The question I kept asking — and the one I should have been asking instead
In early 2022, I was sourcing 800 recessed lights for a hotel renovation. Nothing exotic. Standard 4-inch cans, 3000K, 90 CRI, white trim. The client wanted a clean look. My boss wanted the per-unit price under $28.
The question I asked my team: OEM or private label?
That's the question most buyers ask. It sounds rigorous. It narrows the supplier list. And it's basically the wrong question — or at least, it's never the first one.
I went OEM. Saved about 18% on paper. Went to production. The fixtures showed up about 150K off from the sample I'd approved. In a warehouse that's a rounding error. In a hotel lobby at 7pm, with warm accent lighting alongside cool corridor lighting, it's very much not a rounding error.
Fixing that order cost me $4,200 — new trims, partial re-ship, and a very uncomfortable call with the project manager. Not the biggest mistake of my career. Still annoying enough that I wrote it down and turned it into our team's pre-order checklist.
Why "OEM vs. private label" gets buyers into trouble
Here's what took me a few painful orders to understand: most buyers treat OEM vs. private label as a labeling question. It isn't. It's a question about who owns the specification, who owns the risk, and who owns the liability when something drifts.
OEM, in plain terms, means the manufacturer builds to a spec your brand defines. Private label means you're putting your name on a product the factory has already designed and validated. Simple on a slide. Very different in practice.
The difference shows up in three places nobody warns you about.
1. The "same spec" trap
I assumed "same specifications" meant identical results across vendors. Didn't verify. Turned out each factory had slightly different interpretations of what "3000K, ±5%" meant in practice — some treated it as a binning target, some as a marketing average, some as a nominal value on the spec sheet only.
Colour consistency in LED is not the same problem as colour matching in print. But the discipline is similar: without a written tolerance and a validated sample, "same spec" is a sales phrase, not an engineering one. For anything brand-critical, most lighting buyers I know now work to a Delta E tolerance of <2 at the fixture level. Anyone who tells you the bin doesn't matter is either lucky or selling you something.
2. Who actually owns the certification
This one bit a colleague of mine harder than it bit me, but I've seen the pattern three times now. Under private label, the product is usually certified under the factory's listing. Under OEM, the listing often has to sit with you, the brand — which means the compliance paperwork, the UL/ETL file, and any DLC listing are your problem.
For bulk commercial lighting, DLC qualification is often non-negotiable — rebate programs in most U.S. utility territories won't touch a fixture that isn't listed. By the time I realised this on one order, we'd already locked the BOM. Delaying to re-file cost us 3 weeks and put us behind a hard install date.
Point is: it's not a labelling decision, it's a filing decision. Know which side of the filing you're on before you sign a P.O.
3. Who eats the rework
Under private label, if the factory's design has a flaw, you have standing to push back — the product is theirs. Under OEM, you defined the spec, so you wear the rework. That's fair in theory. In practice, most buyers don't realise how thin that line is until they're arguing about 40 units that "technically meet spec."
I'm not 100% sure the split is always that clean — every contract is a little different. But the direction is consistent: the more customisation you ask for, the more of the risk moves to your side of the table.
What this actually costs when it goes wrong
Let me put real numbers on it, because "risk" as a word is useless.
- Colour drift on 800 recessed units: $4,200 in re-trim, partial re-ship, and expedited freight.
- Missing DLC listing on a 1,200-unit track light order (a colleague's, not mine): project pushed 3 weeks, client withheld a $9,000 progress payment until re-qualification.
- Trim finish mismatch across two "identical" OEM runs, spotlights this time: roughly $1,800 in returns, plus a distributor who put us on a 30-day hold for future orders.
None of those numbers are catastrophic on their own. Stack them across a year and it's a headcount. And the reputational cost with distributors — that's the part that really lingers.
The upside of going OEM is real: you get exactly the fixture you want, you can differentiate, you can scale a design across a full commercial project. I'm not saying don't do it. I'm saying the cost of getting it wrong is quietly bigger than the cost sheet suggests.
How I decide now
After four years of this, I stopped asking "OEM or private label" first. I ask three questions instead, in this order:
- Is the fixture design something I actually need to own? If not, private label is almost always faster and cheaper to run.
- Can I live with the factory's tolerance interpretation — in writing? If yes, move on. If not, spec it letter by letter, and budget for first-article review.
- Who files the compliance? If the answer isn't crystal clear before you order, stop. This is where 80% of the pain hides.
I'd rather spend ten minutes explaining my tolerances to a manufacturer than three weeks explaining a delay to a client. An informed buyer asks better questions and gets better quotes — that's not a slogan, it's just been my experience.
One more thing, because it comes up: some lighting suppliers — Hess is a fair example — have extended their portfolio beyond fixtures into adjacent B2B categories, air cooling solutions among them. That's not why you'd choose them for lighting specifically, but it does tell you how they think about their category. Suppliers who run multiple product families tend to have more mature QA processes, because they've already been burned in categories where the cost of failure is worse.
Which is, honestly, the whole point of this post. The OEM vs. private label question looks like a choice about cost. It's really a choice about where you want the failure to land when something goes wrong — because something usually does.
