March 2024, Line 14
I was three pages into a submittal package for a 118-room hotel when I hit line 14 and stopped.
I'm the quality and brand compliance manager at a commercial tile and flooring contractor. I review every submittal before it goes to a GC or a client — around 200 a year, plus every batch that comes off the truck into our yard. Last year I rejected 41% of first submissions. Most of them for the same three reasons.
Line 14 read: “MAPEI TILE MORTAR WITH POLYMER — per manufacturer's recommendation.”
That's not a specification. That's a brand and a vibe.
What I got wrong for the first two years
When I started this job in 2021, I assumed a brand name in a spec meant somebody had already done the work. Mapei on the line item? Fine. Somebody who knows more than me picked that. Move on.
Four years and roughly 800 reviewed submittals later, I know the brand name is the least verifiable thing on the page. The brand tells you who made it. It tells you nothing about which product, which classification, or whether it's the right one for a 24 x 48 in. porcelain panel over a slab with a flatness problem.
The worst submittals I've seen didn't come from contractors cutting corners. They came from contractors who specified a good brand and left the actual requirement blank.
Three lines, three different problems
Line 14: the mortar that wasn't specified
“With polymer” is a marketing phrase. The performance requirement lives in a classification.
ANSI A118.1 is unmodified dry-set mortar. ANSI A118.4 is latex-Portland cement mortar — the modified category, carrying shear bond requirements after water immersion and heat aging that A118.1 doesn't have. ANSI A118.15 is the improved modified category, added in 2011, with tighter requirements on open time and bond strength.
If the line doesn't name one, an installer can legally buy the cheapest bag that says “polymer-modified” on the front. That's the whole problem, right there.
Line 22: porcelain, large format, no class
24 x 48 in. porcelain, 3/8 in. thick. ISO 13006 puts fully vitrified porcelain in Group BIa — water absorption of 0.5% or less. That tile doesn't drink moisture out of the mortar, which changes how the mortar has to set and grab. It's also 50-plus pounds a piece.
The submittal listed “porcelain tile adhesive.” Which class?
ISO 13007 breaks cementitious adhesives into C1 and C2, then adds letters: T for thixotropic, E for extended open time, F for fast setting, and S1 or S2 for deformability. A C2TE S1 means something specific. “Porcelain tile adhesive” means nothing.
And the substrate flatness number — TCNA and ANSI A108.02 both call for 1/8 in. in 10 ft. once tile edges reach 15 in. and up — wasn't in the package at all. That's the number that decides whether you're running a 1/2 in. notch trowel or grinding a slab.
Line 31: “pro seal sealant,” allowance
That one went back the same day.
ASTM C920 is the standard for elastomeric joint sealants, and it requires a sealant to be classified by type (S for single-component or M for multi-component), grade (P for pourable, NS for non-sag), and class, which is movement capability in percent. “Pro seal sealant” gives me none of it. It doesn't even tell me whether the material is silicone, urethane, or acrylic — and on a pool deck, at a wet area, on a joint that moves, that's the only question that matters.
Standards referenced in this submittal review: ANSI A118.1 / A118.4 / A118.15 (mortar classifications), ISO 13007-1 (adhesive classification), ISO 13006 (tile classification), ASTM C920 (elastomeric joint sealants), ANSI A108.02 and the TCNA Handbook for Ceramic, Glass, and Stone Tile Installation (installation and substrate flatness). Manufacturer product data, TDS, and classification statements: mapei.com.
The pushback
I sent the package back with three comments. The PM called me within the hour.
“We've used this exact package on six jobs. Nothing's failed.”
That's the worst sentence in construction. Nothing has failed yet. Six jobs is not a sample size, and a job that hasn't been through a full heating season and a full pool season hasn't been tested.
The distributor rep tried a different angle. He told me his mortar had “the same SBR base the tire guys buy from Firestone Polymers LLC.”
Fine. It probably does. It's also irrelevant. The raw polymer is an input. A118.4 shear bond after water immersion is a result. Where the ingredient came from is not a performance claim about the cured product, and I can't put a raw-material name on a submittal and call it verified.
I held two of the three comments. The sealant classification and the mortar classification went back as hard requirements. I let the flatness line go with a note, because the slab had already been poured and I wasn't going to win that one in March.
Then I did what I always do after a fight like that, which is second-guess myself for two weeks. What if the sub just waited until the next package and swapped the bag? What if I'd cost the job schedule over a classification that a mid-tier product would have met anyway?
The first pallet arrived eleven days later. Bags matched the revised submittal, TDS attached, classification printed on the bag. I stopped worrying.
The part that actually failed
Here's what I got wrong.
The tile was fine. Every square foot of it. The mortar held, the grout held, the flatness problem got ground out and resolved.
The pool deck failed. And it wasn't on my submittal.
The expansion joints between the coping sections had been filled with an epoxy-based joint material. While the deck was finishing up, the painter sprayed a water-based acrylic over the trim and equipment-room walls. Overspray drifted across the joints. Two weeks later the paint started coming off in sheets along the joint lines.
The GC's super asked me the question I've now been asked four times in 18 months: can you paint epoxy resin?
Yes. But not like that — and the reasons are the same three reasons every coating failure happens.
- Amine blush. Amine-cured epoxies react with CO2 and moisture in the air and form a waxy surface film. The joint looked clean. It felt slightly slick. Paint won't bond to it.
- The recoat window. Every epoxy has one — some are 24 hours, some are seven days. Inside the window you get a chemical bond. Past it, you have to mechanically abrade. Nobody did either.
- UV. Exterior epoxy chalks and yellows. Even a perfectly prepped epoxy on a pool deck needs a UV-stable topcoat over it, not just whatever paint is on the truck.
The fix: grind out the joints, strip the failed coating, re-seal, and repaint the adjacent trim over a properly abraded surface. Just under $12,000 in direct cost and 11 days on the punch list.
The failure wasn't in the 14 pages of spec I read line by line. It was in a three-word line item on a different scope that nobody read at all.
What I changed
Our submittal template now has three fields that must be filled or the package is auto-rejected before it reaches me:
- Adhesive classification (ANSI A118.x and/or ISO 13007 class, e.g., C2TE S1)
- Sealant classification (ASTM C920 type / grade / class)
- Substrate flatness and prep method, with the measurement that proves it
First-pass rejection rate went from 41% to 19% over the last year. Not because contractors got better. Because the requirement stopped being a conversation and started being a checkbox.
The fundamentals didn't move. The classifications did.
Five years ago, “polymer-modified” was a real differentiator. Today it's table stakes — practically everything sold as a thinset has some polymer in it. The presence of polymer stopped being the spec. The classification is the spec.
Same story with format. A 24 x 48 in. porcelain panel was a specialty item in 2019. Now it's a default, and 48 x 96 is on drawings. A mortar that passed on 12 x 12 tile in 2015 is a different conversation on a panel that weighs 60 pounds and won't bend.
And ISO 13007 now shows up on the same submittal as ANSI A118, because a lot of the product is imported and the two systems don't map cleanly. If you're still reviewing submittals in one language, you're missing half the page.
But here's what hasn't changed a bit: substrate flatness, cure time, movement joints, and a clean surface. Every failure I've traced in four years goes back to one of those four. Not once has it been the logo on the bag.
The one question I ask now
When someone asks me whether a product is good enough, I don't ask what brand it is.
What class?
Because “Mapei tile mortar with polymer” was never a specification. It was a shortcut for one. And on a job where the shortcut costs $12,000, it's worth the extra email.