Why bulk sometimes differs from the sample — and the protections that close the gap
By Mohamad Sinno, Charcoal Expert · Last updated
Ask AI for an importer brief:
The short answer
Some of the reasons are arithmetic and some are dishonest, and a buyer who cannot tell them apart makes one of two expensive mistakes: accepting a substitution because variance is normal, or dropping a competent factory over a deviation that was inside the tolerance all along. The separator is not how large the difference is. It is whether the deviation can be traced to an identified batch with a record behind it, and whether it sits inside a tolerance somebody wrote down before the order.
Legitimate causes, and the other kind
| Cause | Class | What is actually happening |
|---|---|---|
| Population size | Honest variance | A sample is a handful of pieces; a container is thousands of kilograms. Any measured spread widens with the sample size, and a single hand-picked piece sits at the good end of it. |
| Moisture uptake | Honest variance | Moisture responds to storage time, weather and the conditions on the packing day. It is the parameter most likely to differ between a parcel weighed in June and a container stuffed in August. |
| Press and cutter wear | Honest variance | Dimensions and therefore packing counts drift across a long run as tooling wears. This is what a stated dimensional tolerance is for. |
| Feedstock batch variation | Honest variance | Shell from different lots carbonises slightly differently, which shows up in ash first. It is bounded by testing rather than eliminated. |
| The curated sample | Not variance | Pieces selected by hand from a run, or pressed specially, so the parcel is not drawn from anything that will be produced at volume. |
| The substituted grade | Not variance | A different material shipped against an approved sample — a different feedstock, a different mix, a lower specification with the same shape. |
| The borrowed certificate | Not variance | A report that belongs to a different batch, a different product or a different company, travelling with goods it never described. |
The four in the first class are bounded by process rather than removed by it, and the three in the second are not variance at all — they are a different product, or a different document. The patterns behind the last two, and the tells that catch them earlier, are on charcoal supplier scams: the patterns.
What varies most, and why
Moisture moves; ash and fixed carbon largely do not. That single fact organises most sample-versus-bulk arguments. It is also why our own retained samples come with a caveat rather than a flourish: Ash and fixed carbon remain stable in open storage; moisture drifts, and any comparison against a retained sample has to say so. A buyer comparing a container against a sealed sample from three years ago is on solid ground for ash and on soft ground for moisture, and should say which they are doing.
Moisture is also the parameter the shipping regime touches. Charcoal is held after production and before packing rather than pressed and boxed the same week — a two-stage weathering regime of at least fourteen days of raw-material curing and at least fourteen days of post-production hold, at least twenty-eight days in total, with the material's temperature recorded on the packing day. That practice exists for a dangerous-goods reason rather than a quality one, and it has a quality consequence: what leaves the factory has spent time in the open air, and what arrived in your sample envelope may not have. UN 1361 and IMDG 42-24 covers the regime itself.
Dimensions drift second. Which is what a per-side tolerance is for: ours is ±0.3 mm per side, and a clause that names it converts a disagreement about size into a caliper reading. Packing counts move with the dimensions, so a count that comes in under the nominal figure is usually a dimensional story rather than a weight one.
The protection stack
Four layers, and they are worth understanding as a sequence rather than a menu: each one catches what the one before it let through. The first two are things a factory does and you ask to see; the second two are things you do.
- Sampling inside the process, not only at the end. In our plant samples are drawn at 6 stages — carbonization, screening, grinding, mixing, extrusion, drying — at random, at least five pieces per 50 kg, recording weight, ash content, moisture, ash colour. A deviation caught at extrusion is a reworked batch; the same deviation caught at the port is a claim.
- A file per batch, and a certificate that names its batch. The batch file holds the laboratory COA and the factory’s own results for the batch: ash content, moisture, burning time, drop test, cracks, temperature. A COA is issued for every production batch, so "which batch" is always a question with an answer. The drying log and the container loading log are not published, and no sample page of either has been supplied.
- A code on the carton that ties the goods to that file. Every carton carries a four-digit batch production code, two letters identifying the packing team, and one check digit. The batch code identifies the production batch whose QC file holds the laboratory COA and the factory’s own results. Inner-box marking: not recorded. This is the layer buyers most often discover they are missing: without it, a complaint is about a container, and a container is not a unit anybody tested.
- An inspection before the balance moves. The quality-control trade is consistent on this and it is the one recommendation worth taking literally: an inspection has leverage only while a payment is still outstanding, so hold part of the balance against a pre-shipment inspection passing, in writing (cmgm.net, quality-inspection service, 19 July 2026; rich-bee.com, sourcing agent, 17 June 2026). The lighter version of the same idea is a shipment sample drawn a few days before loading to confirm batch consistency (goodcantrading.com, sourcing agent; modified stamp 1 July 2026). Both are trade conventions rather than standards, and they are cited as such.
A fifth layer is the drop test, because it is the acceptance criterion a clause can point at and a buyer can repeat.
- Cube. The piece is heated until glowing red, held five minutes after heating, then dropped five times from 1.2 m onto a hard tiled surface. It must not break.
- Hexagonal. Three drops from 50 cm.
What honest variance looks like on paper
It looks like a certificate for the batch that actually shipped, with each value beside the method that produced it, and figures that sit inside a range somebody published before the order rather than after the complaint. Our own published figures are stated as a range across held certificates rather than as a single number for exactly this reason — a supplier who quotes one decimal place for a natural product is describing one test, not a production line. The certificates and the range are on the certificate of analysis page, and what a certificate settles and what it does not is on how to read one.
It also looks like a supplier who answers "which batch?" without hesitating, and who can produce the certificate for that batch rather than the certificate they show everybody. That question, asked once at sample stage, tells you most of what the protection stack above would have told you later.
If it has already happened
Document before you complain, and document against the batch code rather than against a memory of the sample: photographs of the cartons, the code, the sealed counter-sample beside the delivered goods. Then read the window in your own contract, because it is usually shorter than people expect. Ours is stated: within 7 days of receipt; after that the goods are deemed accepted. The remedy is replacement or credit on the next order, and compensation beyond that is settled by agreement in a later shipment, not in cash. The channel is WhatsApp or the export desk email.
If the supplier stops answering rather than disputing, that is a different problem with a different first hour — what to do in the first 72 hours covers the sequence, including how to tell a fraud from a dispute early.
Questions
Is a difference between the sample and the bulk always dishonesty?
No, and treating it that way costs buyers good suppliers. A sample is a handful of pieces and a container is twenty tonnes, so some spread is arithmetic rather than intent. What separates honest variance from a substitution is whether the deviation is inside the stated tolerance, whether it traces to a batch with a record behind it, and whether the supplier can produce the certificate for the batch that actually shipped.
What actually changes between a sample and a container?
Moisture is the parameter most likely to move, because it responds to storage, weather and how long the goods sat before stuffing. Dimensions and packing counts drift with press and cutter wear across a long run. Ash and fixed carbon are the most stable, which is why a retained sample can still answer for them years later while a moisture comparison against the same sample cannot.
Can I inspect before the balance payment is due?
That is the arrangement the quality-control trade recommends, and it is worth writing into the contract rather than assuming: a pre-shipment inspection that happens after the balance has moved has no leverage attached to it. A shipment sample drawn a few days before loading is the lighter version of the same idea, and it is a convention buyers can ask for by name.
How do I prove what arrived was not what was approved?
Against the batch code on the cartons, not against a general impression. Each carton carries a printed batch production code, and that code identifies the batch whose file holds the laboratory certificate and the factory results. Photograph the cartons and the code, compare against your sealed counter-sample, and raise it inside the claims window: ours is seven days from receipt, after which the goods are deemed accepted.
Before the first order
Score the sample against written criteria — the scorecard method — and then make the approved sample the named reference in the contract, using the golden sample clause. Requesting a sample and the certificate for its batch: free charcoal samples, or WhatsApp +62 821 3924 038. What a sample can prove on its own, and what it cannot, is where the guide to charcoal samples begins.