Two quotes for the same 6-inch recessed downlight. Left column: $28.40. Right column: $19.15. Four hundred units, so the gap is $3,700 and the sales guy is already spending it.
The fixture on the left is a Lithonia. The one on the right has your name laser-etched on the trim ring. Both cut sheets say 4000K, 90 CRI, 1,000 lumens, 0–10V dimming. On paper they're twins.
They're not twins. And the difference rarely shows up on the invoice or in the first container. It shows up around month 14.
I review private-label submittals before they leave our building — 40 odd fixtures a month, call it 500 a year. In 2023 I rejected roughly a third of the first deliveries from new OEM vendors. Almost none of those rejections were about the fixture. They were about the fields around the fixture that nobody filled in. That's the part that doesn't get quoted.
The Surface Problem: The Gap Looks Like Free Margin
The comparison always happens the same way. Six columns on a quote sheet: price, lead time, CCT, CRI, lumens, warranty. The $9 gap sits in column one and everything else looks close enough to wave through.
I get why. When you're a recessed lighting distributor or you're quoting track lighting wholesale into a fifty-head job, margin is thin and the unit price is the only lever that moves fast. Private label looks like a no-brainer. Same box, same connector, your name on it, 30% off.
Here's the thing that took me about two years to internalize: the quote sheet isn't comparing fixtures. It's comparing the six attributes someone bothered to type. The attributes that decide whether those fixtures are actually the same product — the ones that determine who pays when something fails — are the ones that are blank.
What the Spec Sheet Doesn't Tell You
"4000K" Is a Range, Not a Number
ANSI C78.377 defines 4000K as a nominal point with a chromaticity quadrangle around it, not a single coordinate. Two fixtures can both print "4000K" on the box and land on opposite edges of that quadrangle. North American production tends to run a little above the blackbody locus; a lot of Asia-sourced product sits right on it.
The control is step size. Five-step MacAdam (SDCM 5) is the common commercial floor. Three-step is what you want when fixtures sit side by side. Put a 3-step and a 5-step in the same recessed row in a lobby and you will see the seam. Not everyone sees it. Your client will.
What most people don't realize is that the binning spec is often one of the most expensive lines in the BOM, and it's the first one that quietly moves when a vendor needs to hit a price point. It moves without a phone call.
The Driver Is the First Thing to Change
The driver is roughly a third of the bill of materials. It's also the least photogenic part of the fixture, which makes it the easiest place to substitute.
What you get in the submittal is an LM-79 report. Reasonable. But an LM-79 is a snapshot: one unit, one date, one lab. It tells you what that specific sample measured. It doesn't bind the next production run, and it says nothing about the driver's internals.
Lumen maintenance is where this gets expensive. L70 and L90 are projections, and the projection comes from IES TM-21 applied to LM-80 data — which is tested on the LED package, not the finished luminaire. Test temperature matters. Drive current matters. The extrapolation assumptions matter. A vendor can hand you an LM-80 report that's technically valid and completely unrepresentative of what happens inside their own housing, because their thermal design isn't the one the data came from.
Two questions cut through most of this. What's the LM-79 lab and date? And what's the driver part number, with a no-substitution clause in the PO? If the answer to the second is "we source from several qualified suppliers," that's a red flag. Qualified by whom, and against what?
The Listing Doesn't Travel With the Parts
DLC qualification is model-specific. UL 1598 and UL 8750 files are specific to a construction. FCC Part 15 is specific to the driver.
Swap the driver, and the DLC listing that got you the utility rebate may no longer cover what's in the box. This is not hypothetical. In 2024 we had a vendor change the driver supplier between runs — same part number on the packing list, different board inside. The fixture was fine. The listing wasn't. We found out when a utility audited a completed project and clawed back the rebate on 220 fixtures.
For anything touching emergency egress it gets worse. UL 924 is a separate listing with its own test cycle. NFPA 101 tells you where the fixture has to go and how long it has to run. Neither one cares what your cut sheet says.
OEM and Private Label Are Not the Same Transaction
Everything I'd read said private label is just the same fixture with a different label. In practice they diverge at exactly three points, and the whole price gap lives in those three points: who owns the listing, who owns the warranty, and who eats the field failure.
When you buy a Lithonia downlight, or a Lithonia Lighting flood light, or a BGS LED area light for a site package, you're buying an entity that will still exist and still answer a phone in year four. When you private label, you are — functionally — the manufacturer. Maybe not legally, but the spec-holder is going to call you, and the only entity that can actually indemnify you is a vendor who has already moved that SKU to a new part number.
What the Blank Fields Actually Cost
Four ways this shows up. They tend to arrive in this order.
- Submittal rejection. The spec-holder rejects on a technicality — no IES photometric file, an LM-79 that's four years old, a CCT that doesn't match the specified step. You redo the submittal, you lose two weeks, and sometimes you lose the job.
- Warranty claims you can't pass through. Failures cluster around months 14 to 20, right after the vendor's attention has moved on. You replace at your cost.
- Rebate clawback. The utility audits, the listing doesn't cover the serial numbers, and the money comes back out of your project.
- The redo.
In 2022 we took a batch of 1,400 four-inch downlights where the CCT was visibly off — measured around 3700K against a 4000K spec, on a job where we'd already installed 900 of them. Normal tolerance for that spec is SDCM 3. The vendor's position was that it was "within industry standard." It was, technically, if you averaged across the batch using a 5-step standard. We rejected the balance and they redid it at their cost. Our cost was the 900 already in the ceiling. That was a $22,000 problem and it did not show up in the unit price.
Here's the ballpark math. The $9.25 gap on a 400-unit order is $3,700. Two field failures out of that order, plus the labor to reach them, plus the truck roll, is over $2,000 (this was the going rate through 2024 — lift rental alone can run $400 a day in some markets). So the decision isn't really about price. It's about which risk you'd rather carry.
The Six Lines That Belong in Every PO
This is the part that makes everything above simple. You don't need a 40-page spec. You need six lines that don't have blanks in them.
- CCT and step size — "4000K, SDCM 3," not "4000K."
- LM-79 report with lab name, test date, and the exact model number it covers.
- Driver manufacturer and part number, with a no-substitution clause.
- DLC listing ID and the covered model number.
- L70 or L90 value and the TM-21 basis — test temperature, drive current.
- Warranty terms with a named entity and a labor allowance.
That's it. Six lines. Every one of them is answerable in an email, and any vendor who can't answer them inside 48 hours has told you something useful.
When Private Label Is Actually the Right Call
I'm not going to argue that branded is always correct, because it isn't. If you're specifying track lighting wholesale for a tenant improvement where nobody will ever file a rebate form, or you're sourcing a custom housing no catalog carries, OEM is the only path. Same for a non-standard distribution pattern on a flood light where the catalog optics just don't reach.
Private label works when three things are true. You move enough volume on a stable SKU to matter — ballpark 5,000 units a year or more. Someone in your building owns QC as a job, not as a side task. And you can absorb a field failure without it becoming a cash flow event.
Here's the part people don't like hearing: if you're under about 2,000 units per SKU, or you don't have a QC function, or the spec was written around a named basis of design, private label is usually the expensive option. Not on the invoice. On the P&L. I've watched three distributors learn that the hard way. Combined cost of the lesson: about $60,000.
Then again, branded isn't a guarantee either. A catalog BGS LED area light or flood light covers most standard site and perimeter packages, but it won't cover a custom distribution or an unusual CCT, and no amount of catalog breadth fixes that. When we hit that wall we go OEM — with the six lines above in the PO and a first-article inspection on the first shipment. That inspection has caught something on maybe 40% of first runs. Not always something fatal. But always something.
The fixture is the cheap part. The blank fields are the expensive part. Fill those in first, and the $9.25 decision makes itself.
