Why Your TPO Roofing System Might Be Leaking (And It’s Probably Not the Membrane)

A call I wish I hadn’t taken

Two weeks ago a contractor I’ve worked with for years called me. “We installed a Johns Manville TPO roofing system on a 40,000 sq. ft. building last fall. Now, after one winter, we’re getting reports of three separate leaks near the perimeter seams. The owner is blaming the material. We’re looking at a $22,000 redo plus lost reputation.”

He was ready to switch brands. I asked him to walk me through the installation process. Turned out they’d been in a rush – the whole roof had to be done in 10 days instead of 14. They cleaned the membrane with Sprayway glass cleaner because it was on the truck from another job. The seams weren’t rolled with consistent pressure. And they skipped the detailed white top spec review because “it’s just a color.”

That call cost him 45 minutes of troubleshooting. The fix – proper seam rework and a surface wash with JM-approved cleaner – cost about $4,000. The real damage was the trust.

The surface problem: blaming the membrane

Most roofing contractors I talk to default to the same diagnosis: “the TPO membrane must be defective.” It’s a natural reaction. When your work fails, you want to blame the material. And honestly, the industry has seen some bad TPO batches over the years. But in my experience—and I review roughly 500 roofing specs per year at Johns Manville—more than 70% of premature leaks trace back to installation details, not the membrane itself.

The problem looks like pinholes, fishmouths, or adhesion failures at the seams. That’s the visible symptom. But the root cause is almost always in the process that happened before the seam was closed.

Deep cause #1: surface contamination you can’t see

TPO membranes are notoriously sensitive to oils, dust, and cleaning agents. I’ve watched crews spray Sprayway glass cleaner on the lap area because it evaporates fast. It works for glass. But on TPO, the ammonia and surfactants leave a residue that prevents the heat-weld from bonding fully. The seam looks fine from the outside, but a year later the weld peels under thermal cycling.

We tested this in our lab in Q1 2024: 50 sample seams cleaned with Sprayway vs. 50 cleaned with JM’s recommended cleaner. After 500 heat-cool cycles, 12% of the Sprayway samples failed peel strength. Zero failures on the JM side. That 12% is what drives callbacks.

(Note to self: I really should write that up as a field advisory.)

Deep cause #2: rhythm and pressure when you roll a joint

How to roll a joint on TPO isn’t taught in a classroom. It’s a skill that requires consistent speed and downward force. Too fast and the weld is under-cooked. Too slow and you burn the membrane. And if you pause mid-seam – which happens when a crew member stops to grab a drink or answer a text – you get a cold zone.

In a rush job, the temptation is to let the robot welder run at high speed. Or worse, use a hand roller and not maintain the pressure spec (typically 15-20 psi). I’ve seen crews “roll a joint” so quickly the seam barely warped. That joint will fail inside 18 months. Exactly what happened in that call I took.

When I compared two identical side-by-side installations – one with disciplined rolling (using a calibrated speed and pressure gauge) and one with “we’ll be fine” speed – the rushed roof had 4× more repair calls in the first year. The total cost of those repairs: $18,000 on a $120,000 job. That $200 savings in installation time became a 15% profit hit.

The hidden cost of picking the cheapest approach

In my opinion, the industry’s obsession with “lowest installed cost” is the single biggest driver of TPO failures. I hear it every week: “We can beat that quote by $10 per square.” But when I dig into the numbers, that $10/sq. saving usually comes from shortcuts: skipping the primer, using off-brand cleaner, or reducing seam overlap. It’s a classic value over price trap.

Let’s do the math. A 40,000 sq. ft. roof at $10/sq. cheaper upfront saves $400,000. But if 5% of the seams fail in the first 3 years, and each leak costs $5,000 to fix (average service call, material, labor), that’s $200,000 in repairs. Plus the owner’s lost business time and potential interior water damage. The “cheaper” roof ends up costing more than the spec-compliant one – and you lose the relationship.

And don’t forget the warranty implications. Johns Manville’s standard warranty requires that all installation follow published guidelines. If a crew used improper cleaning (hello, Sprayway) or didn’t roll the joint within spec, that warranty is void. I’ve had to reject warranty claims on roofs that looked fine cosmetically but failed due to installation deviations. That’s a hard conversation.

The fix: three things that matter more than the membrane brand

1. Clean with the right stuff. Use only approved TPO cleaners – not glass cleaner, not acetone, not a rag dipped in dish soap. I keep a note in every spec: “No Sprayway, no substitutes.”
2. Roll your joint with discipline. Mark the speed control on the welder. Use a pressure gauge on the hand roller. Train every crew member to recognize the correct melt pattern. And never let a crew “roll a joint” without a signed checklist.
3. Don’t underspec the white top. White top TPO (JM’s white top is a high-reflectivity surface) is not just about aesthetics. It reduces roof surface temperature by up to 30°F, which directly lengthens membrane life. Cutting corners to a dark-colored membrane just to save $0.30/sq. is false economy.

And while we’re at it: if you’re installing a Johns Manville duct board package on the same project (rooftop AHU insulation), keep the same discipline. Duct board edges also require proper cleaning and taping. I’ve seen the same crew use Sprayway on duct board – which ruins the vapor retarder – and then wonder why their system sweats.

In the end, my advice to that contractor was painfully simple: fix the seams now, document everything, and get your crew re-certified on JM’s field application standards. It cost him $4,000 and a week of bad press. Worth every penny. He’s now a loyal customer – and he finally understands why I always push for the value-over-price approach.

It took me 5 years and about 150 roof audits to understand that the membrane is rarely the problem. The problem is the process. And that’s where quality inspection pays for itself.