Samples and Sample Approval Process

Every sample request I've answered over the years arrived with the same kind of approval form attached. Three lines: supplier name, sample specification, and a signature box. It takes about eight seconds to fill out. That's the whole of it for most buyers.

I've been on the slate side of this business for a long time — checking, testing, and sending samples from our quarry in Jiujiang to roofs on four continents. And I can tell you honestly: the shipments that end in arguments rarely start with bad stone. They start with that eight-second form. The sample was fine. The bulk delivery was fine. The gap between them simply was never written down — so when the gap showed up, nobody could prove what had been agreed.

This guide walks through the sample approval process the way a quality engineer would write it: as a small document, step by step, with one rule on top. The rule is this: you are not approving a tile. You are approving a gap — the allowed difference between the sample and the bulk order. Get that on paper and the sample starts working for you.

Step 1 — Same Source, or It's Not a Sample

The first rule sounds obvious and is broken constantly: the sample has to come from the same lot you're buying. Not from the stockroom. Not from a previous order. Not from the "best piece in the warehouse that happens to be the same colour."

Slate is a natural product. A quarry wall changes as you cut down through it — the first metre can be a shade darker, the next lighter, the layer after that a little grainier. A batch produced in April and a batch produced in August from the same quarry face can look perfectly similar in a phone photo and visibly different on a roof in daylight. So a sample only means something if it carries a batch reference that ties it to your production run.

Ask for the link, in writing: "Which batch does this sample come from, and will my production come from the same batch?" A supplier who takes the question normally will answer with a batch number and can pull a few more pieces from the same run when you ask. A supplier who hesitates is telling you something about their stock control — which is also a sample, of a different kind.

This is why our guide to reading test reports makes the same point about paper: a test report is only meaningful if it carries a batch number that links to your order. The sample and the report work on the same logic. A sample without a batch is a decoration. A report without a batch is a brochure.

Step 2 — Take Three Pieces, Not One

One piece of slate tells you almost nothing about a roof. Slate arrives in batches, and batches have a range — a range of colour, a range of thickness, a range of surface texture. If the supplier sends you a single beautiful tile, what you've learned is that at least one beautiful tile exists in that factory. That's not a prediction about your container.

Three pieces is the smallest honest sample set:

• Piece one — the reference. This one you keep, sealed, until the container lands.
• Piece two — the factory keeps it, sealed the same way, and the production line works from it.
• Piece three — the sacrificial one. You wet it, knock it, measure it, and if the project ever needs a destructive test, it dies without costing anyone anything.

If your sample is a whole series of small tiles (say 500×250), ask for at least three pieces and preferably five. Lay them all out on a table, in daylight, and look at the spread between them — not at the prettiest one. The difference between your lightest piece and your darkest piece is the actual colour range of the batch. That range, not the pretty piece, is what you should compare later to the delivered roof.

Two details that make the three pieces honest:

• Photograph the set together, in the same frame, in daylight. The photo doesn't replace the physical sample — it's your memory when the sample is 9,000 km away.
• Ask for one piece from the middle of the production, not the "best" piece. Some suppliers, old habit, will grab the nicest tile in the stack. That tile is a floor sample, not a production sample. If you want to know what your roof will look like, you want the middle of the range — see our guide to sorting and mixing slate for how much a roof tolerates variation and when it notices.

If the supplier balks at three pieces: "we only send one sample per request" — that's a bigger signal than the sample itself. A serious factory cuts pieces by the metre, not by the piece.

Step 3 — Put Numbers on the Card

Here's where most approvals go wrong, in my experience. The sample card has a colour name and a photo. "Grey, natural cleft." No numbers. And then a 900 m² order arrives in a harbour in Europe, and the colour is "grey" — in the sense that it is grey. But it's a different grey. The argument then takes months, because neither side can point to a number that settles it.

Numbers end arguments. So, before you sign the card, get numbers on it:

Thickness deserves a special word, because it's the number that changes everything else on a roof — the weight, the headlap calculation, the fixings. A 4.8 mm sample measured with a caliper, written on the card, is the single fact that survives a six-month production cycle. Our own factory is CNC-trimmed to tight tolerances, so when we send samples, the card genuinely says what the batch will do — but don't take any supplier's word for it. Measure the sample yourself, with your own caliper, and write your own numbers beside theirs. Two sets of numbers on the card is a beautiful thing; it means both sides measured the same piece.

Step 4 — Seal It, Sign It, Date It

The physical act of sealing the sample is where the sample stops being a favour and starts being a reference. The ritual is deliberately simple:

• Keep one piece on your side — the one you'll put on the table when the container opens.
• The factory keeps its copy — visible in the production hall, signed and dated.
• Both copies carry the batch number, the size, the thickness, and the date — written on the tile itself or on a card fixed to it. A sample that can't be told apart from a random tile is not a sealed sample, it's a pebble.

This is the moment many buyers think "I could just photograph the colour." And that's exactly the thought that costs money later. A phone photo compresses colours, shifts white balance, and flatters everything. The physical piece is the only copy that survives the season, the lighting, and the argument.

One more habit, from years of watching containers open: keep the seal unbroken for as long as you can — meaning, if you need to compare at delivery, compare first, unseal after, and photograph the sealed piece beside the delivered tiles before you open it. The photograph of the two pieces together, before anything is moved, is the one image that can't be disputed. You don't need to unseal it to compare colour; you need it to exist.

Step 5 — Write the Approval Record (a Small Form)

Now the fast form gets replaced with a real one. The approval record doesn't need to be long — a standard form fits on one page — but it has to be complete enough that a third person, opening your container six months later, can tell what was agreed. Ten lines is enough:

That's it. One page. Most of these fields take a minute each, and a supplier who has been through real approvals before will fill them without blinking. If a supplier reacts to the form with "that's not how we do it," you just learned something cheaper than a container load of it.

Step 6 — The Delivery Check: Close the Loop

The sample approval process doesn't end when you sign. It ends when the container is opened, the palette is cut, and the first three tiles come out onto the table.

Take out the sealed sample. Lay a delivered tile beside it — not in the shade of the container, out in daylight, or at least in the same light for both. Do the three-minute check:

• Colour: Does the delivered tile sit inside the range of your sample set, or does it look like a different series? One tile that's a bit off is a tile. A whole pallet that's off is a batch — and that's when your approval record earns its place.
• Measure: Caliper on thickness, ruler on length and width. Compare with the numbers on the card, not with the number in the quote.
• Sound: a clean slate rings. Set the delivered tile on two fingers, knock the centre with a knuckle, and compare the ring to the sample. A dull thud on a delivered tile means an internal crack — and the ring test needs no training at all.

Our arrival inspection checklist covers this whole moment in detail — crate by crate, tile by tile. The point of the approval process is that the inspection at delivery is not a mystery; it is simply the last line of the form. You approved the gap. Now you check the gap.

The Traps in the Sample Box

Experience compresses into a short list. Five ways a sample can be polite and the delivery rude:

• The cherry-picked sample. The nicest tile in the warehouse, photographed in warm light, sent to you as "our standard." The check: always order the sample from the current production run, not from stock, and ask for the run's average, not its star.
• The stock sample. The sample is from last month's batch, and your production is from this month's quarry face. Two faces, two colour worlds. The fix: the batch number on the sample must match the batch number on your order confirmation.
• The lighting lobby. A sample photographed under warm indoor light becomes "warm grey" in the photo and "cold grey" in daylight. The fix: judge colour in daylight, both wet and dry. Wet slate is the colour you'll see after rain, and it is always darker.
• The thickness star. The sample measures 5.4 mm at its thickest point, so the card says "5–8 mm" and the delivered batch sits at 4.3–5.2. The fix: your own caliper on the sample, four corners, written on the card. This is the single most common silent downgrade in slate imports.
• The expired sample. You approved a sample in spring; production happens in autumn. Slate from the same quarry can shift across the seasons. The fix: if more than two months pass between approval and production, re-request the sample. It costs a little freight and a few days — cheap insurance for a container.

Notice that all five traps are fixed by the same six steps above. That's the design. The process isn't bureaucracy — it's the difference between a "sample request" and a sample approval that survives until the container opens.

Where Samples and MOQ Meet

One last connection that catches people: the sample approval isn't the end of the story — it's the beginning of the ordering one. A sealed sample is exactly what turns a trial order or a mixed container into a safe one. When you can't afford a full container, the sealed sample from the trial order becomes the reference that a later full container is measured against. The trial order is your batch; the seal is the memory of it.

And when you finally compare quotations across suppliers, add one line to the comparison: "what does the sample process cost and include?" A supplier who answers with a clear card, a batch reference, and a sealed sample is quoting you more than a product. They are quoting you a process that protects your money. The same supplier's price-per-m² often doesn't look as cheap. It should still win.

Related Reading

• How to Source Roofing Slate from China — where the sample request fits inside the whole sourcing journey, checkpoint by checkpoint
• Questions to Ask a Roofing Slate Supplier — the ten questions that include "will you seal my sample and produce against it?"
• Roofing Slate Inspection Checklist — the delivery-side check that closes the approval loop, crate by crate
• How to Compare Roofing Slate Quotations — add "what does the sample process include?" to your quote comparison
• Roofing Slate MOQ Explained — why the sealed sample is the anchor of a trial order and a mixed container