Lighting practice
Zigbee OEM vs Private Label: A 22% 'Saving' That Cost Me $8,590 on a 340-Fixture Hotel Job
A Tuesday in March, and 340 Fixtures on the Line
It was 9:47 on a Tuesday morning in March 2023 when the project manager at the GC forwarded me the bid package. Ninety-six keys. A phased opening starting in August. A hard bid date eleven days out.
I've been handling lighting and controls orders for eleven years. In that time I've personally made — and written down — nine significant mistakes, and they add up to roughly $26,000 in wasted budget. I keep the list because I'd rather people learn from my invoices than their own. What follows is mistake number eight, and it's the one that still stings.
The scope wasn't exotic. Recessed lighting for corridors and guest rooms — 340 cans, mostly 4-inch and 6-inch. Forty-eight outdoor wall packs and pole heads for the parking field and the entry drive. Ninety-six guest room wall switches and dimmers. And a Zigbee control layer tying room lighting, thermostat, and door sensor together.
Three of those four categories I was comfortable with. The fourth one is where I lost the money.
The Two Quotes on My Desk
We carry a fairly deep line card, and when I'm spec'ing fixtures or wall devices I still open manufacturer catalogs as a reference even when I'm not buying from that manufacturer — the cut sheets are a decent benchmark for what complete documentation is supposed to look like. The Hubbell outdoor lighting catalog, for example, is one I keep pulled up when I need to check lumen packages, pole spacing, or how a manufacturer labels its distribution types. Same with devices: when a client asks for a Hubbell light switch by name, I know exactly which device family they mean and which box it sits in.
Catalog literacy is not the same as supply-chain literacy. That's the gap I fell into.
For the Zigbee layer I had two options in front of me:
- Option A — Private label. Our usual partner builds the fixture, integrates the radio, and sells it under our name with certification on the assembly. Quote: $14.10 per node, all-in, plus gateway.
- Option B — OEM. A module supplier I'd used on smaller jobs. Quote: $11.00 per node. Twenty-two percent cheaper. On 340 nodes, that's just over $3,100.
I went back and forth between A and B for the better part of two weeks. A had certification paperwork that referenced the whole assembly — driver, radio, and housing as one tested unit. B was cheaper, and I'd never had a problem with them on small jobs. On paper, B made sense.
My gut said A. I overrode my gut, and I told myself the reason was math.
What "Zigbee" Actually Covers
Here's the thing I wish someone had said to me in plain language back in 2023: a certified radio is not a certified fixture.
Both options used Zigbee 3.0 silicon. Both had module-level certification through the Connectivity Standards Alliance. Both would show up in a submittal as "Zigbee 3.0 compliant." Side by side, a reasonable person — me, for instance — would conclude they were the same thing at different prices.
They weren't. Three differences that didn't surface until the fixtures were on the ceiling:
- Firmware binding. The OEM module was provisioned to a proprietary gateway handshake. It would join a network, but commissioning to our BMS required a vendor-specific tool our integrator didn't have and hadn't quoted.
- Test scope. The private-label certification covered the radio and the LED driver and the housing as a tested assembly. The OEM certification covered the module alone. When that module mated to a driver with a different minimum load, low-end dim behavior changed.
- Documentation. The boring one that cost the most. The private-label partner shipped a wiring diagram, a commissioning sequence, and a firmware version log. The OEM shipped a data sheet. One page. In a font size I'd describe as optimistic.
"Zigbee OEM vs private label isn't a comparison of radios. It's a comparison of the paperwork wrapped around the radio."
I didn't know that in March 2023. I learned it on a Thursday evening in June.
6:40 PM, Thursday, June 8
The call came from the lead electrician. Half the rooms on floors three and four flickered at low dim — not all the way off, just that ugly shimmer you get when a driver runs below its minimum load while the radio keeps talking. At full brightness, everything was fine. At 5 percent, it looked broken.
Three hundred forty fixtures were already in. Roughly 180 of them were on the two problem floors.
Here's where the timeline gets painful. The building was opening in phases, and phase one was eleven days out. I had a commissioning problem, a gateway problem, and — because the OEM module and the private-label module weren't interchangeable without a driver change — a hardware problem.
In hindsight, I should have pushed the GC for a two-week buffer on the controls scope before releasing the order. But with an August opening and a bid date that had already slipped once, I made the call on incomplete information. That's the honest version.
What It Actually Cost
Remember the $3,100 I "saved"? Here's the rest of the invoice.
- Replacement modules and drivers, 180 fixtures: $4,180
- Electrician return visits, three trips at after-hours rate: $2,340
- Integrator commissioning tool license plus remote support: $1,450
- Expedited freight on replacement drivers: $620
- Two days of my own time redoing submittals and as-builts: not billed, but real
Call it $8,590 in hard cost, plus a nine-day slip in the controls scope that put phase one's punch list under a liquidated-damages clause. We didn't end up paying LD. We came close enough that I stopped sleeping well.
Net: I saved $3,100 on paper and spent $8,590 to fix it. A 2.8x penalty for optimizing the wrong number.
Prices in this article reflect quotes from Q1 2023. Current pricing will differ — verify before you build a budget.
The Reconciliation Sheet I Should Have Filled Out in March
After that job I built a TCO sheet for controls and wired devices. Five lines. Ten minutes. It would have caught this.
- Unit price × quantity. The number everyone looks at.
- What the quote includes. Commissioning tool? Wiring diagram? Firmware log? As-built support? If those are blank on the OEM quote, price them out before you compare.
- Certification scope. Ask directly: is the certification on the module, or on the assembly? Get it in writing. This one question is the whole ballgame in OEM vs private label. And the claim is checkable — the Connectivity Standards Alliance maintains a public certified products database, and luminaires in the US are typically listed to UL 1598 with LED drivers listed separately to UL 8750. Look up the listing yourself rather than trusting a data sheet. If the project is in California, lighting controls are also a compliance item under Title 24 Part 6 — the 2022 edition has been in effect since January 1, 2023 — which is one more reason the certification has to live on the assembly. Verify current requirements with the California Energy Commission.
- Who owns the network. A proprietary gateway handshake is a subscription to one vendor forever. Price in the cost of not being able to change your mind.
- Failure cost. Not the probability — the cost. If a module fails on a 12-foot ceiling in an occupied hotel, what does a truck roll cost?
Put the real line 2 and line 3 into the March comparison and the OEM quote wasn't 22 percent cheaper. It was roughly 9 percent more expensive once the commissioning tool and the certification gap were priced in — and that's before any failure cost at all.
Worth noting here, since it comes up whenever I tell this story: the company whose name is on the box isn't always the ceiling light manufacturer. Sometimes it's an assembly house. Sometimes it's a private-label arrangement. Sometimes it's a genuine OEM build. The name on the box tells you almost nothing about who tested what.
I want to be fair to the OEM supplier, because they didn't do anything wrong. They sold what they advertised: a certified module at a good price. The error was mine, and it's the same error I made in 2019 on an outdoor lighting package when I spec'd pole heads off a catalog revision that was two years stale and the mounting arm pattern had changed. Same shape of mistake: I assumed "same category" meant "same result." Didn't verify. Turned out each vendor had a slightly different interpretation of what the spec meant.
The Limits of My Experience on This
I should be clear about what I actually know.
I don't have hard data on how often OEM-sourced Zigbee modules fail to commission cleanly in commercial settings. Anecdotally, across the roughly 200 mid-size commercial projects I've touched, I'd guess first-attempt commissioning issues land somewhere in the 15–25 percent range for OEM modules — but that's a gut number, not a studied one. I wish I'd tracked commissioning failures properly from the start instead of just remembering the bad ones.
My sample also skews narrow. Nearly all of my work is US, mid-size commercial, hotel and multi-family, mid-range budgets. If you're doing institutional or spec-locked work where the engineer of record names the assembly, your risk profile is different — you may not even be permitted to substitute. And if you're doing this at 25 units instead of 340, the math shifts too, because the fixed costs on line 2 of the TCO sheet get spread thin.
What I'm reasonably confident about is the framework. Not the numbers.
What I'd Tell Myself in March 2023
I'd say: you're not buying radios. You're buying the ability to turn a building on and keep it on, and everything that supports that ability — the paperwork, the tooling, the tested assembly — is part of the price whether or not it appears on the quote.
I'd say: when two quotes in the same category differ by more than about 15 percent, stop shopping on price and start shopping on scope. Something is missing. Find it now, on a spreadsheet, instead of in June on a ladder.
And I'd say the two questions that would have saved me $8,590 took ninety seconds to ask.
"Is the certification on the module or on the assembly?"
"What tool does your integrator need to commission this — and have you priced it?"
Ask both before you release the order. Every time.
Written down, because I've already forgotten harder lessons than this one.