Stone Guides

Stone is not waterproof: the layer beneath decides

26 October 2026  ·  17 min read  ·  GranTerra Stone, Hui'an

Granite linear drainage grilles strapped on a pallet in the workshop before packing

A granite floor is not waterproof. It is water-resistant, which is a different claim, and the difference is the reason so many beautiful bathrooms are opened up again eighteen months after handover.

The stone you see is the last layer of a stack. Everything that actually keeps water in the right place sits underneath it, and the stone’s only job in that arrangement is to be a hard, clean, drainage-friendly surface that lets water reach the outlet without pooling on the way. When a wet area fails, the failure is rarely the stone. It is the interface between the stone and the layer beneath, and the people who own that interface are usually different trades who never met.

This is a note about that interface. Not about falls in degrees or the depth of a bedding layer, which belong to whoever is doing the waterproofing. About who is responsible for what, where the water is allowed to go, and which decisions have to be made before a single piece is cut.

The stone is the wear surface, not the membrane

Granite absorbs very little water. That is true and it is measurable, and it is also the most misused fact in the trade. “Granite does not absorb water, so the stone is the waterproofing” is a sentence that has cost a great many clients a great deal of money.

Low absorption is not zero absorption. More importantly, water does not need to pass through the face of a stone to get behind it. It goes around. Every joint between two pieces is a capillary. The joint between the stone and the wall is a capillary. The gap at the edge of a drain is a capillary. Water sitting on a surface will find a route down through the joints long before it bothers to soak into the material itself, and once it is below the stone it has nowhere to evaporate to.

There is a second route that gets forgotten entirely: from underneath. Ground moisture, a leaking pipe, a neighbouring wet area with a failed membrane, a planter on the other side of the wall. Stone is a good recipient of water from below because it is dense enough to hold a reservoir quietly for a long time before anything shows on the surface.

So the honest position is this. The stone is what you walk on. The membrane is what stops the water. They meet at a plane, and that plane is where a wet area is won or lost. If your specification treats the stone as the defence, you have removed one line of protection from a system that only ever had two.

What actually sits under a stone floor

Reading upward from the structure, a stone wet floor is roughly this: the structural slab, a screed or mortar bed that carries the fall, the waterproof membrane, a protection layer over the membrane, the bedding or adhesive that holds the stone, and the stone.

Each layer has a characteristic failure, and each failure shows up differently on the surface. It is worth knowing the signatures because they determine who pays.

Structure. Deflection. A floor that moves slightly under load will crack the rigid layers above it in a pattern that follows the structure, not the joints. You see long, straight cracks crossing several pieces.

Fall. Water that stands, or water that leaves too slowly. Puddles that stay for hours, or a drain that gurgles. Often no leak at all, just a floor that never quite dries.

Membrane. Damp patches with a hard edge, appearing first at the lowest point of the ceiling below or at the skirting on the far side of the wall. The pattern drifts and grows.

Protection layer. Damage done during construction, hidden until the building is in use. Punctures from a dropped trowel, a screw through a batten, a chipped edge where a wheelbarrow hit the upstand. These do not show up in a flood test done before the stone went down.

Bedding or adhesive. Hollow spots, then movement, then a loose piece. A hollow-sounding stone in a wet area is not a cosmetic issue, it is a route.

Stone. Wear, etching, staining, and slip. All real, none of them water damage to the room below.

When a client tells us there is a leak, the useful question is not which stone it is. It is which of those six is failing, and the answer is almost never the sixth.

The fall belongs to the layer below

Falls are formed in the screed and the membrane, not in the stone. This sounds obvious and gets contradicted on site constantly, because the person laying the stone is standing there with a level and it is very tempting to take up a small error with the bedding.

Taking up error in the bedding does three things. It varies the thickness of the mortar under the stone, which changes how the stone behaves under load and under heat. It creates a void where the bed is thinnest. And it puts the finished surface slightly off the plane the drainage was designed around, which is exactly the plane the water needs.

The result is a floor that looks flat and drains badly, or drains well in the middle and holds a film of water at the edges, which is where the joints are, which is where the water then goes down.

The correct sequence is the reverse of what site pressure usually produces. Decide the outlet position first. Set the falls from the outlet out to the perimeter. Test them before the membrane goes down. Only then does the stone get measured, because the stone is being cut to a surface that already exists, not used to create it.

This is also the reason a wet-area order needs the drainage layout before it can be cut. A stone floor is set out from the drain. If we are given a room and no outlet position, we are guessing, and the guess will be expensive in a way that only appears once the screed is already poured.

Three junctions that leak first

In a wet area there are three places where the waterproofing stops being a plane and becomes a corner, and those three places account for most of the failures we hear about. All three are geometry problems before they are detailing problems.

The doorway. The threshold is where a wet floor meets a dry one, and it is the only place where the membrane has to change direction and then stop. Water on the wet side pushes against that stop with the whole depth of the screed behind it. A threshold that is simply a stone bar laid across the line, with nothing under it holding the membrane up and over, is a dam that will be overtopped eventually. The detail that works is the one where the membrane turns up and over into the threshold before the stone goes down, so the stone is sitting on a continuous surface rather than capping an open edge.

The wall-floor corner. Nearly always built as a change of plane with a small radius or a fillet, because a sharp internal corner is where a membrane is thinnest after it has been dressed into place. The stone detail then has to respect that: a skirting or a wall panel that lands on the floor should not be fixed in a way that puts a fixing through the upstand.

The outlet. Covered separately below, because it is the one junction where stone and plumbing meet directly and where the stone supplier has real influence.

The common thread is that all three are decided at drawing stage, not on site. Once the screed is down, the only lever left is sealant, and sealant is not a junction. It is a maintenance item with a service life. The drainage side of the same room is covered here, and the two only work when they are read together.

A drain is an interface, not a hole

A linear channel or a point drain in a stone floor is not a hole cut in a piece. It is a connection between two systems that have completely different tolerances, and it is the single place where the stone and the waterproofing have to agree about exactly where the opening sits.

Three things follow, and all three cause trouble when they are left to the end.

The position has to be fixed before the stone is cut. Not “approximately here”. The outlet body has a flange, the flange has a footprint, and the stone has to stop short of that footprint in a line that looks deliberate. If the drain position moves after the stone is made, the stone does not move with it. What happens instead is that the opening is enlarged on site, the flange is partly covered, and the waterproofing connection is improvised around a gap that nobody designed.

The grille is part of the design, not hardware. A stone grille in the same material as the floor reads as part of the floor. A stainless grille reads as a fitting. Both are legitimate, but the choice changes how the stone is cut around the opening and who supplies what. Deciding late usually means the stone has been cut for one and is receiving the other.

The stone at the drain has to be supported, not just laid. The edge of an opening is where a piece loses its continuous bearing. If the piece over the channel is a long span with nothing under its middle, the first heavy item dropped on it will tell you. This is a reason to keep openings near joints rather than in the middle of a long piece.

Our own position on this is narrow and worth stating plainly: we cut the stone opening to the outlet you specify, and we will cut it in the works rather than leaving it to site, because a factory cut is square and a site cut is whatever the saw and the schedule allowed that morning. The grilles we produce are cut in the same material as the floor they sit in, which is the only way the opening reads as part of the surface rather than a hole in it. But we need the outlet, its flange, and its position on a drawing. Without those three, the opening is our guess.

The pool is the same problem with water on the other side

A swimming pool inverts the whole argument. In a bathroom, water sits on top of the stone and tries to get down. In a pool, water stands against the stone with depth behind it, and tries to get in and under.

That inversion changes which failure you get first. A bathroom fails by leaking downward, slowly, into a room below. A pool fails by pieces coming loose, because a piece of stone at the waterline is being pushed at its joints by a head of water every time the pool is filled, and pulled at by every bather who uses it as a step. The visible symptom is a hollow sound, then a rocking piece, then a missing one.

Three consequences follow for the stone itself.

The waterline is the hardest zone in the whole pool, and it is not the same specification as the floor of the pool. It gets wet and dry repeatedly, it gets sun, it gets chlorine or salt, and it gets cleaned. A surface that is perfect underwater can be the wrong surface at the waterline.

The coping is a movement detail before it is a decorative one. It sits across the joint between the pool structure and the surrounding deck, two elements that move differently. The coping has to be able to accommodate that without transferring it into the pool shell, and it has to shed water away from the pool rather than into it.

The stone cannot be the seal. Around a pool, the watertightness is the shell and its coating or membrane. The stone is a lining. If a specification relies on the joints between pool stones being sealed well enough to hold water, it will be re-sealed every season until somebody rebuilds it.

For the surface and edge decisions themselves, the wet edge is covered here, and the coping line we produce is set out on its own page. What matters at this level is the sequence: shell, coating, stone. Any other order of confidence is a leak with a delay on it.

A wall has a different problem

Vertical stone in a wet area is not a floor standing up. It sheds water by gravity rather than collecting it, so the failures move to the edges and the joints.

The wall’s protection is a drained cavity or a continuous coating behind the stone, depending on the system. Where it is a cavity, the water that gets past the face is supposed to run down inside the cavity and out at the base, and the base has to actually be open. A cavity that has been sealed at the bottom, or filled with mortar to make the skirting sit nicely, is a bucket.

Where the wall meets the sky, the coping or capping is the whole game. It has to overhang, it has to fall away from the wall on both sides, and it has to have a drip so that water leaves the stone rather than tracking back underneath. A coping that is set flush with the wall face, however neat it looks from the ground, puts the wall’s head directly in the weather. The wall guides cover the coping as a roof, and the same principle applies at a parapet.

The same logic, smaller, applies to window heads and sills, and to any horizontal ledge. Every horizontal surface on a vertical element is a small roof, and every small roof needs a direction and an edge.

Trapped water does more damage than moving water

This is the counter-intuitive one, and it is why sealing everything is not the answer.

Stone is a vapour-permeable material. It lets moisture move through it slowly in both directions. If a surface is sealed on the front and the back, or if the joint between stone and everything else is filled with a hard, impermeable compound all the way round, the moisture that inevitably gets in has no route out except through the stone itself, which is now sealed.

What follows is a slow accumulation. Salt carried in solution reaches the surface, the water evaporates, the salt stays, and the face blooms white. That is efflorescence, and its mechanism and its cures are set out here; the point for this note is that sealing the face is one of the ways it gets worse rather than better.

The same accumulation shows up as dark patches that do not dry, as a dull haze that will not polish out, and in freeze-thaw climates as surface spalling where a saturated stone is asked to hold ice.

So the detailing principle is not “seal everything”. It is: keep water out where water is coming from, and let whatever gets in leave. Joints that are meant to be movement joints should be flexible and should stay flexible. Joints that are meant to be closed should be closed. The mistake is treating every joint as though it were the same joint, and the different types are set out here.

The related mistake is treating sealant as a structural element. A bead of sealant is a line of material sitting in a gap, and if that gap is going to move, no bead will hold it still. It manages a joint that was already designed. It cannot substitute for one that was not.

What a flood test proves, and what it does not

A flood test is the standard proof that a wet area’s waterproofing is continuous. It is genuinely useful and it is genuinely limited, and knowing the limits avoids the false confidence that follows a pass.

It proves the membrane has no holes and no open laps, at the depth of water used, at the temperature on the day, for the duration of the test.

It does not prove the falls are right, because a pond of water with no leak is exactly what a wrongly-falled floor also produces. It does not test the upstands above the water line, which is where more failures start than anywhere else. It does not test anything that happens after it, which is where the real risk sits: the protection layer being punctured by the next trade, a fixing driven through a membrane, a threshold installed over a membrane that was never turned up.

And it tells you nothing about the stone, because the stone is not in the test.

The practical consequence is that the test should be done before the stone goes down, it should be recorded, and the record should be part of what the client keeps. If a leak appears two years later, the flood test result is what turns an argument into a diagnosis.

What we need before a wet-area order is cut

We cut stone. We do not install waterproofing and we do not design drainage. Because of that, the division of responsibility has to be clear before the order is released, and these are the items that decide whether the stone can be made correctly.

  1. The drainage layout. Outlet type, outlet position, and the flange footprint. A marked-up plan is enough; a dimension from two fixed walls is better.
  2. Where the falls start and finish. We do not need the levels, but we do need to know the high and low points, because they decide where joints can and cannot land.
  3. The threshold detail. Whether the stone runs through the doorway or stops at it, and whether there is a stone upstand. This determines whether the threshold is one piece or two, and which side carries the fall.
  4. The wall connection. Skirting, wall panel landing on the floor, or a shadow gap. Each one changes the last course at the perimeter.
  5. Whether the pool, if there is one, is a lining or a structural finish. This changes both the joint strategy and what the stone has to be able to do at the waterline.
  6. Any drawing that already shows a joint layout. If your design team has set out a grid, we will follow it. It is far cheaper to follow a grid than to correct one.

None of these are about the stone. All of them are about where the stone has to stop.

Where our responsibility stops

It is worth being direct about this, because the most expensive projects are the ones where nobody said it out loud.

We are responsible for the stone: the material, the finish, the cutting, the consistency across the order, the packing, and the accuracy of what leaves the works against the drawing we were given. We can advise on what a surface will do under water, on which joint widths are realistic, on how a pool coping behaves when the deck moves, and on where a stone opening should land relative to a drain. We will flag a detail we think will fail, in writing, before the order is released.

We are not responsible for the waterproofing, and we cannot warrant a wet area. No stone supplier can, and any that offers to is warranting another trade’s work with no control over it. What we can do is make sure the stone is not the reason it fails, which means being precise about where our edges are and asking for the information that puts them there.

If you are at the stage where the drainage is set but the stone is not yet drawn, that is the best possible moment to talk. Send the layout, the outlet, and the threshold detail, and we will come back with what the stone can and cannot do around them — before anything is cut.

Send us the plan.

Written in the yard

Every figure on this page comes from our own order sheets. Where something is a trade norm rather than a number we can hold, we say so — send a spec and we will confirm it against a real block before you commit.

Next step

Send the drawing. We do the rest.

Reply within one working day, in English or Chinese. Send a spec, a drawing or just a photo of the sample you are matching — we will tell you honestly whether we can hold it.

MOQ 1 × 20FT  ·  FOB XIAMEN  ·  SAMPLES IN 7 DAYS

Chat with us on WhatsApp