Usually not, and this is worth settling before anybody argues about materials. A rigid crack filler does one job well: it fills a gap and it goes hard. If the concrete either side of that gap stays where it is, the repair holds indefinitely and you will never see it again. If the concrete moves, the filler is being asked to do something no rigid material can do, which is to stretch. So the crack reappearing is not evidence that the wrong product was used, or that it was mixed badly, or that the coating on top was too thin. It is evidence that the slab moved after the repair went in. That is a completely different problem with a completely different answer. The uncomfortable version of this: a coating is a few thousandths of an inch thick. Concrete slabs weigh tons and move with temperature, moisture and load. Any conversation that starts from the coating holding the slab together has started in the wrong place.

How do I tell an active crack from a dormant one?

By watching it over time rather than by looking at it once, and this is the test that gets skipped. The cheap version is a crack monitor or a rigid gauge fixed across the crack, read over several weeks and ideally across a temperature swing. A dormant crack does not change. An active one opens and closes, usually widest when the building is coldest. The rough version, when there is no time for that, is to look at what the crack is doing rather than how wide it is. A crack that runs the full width of a bay in a straight line, or that lines up with a column, a re-entrant corner or a doorway, is usually structural or a shrinkage crack that found its own line, and those often keep moving. A short, wandering, hairline crack in the middle of a slab that has been down for years is usually finished. What you must not do is decide by looking once and assuming. A crack that is dormant in August may open in January, and by then the floor is finished and the argument is about who pays.

What is the difference between a crack and a control joint?

A control joint is a crack somebody planned, and treating one like the other is a common and expensive mistake. Control joints are cut into a slab specifically so that shrinkage cracking happens in a straight line where nobody minds, rather than randomly across the floor. The slab is meant to crack there. It is doing what it was designed to do. That means a control joint will always move to some degree, so filling one with a rigid material and coating over it is asking for the crack to come back through your new floor along a dead straight line. It looks worse than a random crack because it is obviously deliberate. The standard answer is to honor the joint: fill it with a semi rigid joint filler to support the edges against traffic, and either leave the joint visible in the finished floor or re-cut it through the coating afterward so the movement has somewhere to go. Which of those you choose is a decision to make with the client before coating, because one of them changes how the floor looks.

Why does routing the crack matter so much?

Because a hairline crack has almost no room in it, and a repair with no material in it has no strength. Running filler across the top of an unrouted crack wets the surface and fills the first millimeter or two. There is nothing gripping the sides deeper down, so the repair is a skin. It is also why an unrouted repair is so cheap in material terms, and that is the honest reason it gets skipped: opening the crack out to a proper kerf multiplies the filler volume enormously. On a 30 ft crack, a hairline barely 1 mm wide and 6 mm deep takes about 63 mL of filler once you allow for waste. Rout that same crack out to 6 mm wide and 12 mm deep and it takes 757 mL, which is roughly twelve times as much. Nobody skips the routing because they do not know it matters. They skip it because of that number. The kerf dimensions themselves come from the filler manufacturer, not from us, and they vary with the product. What does not vary is the principle: a repair needs a cross section to develop strength in, and you cannot get one without cutting it.

Can a dormant crack be made invisible?

Often yes, and this is the good outcome worth aiming for. Rout the crack out to a kerf, vacuum it properly so the filler bonds to concrete rather than to dust, fill it slightly proud with a rigid polyurea or epoxy crack filler, and grind it flush once it has cured. On a dormant crack in a floor that is going to be coated, that repair can genuinely disappear. Two things make it disappear rather than merely improve it. Broadcasting sand into the wet filler gives the coating something to key into across the repair, so the patch does not read as a different texture. And on a solid color floor, a repair is often still faintly visible through a thin build, so a decorative broadcast or a heavier build hides what a thin coat will not. The honest caveat is that all of this rests on the crack being dormant. Every step above is wasted on a moving crack, and doing it beautifully does not change that.

What are my options on a crack that is still moving?

Three, and the first thing to accept is that making it invisible is not one of them. Honor it as a joint. Saw the crack out to a clean, straight kerf, fill it with a flexible sealant rather than a rigid filler, and let it stay a visible line in the finished floor. This is the most durable answer and it converts an ugly random crack into something that looks deliberate. Re-cut through the coating. Fill and coat as normal, then saw a joint back through the new floor directly over the crack. The movement then happens in the saw cut instead of tearing the coating. This gives you a cleaner looking floor than the first option and it needs the crack located accurately before it disappears under the coating, so mark it. Accept and manage it. On some floors, particularly older industrial slabs with a lot of movement, the realistic answer is to fill, coat, and tell the client plainly that the line will come back and will be maintained rather than cured. That is a legitimate choice when the alternative is a slab replacement, but it has to be a decision the client makes in advance rather than a surprise. What unites all three is that the client is told before the coating goes down. A moving crack that reappears after a conversation is a maintenance item. The same crack reappearing after no conversation is a defect claim.

Could this be moisture or settlement rather than movement?

It could, and the distinction changes who needs to be involved. Most telegraphed cracks are shrinkage and thermal movement, which is a flooring problem with flooring answers. Concrete shrinks as it dries and moves with temperature, and cracks open and close accordingly. Settlement is different. If the crack has a step in it, so one side sits higher than the other, or if it is opening progressively rather than cycling with the seasons, the ground under the slab is moving. That is a structural question and it belongs with a structural engineer, not with a floor coating. Coating over a settling slab is money spent on something that will fail again. Moisture is the third possibility and it is worth ruling out because it often travels with cracking. A slab passing vapor can push a coating off around a crack as well as through it, so if the coating is also blistering or debonding near the line, test the slab rather than assuming the crack explains everything. The practical rule: cycling with the weather is normal movement, a vertical step or steady widening is not, and either of those with blistering alongside means test for moisture too.

Crack Telegraphing: Why the Crack Came Back

A crack that reappears through a new floor is almost never a coating failure. The concrete underneath is still moving, and no coating a few thousandths of an inch thick is going to hold two slabs together. What decides whether the repair holds is a question that gets asked far too late: is this crack still moving?

Quick answer

The filler almost certainly did not fail. The slab moved. A coating is a few thousandths of an inch thick and concrete slabs weigh tons, so nothing you put on top is going to hold two pieces of concrete together. Everything depends on one question that usually gets asked too late: is this crack still moving? A dormant crack can be repaired until it is invisible. A moving one has to be honored as a joint, not hidden.

How to tell it is telegraphing and not a defect in the film

Look at where the line runs rather than at the line itself. A telegraphed crack follows the concrete, so it carries on straight across bays, runs to a wall, a column or a doorway, and reappears in exactly the same place after every repair. A fault in the coating stops at the edge of whatever was coated.

The other tell is what the coating is doing either side of it. In a telegraphed crack the film is sound and firmly stuck down right up to the line. If it has lifted, blistered or peeled back from the crack, that is a bond failure and you want peeling instead. And if the cracking is a fine crazed pattern all over rather than one line, the slab is not involved at all: that is wrinkling and alligatoring, the film moving against itself as it cured.

The one question that decides everything

Dormant

Width does not change across a temperature swing
Often short, wandering, in a slab that has been down for years
Can be repaired until it is genuinely invisible
Rout, vacuum, fill proud with a rigid filler, grind flush

Active

Opens and closes, usually widest when the building is coldest
Often runs the full width of a bay, or lines up with a column or corner
Cannot be hidden permanently by anything
Has to be honored as a joint, or re-cut through the finished floor

The test is a crack monitor or a rigid gauge fixed across the crack and read over several weeks, ideally across a temperature swing. That is inconvenient and it is the step that gets skipped, which is why so many of these arguments happen after the floor is down. A crack that is dormant in August may open in January, and by then the conversation is about who pays rather than what to do.

A control joint is a crack somebody planned

Control joints are cut into a slab on purpose so that shrinkage cracking happens in a straight line where nobody minds. The slab is meant to crack there, which means the joint will always move to some degree.

So filling a control joint with a rigid material and coating straight over it produces a crack back through your new floor along a dead straight line, which looks worse than a random crack precisely because it is so obviously deliberate. Support the joint edges with a semi rigid joint filler, then either leave the joint visible or re-cut it through the coating afterward. Both are legitimate. Only one of them is a surprise to the client, so decide it with them beforehand.

Why routing matters, and the number behind why it gets skipped

A hairline crack has almost no room in it. Running filler across the top wets the surface and fills the first millimeter or two, and a repair with no cross section has nothing to develop strength in. It is a skin.

Opening the crack out to a proper kerf fixes that, and it is worth being blunt about the cost, because it is the honest reason the step gets skipped. Take a 30 ft crack. As a hairline barely 1 mm wide and 6 mm deep it takes about 63 mL of filler once waste is allowed for. Routed out to 6 mm wide and 12 mm deep, the same crack takes 757 mL, which is roughly twelve times as much material. Work your own out with the crack repair calculator, and put the routed kerf in rather than the hairline you can see, because the routed cross section is what you actually have to fill.

The kerf dimensions themselves come from your filler manufacturer and vary by product. The principle does not: you cannot get a cross section without cutting one. For the estimating side of this, including where to measure width and whether to round up to the next cartridge, see estimating crack filler volume.

Repairing a dormant crack so it disappears

Rout it out to the kerf your filler manufacturer specifies
Vacuum it properly, so the filler bonds to concrete rather than to dust
Fill slightly proud with a rigid polyurea or epoxy crack filler
Broadcast sand into the wet filler, so the coating keys across the repair and the patch does not read as a different texture
Grind flush once it has cured
Build enough coating over it. On a solid color a repair is often faintly visible through a thin film, so a decorative broadcast or a heavier build hides what a thin coat will not

What to do with a crack that is still moving

Honor it as a joint. Saw it out to a clean straight kerf and fill with a flexible sealant rather than a rigid filler. The line stays visible, and an ugly random crack becomes something that looks deliberate. The most durable answer
Re-cut through the coating. Fill and coat as normal, then saw a joint back through the new floor directly over the crack, so the movement happens in the saw cut instead of tearing the film. Cleaner looking, and it needs the crack marked accurately before it disappears under the coating
Accept and manage it. On an older industrial slab with a lot of movement, fill, coat, and tell the client the line will come back and be maintained rather than cured. A legitimate choice when the alternative is replacing the slab, and only when it is the client decision rather than your surprise

All three depend on the same thing: the client hearing it before the coating goes down. A moving crack that reappears after that conversation is a maintenance item. The same crack reappearing after no conversation is a defect claim.

When it is not a coating problem at all

Most telegraphed cracks are shrinkage and thermal movement, which is a flooring problem with flooring answers. Two things put it outside that.

A step in the crack, or steady widening rather than seasonal cycling, means the ground under the slab is moving. That is a structural question for a structural engineer, and coating over a settling slab is money spent on something that will fail again.

Blistering or debonding alongside the crack means test the slab for moisture with the concrete moisture calculator rather than assuming the crack explains everything. A slab passing vapor can push a coating off around a crack as well as through it.