cut from each batch, so that the buyer's own laboratory can test what actually shipped
Where to go next
→ our seating verification scope, covering foam, covers and frame checks against your specification
→ what a report records when a requirement could not be verified
How the report was written
The report was structured around the requirement rather than around the goods, because the buyer had written the requirement and would read the report against it. Fire performance was graded first, and it was graded as a critical defect at AQL 0, where one finding rejects the lot, in the same class as an uncertified gas lift or a tempered glass without a mark. The report recorded the three documents, what each one actually is, the two batch codes found on the goods, and the fact that no document in the file named either of them.
Density was graded as a major defect at AQL 2.5. The lot was 240 chairs, which is code letter K under ISO 2859-1 at general inspection level II, so the sample was 32. A sample of 32 at AQL 2.5 accepts 2 and rejects at 3, and 9 of the 32 sampled chairs were below the 33 kg per cubic metre minimum, so the lot was rejected on that characteristic as well.
The report said what the evidence did not establish, and it did not say the foam was flammable, because nobody tested it and nobody should have implied otherwise. What it said was that fire performance could not be verified for 170 of the 240 chairs because the report on file belonged to a different batch of foam, and that for the other 70 it could not be verified at all because no test evidence existed for the material in them. Those are different findings and they carry different remedies.
The recommendations were specific rather than general. Segregate the 70 chairs carrying batch B-8890, which is 3 kg per cubic metre under the ordered minimum and has no fire evidence. Test one model on a witnessed sample from batch A-2455 at an independent laboratory, EN 1021-1 and -2 on the completed chair, because one test covers all 170 chairs if the material and the construction are identical, and the cost is per model rather than per chair. Release the $14,280 balance only when the report naming the delivered batch and the test result are both in hand.
The cost of the verification that produced all of this is two man-days of full verification at $299 a day, so $598 against a balance of $14,280. That comparison is the argument for booking the inspection rather than relying on the factory's file, and it is the argument that had to be made before the container was closed, because after it is loaded none of it is reversible.
Where to go next
→ how to read an AQL sampling plan before the goods are opened
What the specification should have said
The order was not badly written. It asked for the right standard on the right object, and if the factory had held a report for the batch it delivered, the case would not exist. What was missing were the lines that make a certificate bindable to a container, and those lines are the difference between a fire certificate and a piece of paper about fire.
Seven lines in the material schedule would have caught this, at no cost, before a single chair was upholstered.
| Line | What to write | What it stops |
|---|---|---|
| Standard, by destination market | EN 1021-1 and -2 on the completed chair for the European properties, and California TB 117 Parts 1 and 2 on foam and cover for the California property | An automotive interior standard offered as a furniture certificate |
| Level of the test | Chair level, with the cover, filling, barrier and seam identified in the report | A moulded foam slab report offered as a seating report |
| Batch-locked evidence | The report must state the batch code of the foam supplied on this order, and the foam mill and grade code must be named | A genuine report for a production run that happened months earlier |
| No substitution | A change of foam mill, grade or batch requires written approval before cutting, and a new report | Two mills and three batches in one order, one of them with no evidence at all |
| Witness sampling | Two labelled samples cut from each batch at the factory and sealed, one retained by the factory and one dispatched to the buyer's laboratory within 7 days of loading | No way to test the material that actually shipped, ever |
| Density, with a method | Minimum 33 kg/m3 in the cushion core, verified on a 100 mm cube weighed on a scale reading to 0.1 g | A lighter grade filling the order and passing every visual check |
| Evidence before payment | Balance released against the batch-locked report and the witness test result, not against a supplier's declaration | Paying a balance against evidence that belongs to another consignment |
Where to go next
What to carry to the next order
Nothing on this order was mislabelled and nothing was fabricated. The factory built 240 lounge chairs to the drawing, upholstered them in the correct cover, packed them correctly and produced paperwork that a hotel buyer would reasonably have accepted as evidence of fire performance. One certificate was genuine, 70 chairs carried foam under the ordered minimum, and not one chair in the container could be traced to the report.
The thing worth carrying away is that a certificate is not a material, a material is not a batch, and a batch is not the goods. Each of those three links is broken by a different routine: the standard by a careless substitution, the material by a lighter grade, the batch by a reorder that nobody documented. A hotel furniture inspection that reads the report and stops there has checked that the factory owns a piece of paper.
Two habits cover all three. Ask for the batch code on the offcut and reconcile it to the delivery note, which costs five minutes at the cutting table. And cut a 100 mm cube and weigh it, which costs five minutes and a bathroom scale and separates two populations in an order that looks completely uniform. If the report does not name the batch that came off the table, the report is not about the chair, whatever laboratory printed it.
And if the paper does not add up, the answer is not an argument and it is not a burn test on the factory floor. The answer is a witnessed sample, sealed and labelled at the cutting table, and a test at an accredited laboratory. That is a few hundred dollars per model. The balance on this order was $14,280, and the container was still open when the report was written.
Where to go next
→ our seating verification scope, covering foam, covers and frame checks against your specification
Have an order you need verified? Send us your PI and factory address. Three levels, published prices, $149 per man-day from.