Your Roof Coating Restoration Failed — but the Coating Wasn't the Problem

The call I get more often than I'd like

In February 2024, a facilities manager called about a roof coating restoration that had failed after eleven months. Blisters the size of dinner plates. Peeling in sheets. He was ready to blame the coating manufacturer — and honestly, so was I.

Then I asked a question that changed the conversation: "What exactly did you apply, and what did the spec sheet say?"

The coating wasn't the problem. The problem was what was under it, and what had been used to prep it. I get calls like this a few times a year. They almost always start the same way: "Your product failed." Rarely is the product the culprit.

The problem was never the coating

Here's what I do: I'm a quality and brand compliance manager at Evonik, a specialty chemicals company. I review product documentation, certificates of analysis, and technical data sheets before they reach customers — roughly 200+ items a year. I've rejected about 12% of first deliveries in 2025 on documentation issues alone. Not because the chemistry was bad. Because the paperwork didn't line up.

Let me be clear about where Evonik sits in the chain. We don't manufacture finished roof coatings. We supply the raw materials — coating resins, adhesive polymers, specialty additives — that go into them. When one of our resins ends up in a coating system, we're accountable for what's written on the document. And too often, that's where things go wrong.

Take the roof coating failure. It came down to two things: a substitute resin and an acid etch that was never qualified for that surface.

The contractor had switched to a "comparable" coating system a few months earlier. The manufacturer of that coating had quietly swapped the base resin for something cheaper. On paper, it looked fine. Same viscosity range. Same solids content. The Evonik logo on the documentation was there, and the product name was close. But the lot number didn't trace back to any of our production sites — including Evonik's Marl Chemical Park in Germany, where we make a large share of our coating resins.

That's the first thing I check. A logo on a certificate of analysis is a starting point, not proof. The lot number has to trace back to a real production batch with raw material records, process parameters, and release testing. If it doesn't, you're not buying what you think you're buying.

The substitute resin also had never been tested for UV stability or moisture permeability. Those properties don't show up in a basic spec sheet. They show up on a roof, eleven months later.

The "close enough" trap

Why does this keep happening? In my experience, it's not incompetence. It's a specific mindset: "It's basically the same."

I watched a purchasing agent order "the standard epoxy" for a defense-market adhesive application. Not the grade with certification — the standard grade. He thought the difference was marketing. It wasn't. Adhesives for defense markets don't just need to bond well. They need documented provenance: raw material traceability, batch records, verified quality tests. Without that, the entire lot is suspect, no matter how well it performs in a pull test.

Defense procurement has its own language. If a contract references a particular adhesive standard, the product has to meet that standard — and the documentation has to prove it. A broker's "equivalent" sometimes even has identical chemistry. But without the paperwork, it's unusable. That order sat in a warehouse until it hit its shelf-life limit. The production line lost six weeks waiting on a certified replacement.

And then there's the acid. Someone asked me recently, "What is the chemical equation for sulfuric acid?" It's a fair question, and the answer is simple: SO₃ + H₂O → H₂SO₄. Anyone who passed high school chemistry can write it. But knowing the equation isn't the same as knowing the grade.

Sulfuric acid comes in different concentrations and purity levels (think etch-grade versus reagent-grade). The chemical formula is identical. The trace impurities are not. In the roof coating case, the crew used a by-product acid for surface preparation because it was cheap and "did the same thing." It etched the concrete. It also left behind trace metals that undermined adhesion. The coating bonded fine on day one. By month eleven, it was peeling because the surface chemistry underneath was wrong.

What skipping verification actually costs

Let me put some numbers on this, because "cheaper" tends to disappear when the invoice arrives.

• The roof restoration: the redo cost $22,000, plus a month of delayed occupancy. The substitute resin had saved the coating manufacturer roughly $1.80 per gallon.
• The defense adhesive order: six weeks of line downtime. Programs like that run on quarterly schedules. Miss a window and you're not just late — you're disqualified.
• The acid etch: 8,000 square feet of coating stripped and redone. The contractor paid for it, and their insurer dropped them afterward.

I've got my own scar tissue here. I once accepted a verbal delivery commitment because "we've worked with this supplier for years." That was the one time the verbal agreement got forgotten. The delay cost us $4,000 in expedited freight on an $18,000 project. The supplier didn't cover it. We did — because we didn't get it in writing.

We didn't have a formal verification process for incoming material certificates back then. The third time I saw the same avoidable failure, I built a checklist. Should have done it after the first.

Build verification into your timeline

Look, I'm not saying budget options are always bad. I'm saying they're riskier. And risk has a price — you pay it either in verification or in failure.

Here's what the fix looks like in practice:

1. Verify the product code and lot number against the manufacturer's system. A phone call or an email, not a visual check of the label.
2. Confirm the certificate of analysis matches the grade you ordered — including impurity limits, not just the main assay.
3. Get delivery commitments in writing. Especially for time-sensitive work.
4. If a supplier claims a product is "eco-friendly" or "low-VOC," ask for the test data. That's not just my opinion. Per FTC Green Guides (ftc.gov), environmental claims need to be substantiated. I've seen products marketed as "recyclable" that were anything but.

Most of these checks take fifteen minutes. The problem isn't time. It's that people skip them when they're in a hurry — and the hurry is exactly when you can't afford to skip them.

Why do rush fees exist? Because unpredictable demand is expensive to accommodate. But in my experience, an expedite fee is a bargain compared to a failed batch. In March 2024, we paid $400 extra for guaranteed delivery on a specialty polymer. The alternative was missing a $15,000 commitment. After getting burned twice by "probably on time" promises, we now budget for certainty when the date matters. I'd rather pay $400 for guaranteed delivery than $4,000 for "probably on time."

Now a caveat: I can only speak to the specialty chemicals side of this. A small painting contractor doing residential work doesn't need a full traceability chain. The calculus is different when the cost of failure is a driveway versus a defense contract. But the principle holds: know what you're actually putting on the surface.

The bottom line

Every avoidable failure I've seen this year traces back to the same moment: someone looked at a spec, decided "close enough," and moved on. The coating was fine. The resin was probably fine too. The system around them wasn't.

Verification isn't bureaucracy. It's the difference between knowing what you have and hoping you have it. And in this industry, hope is not a quality plan.