Failure analysis
Six reasons a parkade membrane fails years before it should.
A ten-year system that gives you four is almost never a product problem. Here is what actually goes wrong, in the order we find it.
2026-08-02
The short version.
A traffic membrane rated for ten to fifteen years that fails in four did not wear out. Something underneath it was wrong before the first coat went down — and it is nearly always one of six things.
01
The substrate was never prepared to profile.
Every manufacturer publishes a required surface profile — a measurable roughness the concrete must have before the primer goes on. A swept, pressure-washed slab does not meet it. Achieving it means diamond grinding or shot blasting, and it is the single most expensive line item that can be quietly shaved out of a bid without anyone noticing until the membrane peels in sheets.
You can usually tell after the fact: a membrane that failed on a properly prepared slab tears and leaves material bonded to the concrete. One that failed on an unprepared slab lifts clean, with the underside looking almost new. That clean release is the signature of a bond that never formed.
02
Nobody tested the slab for moisture.
Concrete holds water, and water moves toward the drier side. Put a relatively impermeable coating on a slab that is still driving vapour upward and the vapour collects at the bond line until it lifts the coating off in blisters. This is common on slabs-on-grade with no vapour barrier, on younger concrete, and on any deck with water arriving from below or behind.
The test is an in-situ relative humidity probe placed in the slab per ASTM F2170. It costs very little relative to the project, and it is the most common reason we tell a client the scope is bigger than they hoped — either a moisture-mitigating primer or a system that tolerates the condition.
A slab that failed a moisture test does not mean you cannot coat it. It means you cannot coat it with what you were planning to use. Those are very different conversations, and the second one only happens if somebody tested.
03
The wrong category of product went on the deck.
A suspended parkade slab deflects under vehicle load and it cracks. Traffic membranes exist because they are elastomeric — they stretch over that movement instead of splitting along it. A rigid epoxy, however tough, has effectively no crack-bridging, so every crack in the slab telegraphs through within a year or two.
We still see epoxy specified on suspended decks, usually because somebody priced a hard-wearing industrial floor and assumed harder meant better. On a warehouse slab-on-grade, that instinct is right. On a deck that moves, it guarantees failure.
04
The details were treated as trim rather than as the job.
Leaks almost never start in the open field of a deck. They start at drains that the membrane was not properly terminated into, at expansion joints that have given up, at membrane turn-ups at curbs, columns and walls that were run too short, and at moving cracks that were coated over rather than detailed with reinforcing fabric.
Detailing is slow, fiddly work that consumes a disproportionate share of labour hours and shows up nowhere in a square-foot price. That is exactly why it is the first thing compressed when a bid gets sharpened, and exactly why it is where failures cluster.
05
The drainage never worked, and the membrane paid for it.
A deck with inadequate fall to its drains holds standing water. A membrane in permanent contact with ponded water works far harder than one that sheds and dries between rain events, and in the Lower Mainland that is most of the year. Add freeze-thaw in Whistler or Squamish and ponded water becomes a mechanical attack as well as a chemical one.
Slope correction before the membrane goes down costs real money and it is frequently declined. That is a legitimate decision on a tight budget — provided everyone understands they have shortened the service life of the system they just paid for, rather than believing they saved money.
06
It was installed in conditions the system does not permit.
Every one of these systems has published limits: minimum and maximum substrate and ambient temperature, maximum humidity, a dry substrate, and recoat windows that must be respected in both directions. Coat too soon and the layer beneath has not developed; coat too late and you get intercoat adhesion failure that will not show up for a year.
Schedule pressure is what breaks these rules. A crew that is one day from finishing, with rain forecast and a strata expecting stalls back on Monday, faces real pressure to keep going. The honest answer is to stop — and the reason we document conditions per phase is so that decision is a recorded fact rather than an argument later.
Questions
Related questions.
How long should a parkade traffic membrane actually last?
On a suspended slab, ten to fifteen years for a correctly specified and correctly installed system, and often longer on a deck that is protected from direct sun. The variables that move that number most are the quality of the surface preparation, whether the drainage genuinely works, and whether the wear coat gets refreshed once during the system's life.
Can a failing membrane be patched, or does the whole deck need redoing?
It depends entirely on why it is failing. Localised failure at a detail — a drain, a joint, a turn-up — is often repairable in place, and that is a much smaller project. Widespread adhesion failure from poor preparation or a moisture problem is not repairable, because the same condition exists everywhere. Cores and adhesion testing tell you which situation you are in; a walkthrough does not.
Is a more expensive membrane a longer-lasting membrane?
Not reliably. The premium in a fast-set system buys you speed, not service life — you pay for the ability to hand a ramp back the same day. Money spent on preparation, detailing and drainage correction buys service life. If the budget is fixed, spending it on the substrate rather than on a higher-tier product is usually the better decision.
Our membrane is under warranty. Does that cover this?
Read what the warranty actually covers. Manufacturer warranties typically cover the material, and are conditional on the system having been installed to published requirements — documented conditions, coverage rates and recoat windows. A workmanship warranty from the contractor is a separate thing. Where installation records do not exist, warranty claims tend to go nowhere, which is the practical argument for insisting on documentation while the work is happening.
Related
Keep reading.
Coating systems compared
Tremco, Sika, MAPEI and BASF systems side by side — and how we choose between them.
Learn more →For strata councils
Your depreciation report flagged the parkade membrane. What that scope really involves.
Learn more →Parkade waterproofing
How we install traffic membranes on occupied buildings, phased so the parkade keeps working.
Learn more →Next step
Not sure which of these applies to your deck?
Send photos and whatever documentation you have. We will tell you what we think is going on before anyone talks about price.