Elevator Insight

From a Rejected Otis Elevator Cleveland Batch: What I Learned About Ductwork OEM and HVAC Sourcing

2026-09-29 · Martin Kowalczyk

The Tuesday the 'Standard' Word Got a Batch Rejected

September 2024. A Cleveland contractor I'd been working with for over two years sent me a photo before 8 a.m. A ductwork flange sitting on his shop floor, tape measure laid across the corner. It was off by roughly 4 mm. Not catastrophic. Just visibly wrong next to everything else going into the same lobby.

I pulled the order. Placed two months earlier, fabricated by a ductwork OEM partner we'd been using since 2022. Spec sheet said, and I quote, 'conforms to standard duct tolerances.'

That word again. Standard.

For context: my role is quality and brand compliance at Otis. We supply elevator-related MEP components, HVAC assemblies, ductwork, and railing systems to contractors across the Midwest. The project in question sat inside a building with an Otis elevator well at the center—which is why the search term 'otis elevator cleveland' keeps landing in my inbox, and usually, it points to the wrong company. The visitor thinks we're Otis Elevator. We're not. Same name, different business. But that's a different headache. What happened with that flange is the useful part.

What We Meant vs. What They Heard

Our purchase order said 'standard.' The fabricator read SMACNA standard. The client's contract specified something tighter. Both parties were technically correct. The specs just didn't line up.

That's when I noticed the real problem wasn't this one order. It was that our entire approach to HVAC systems OEM vs private label sourcing had been built on an assumption we never actually verified: that 'OEM' meant we controlled the spec.

It doesn't. Not automatically.

Here's the difference, plainly. When you buy ductwork OEM, you're meant to get a factory product built to your specification—your drawings, your tolerances, your brand. Private label usually means taking a shelf product, slapping your name on it, and shipping. OEM gives you more spec control when you actually enforce it. Private label gives you speed.

We thought we'd chosen OEM. What we'd actually chosen was a supplier who printed our logo on their default output. Every PO said 'standard.' No spec number. No revision reference. No locked tolerances.

So we paid for OEM. We received private-label quality. That was the surprise—and not the pleasant kind.

The Glass Railing That Almost Became a Second Problem

Same client. Same lobby. They'd also ordered a section of frameless glass railing for a walkway overlooking the atrium.

While we were still sorting out the ductwork, the railing glass was already moving through sample approval. I'd rejected two lots before signing off on the third. First lot came in at the wrong thickness—8 mm instead of 10 mm, packed and shipped anyway. Second lot had clamp spacing that met 'industry standard' but didn't match the stamped drawings. Two rejections. Not one. I keep having to correct myself on that, because it felt like three.

Same root cause. Same word doing damage. Standard.

At that point I called the whole project to a stop. Every shipment, every supplier, on hold. I put the client's specs and each supplier's interpretation side by side on a shared screen and made everyone look at the gap. It was not small. For the ductwork it was a few millimeters. For the glass railing it was structural: post spacing, embedment depth, wind load assumptions. If that third lot had shipped without me catching it, the installers would have been drilling holes in concrete for hardware that didn't match the panel.

What I Didn't See Coming

I expected to find a bad supplier. That's not what happened. Every supplier on that project was competent. The failure was on our side of the table.

We never version-controlled the specs. We never attached drawings to POs. We let 'standard' do the work that a spec number should have been doing. And we chose the fabricator largely on unit price—$6 cheaper per flange than the next option. On a 600-piece order, that's $3,600 in savings. The redo cost roughly $4,100 after rush shipping, crating, and time lost on the schedule.

That $3,600 didn't just evaporate. It turned into a $500 loss plus two weeks of stress. And it nearly turned into a much bigger problem with the railing glass.

Looking back, I should have required a revision-controlled spec sheet attached to every PO from day one. I didn't, because in 2022 when we onboarded that fabricator, their samples were solid and their price was attractive. My decision then was reasonable given what I knew. The knowledge just hadn't caught up to the risk.

The Week of Second-Guessing

After I froze shipments and told the client we'd re-fabricate the flange batch, I kept second-guessing myself. What if freezing everything was overkill? What if the client walked because we'd delayed their schedule? What if my spec rewrite ended up so strict that no supplier would bid on it?

I didn't relax until the first replacement batch arrived on October 22, 2024, and passed inspection at the job site. That was the signal. Until then, every email notification from the fabricator made me wince.

What the New Protocol Actually Changed

By January 2025 we had it in writing:

  • Every PO carries a spec revision number. The word 'standard' is banned from our purchasing templates.
  • Drawings get reviewed and signed off before any supplier contract is executed.
  • First-article samples are signed by a compliance reviewer, not by production.
  • Any order above 100 pieces gets a spot-check before it leaves the supplier's dock.

Roughly 70% of first-delivery rejections disappeared in the following quarter. Give or take—I don't have the exact figure in front of me, but it's in that neighborhood.

One more thing worth flagging. Per FTC guidance (ftc.gov), claims about origin, compatibility, or OEM status have to be truthful, substantiated, and backed by evidence. That includes the way 'OEM' and 'private label' get thrown around by suppliers. I've seen contractors buy 'OEM' parts that were really relabeled generics. That gray area costs buyers more than any single quote ever will.

Where It Landed

Everything shipped by April 2025. The client re-upped for a second phase. The re-fabricated ductwork took 11 days. The glass railing went in on schedule—the only one of the three lots that didn't get rejected, because this time we actually looked before it left the dock.

Would I tell a contractor to always pick the pricier option? No. That's not the lesson. The lesson is that ductwork OEM vs private label isn't a decision about price. It's a decision about who owns the spec. Pick the one that writes it down. The $6-per-piece savings from the cheaper fabricator was never real. It was borrowed, at interest.

I still think about that flange sitting on the shop floor with a tape measure across it. It cost us $4,100 to learn a lesson we could have learned for free—just by writing down a number instead of a word.

Share LinkedIn Email
Martin Kowalczyk
Martin Kowalczyk

Martin Kowalczyk is a sustainable-materials and construction-procurement analyst covering envelope products, structural materials, interior finishes, mechanical systems, prefabricated assemblies, and site infrastructure. He uses ISO 21930 construction-product rules and ISO 14025 environmental-declaration principles while comparing declared units, life-cycle modules, global-warming potential, recycled content, VOC evidence, reference service life, installed cost, and replacement cycles. His sourcing guides help architects, developers, and contractors compare technically equivalent options without confusing product performance, environmental disclosure, compliance evidence, and whole-life value.