A floor that has let go tells you why, if you know how to read it. The pattern of the failure, where it started, what is on the back of the piece that came up, all of it points at a cause. Taylor looks at these regularly, and the diagnosis matters because it decides whether the repair is a patch, a recoat or a full removal.
So this is a post-mortem. Pick up a loose piece, turn it over, and work through it.
First: look at the back of the chip
This single observation narrows it down faster than anything else.
Concrete stuck to the back. If gray concrete came up with the coating, the bond to the slab was stronger than the slab itself. That is not a coating failure. That is a weak or deteriorated surface layer on the concrete, often a slab that was overworked or finished with too much water in the mix, leaving a soft skin that was never going to hold anything.
Clean back, bare concrete below. The coating never bonded. This is the big category and it is almost always preparation: an unground slab, a sealer that was never removed, dust left on the surface, or a floor that was coated while it was still wet.
White chalky residue on the back. That is efflorescence, salts carried up out of the slab by moisture. The coating bonded to a layer of powder that then let go of the concrete. This is a moisture failure and it is the most expensive one to ignore.
Layer separated from layer. If the clear came off the base but the base is still stuck down, the recoat window was missed. Resins have a time range in which the next coat bonds chemically. Go past it and the previous layer has to be scuffed mechanically first. Skip that and the layers are only touching, not joined.
Then: look at where it failed
The map of the failure is the second half of the diagnosis.
Circles where the tires sit
Hot tire pickup. A thin coating on an unground slab softened under a hot tire and the rubber won. Very common on big-box kits in this climate.
Random patches with no relationship to traffic
Moisture, nearly always. Vapor coming up through the slab collects at the bond line and pushes the coating off wherever the pressure is highest, which has nothing to do with where you walk. Patch failures in a basement are the signature.
Bubbles that appeared during or just after installation
Outgassing. Air in the slab’s pores expands as the concrete warms through the day and pushes up through wet resin, leaving pinholes or domes. Preventable by coating on a falling temperature rather than a rising one, and by priming properly.
Wear lines down the middle, intact at the edges
Not a failure, strictly. That is a floor that was built too thin for its traffic, or one where the sacrificial clear has simply done its job and worn through. Usually recoatable without full removal, which makes it the best news on this list.
Yellowed and chalky all over, outdoors
UV degradation, because a standard epoxy was left as the top layer where an aliphatic topcoat belonged. The floor may still be bonded fine and simply need the right clear over it.
The four causes behind almost all of it
Having read a dozen of these, the list collapses to four.
No real grind. The most common by a wide margin. A coating laid over a hard-trowelled or sealed surface has nothing to key into. Around Anniston, Oxford, Jacksonville and Piedmont this is especially common, because a lot of the housing from the fifties through the eighties has tight, power-trowelled garage slabs that look perfect and reject coatings. Acid etching does not open them.
Untested moisture. A below-grade or ground-level slab was coated on the assumption it was dry. The testing and prep side of the trade exists to turn that assumption into a number before anybody opens a bucket.
Wrong product for the exposure. Epoxy left on top outdoors. A thin residential system on a shop floor taking point loads. A smooth gloss around a pool. Each of those is a product doing a job it was never built for.
Too thin. Low solids content, rolled fast, one coat. What looked thick going on dries to almost nothing, and there is no material there to absorb heat, wear or impact.
Three of those four are preparation and specification. Only one is the product in the can, which is the opposite of how most people shop for a floor.
What remediation actually involves
Here is the part people do not want to hear: you cannot usually patch it.
The temptation is obvious. Four circles failed, the rest looks fine, so grind the circles and recoat them. The problem is that the sound area is sitting on exactly the same unprepared slab as the failed area. It has not failed yet. It is going to, on its own schedule, and the new patches will have a visible edge where they meet the old coating.
So restoring a failed floor means treating the whole slab as one surface again. The old coating comes off by grinding, which is a mechanical process rather than a chemical strip, and the diamonds keep going until there is a uniform open profile across every square foot, including the parts that were still stuck down.
How hard that is varies enormously. A thin, already-failing coating comes off relatively willingly. A sound, thick, well-bonded system that simply needs removing for another reason is slow work, and that unpredictability is why removal is the hardest part of a job to pin down in advance.
Once the floor is bare, the slab gets assessed properly. Spalls and pits are filled, cracks are addressed, moisture is tested if the location warrants it. Then it is coated like any other floor.
The one case where a patch is honest
There is an exception worth naming, because blanket rules make people suspicious. If the coating was properly built on a properly ground slab and the damage is local and mechanical, something heavy was dropped, a welder’s spark pitted it, a jack stand gouged it, then a local repair is a reasonable thing to do. Nothing about the surrounding floor is compromised, the bond is sound everywhere else, and the repair is cosmetic rather than structural.
The way to tell the difference is the back of the chip again. Damage from above has concrete or sound coating under it. Failure from below has a clean back or a chalky one. One is an accident, the other is a system that was never going to hold.
Why the second floor costs more than the first should have
There is no way to say this gently. You pay for the removal, you pay for the repairs that the failure caused or revealed, and then you pay for the floor. The removal is work that would not have existed if the first job had been done right, and it is often the most labor-intensive part of the whole project.
That is the real argument for the grind, the moisture test and the proper topcoat the first time. They are not upsells. They are the difference between buying one floor and buying two.
Mr. Shanks installed metallic epoxy floor coating in our home and did an outstanding job. He was honest, knowledgeable and communicated very well.
Cody S.
If you are standing on one of these right now
Take a photo of the failed area, pick up a loose piece and photograph the back of it too. Those two pictures tell Taylor most of what he needs before he arrives, and they are worth more than any description over the phone.
Taylor holds an Alabama license, number 2602341, and is insured. He is on every job himself, and he will tell you honestly whether a floor needs full removal or whether you got lucky and it only needs a recoat.
Photograph the failed patch and the back of a loose piece, then call. Taylor can usually tell you what happened before he gets there.