Ask ten people in the LED business what you should check before you buy, and you'll get ten different answers. Nobody's exactly wrong. They're just answering for a different kind of buyer.

I run facilities and purchasing for a 140-person company. I took over in 2021, and since then I've handled roughly $310,000 a year across about a dozen vendors — including a full lighting retrofit at our main office and two warehouse spaces. I report to operations and finance, which means I hear about it from both directions when something goes sideways.

Here's what I wish someone had told me in year one: there's no universal LED checklist. There are at least three kinds of buyers, and each one needs to check something different.

Same products on the shelf. Completely different risk. Let's walk through them.

Scenario 1: You're lighting your own space

This is me. Forty-two fixtures in the main office, sixty-odd light bars and strips across the warehouses, plus a batch of smart bulbs in the conference rooms that I've mostly made peace with.

If this is your seat, the metric that matters is total installed cost, not price per fixture. Everyone says that. What they don't say is how much of it is labor, and how much of it is the second purchase you make in year three.

In 2021 I focused on price per panel and completely missed driver replaceability. Our first batch had integrated drivers. When two of them failed at month fourteen, the whole fixture came down. If the driver had been a replaceable component with a standard form factor, I'd have swapped a $12 part. Instead I paid an electrician to deal with it, twice, on a lift.

What I check now:

The counterintuitive part, at least for a single-site buyer: standardize aggressively, even if it costs more per unit. I used to chase the best price on each fixture type. We ended up with seven SKUs and no spares for any of them. Last year we cut it to three and bought spares for each. Cost per fixture went up around 6%. My maintenance calls went down by more than half. I wouldn't reverse that trade.

Scenario 2: You resell or install for other people

Your problem isn't light output. It's the callback.

If you're a light bar distributor, or you're installing for clients, the fixture that fails at month nine costs you the replacement plus the truck roll plus the relationship. Whatever you saved on unit price disappears right there.

What shifts when you're in this seat: listings you can actually look up — UL or ETL with a file number a customer or inspector can verify. ETL is Intertek's listing mark, issued under the same NRTL system, and generally accepted the same way by local inspectors. Either is fine. "CE" alone is not a North American listing. Then binning consistency across reorders, which is the one that catches distributors out. Order 200 units in March and 200 in September and the second batch looks slightly different. Ask what bin tolerance the factory holds, and whether it's contractual or aspirational.

Lead time variance matters more than lead time. A supplier who quotes four weeks and delivers in four weeks is worth more than one who quotes two and delivers in six. And packaging: carton labeling, barcodes on the individual unit, a master carton that matches what's printed on it. Sounds trivial until it's 300 units that have to be relabeled in your own warehouse on a Friday.

For light bars specifically, the things that go wrong are voltage drop on long runs, cut increments that don't line up with connector spacing, adhesive that peels off a textured ceiling in a warm room, and connectors that mate on one order and not the next. On 12V versus 24V — 24V handles longer runs better, but I'd verify that against your actual run length rather than treating it as a rule. IP rating matters too: IP20 for dry indoor, IP65 for anything semi-exposed or wash-down. Don't let anyone sell you indoor tape for a loading dock.

The counterintuitive bit here, and I didn't believe it at first: the lowest-priced distributor almost always ends up costing you more. Not necessarily because of the product — because of the returns process, the relabeling, and the fact that you're the one absorbing the customer's frustration. I'd rather pay more for a supplier who picks up on the first call. Actually, closer to 12% more — I ran the numbers on our worst return month and that's roughly where it landed.

Scenario 3: You put your own name on it

Now you're the brand of record, and that changes what you're buying. You're no longer buying fixtures. You're buying documentation, consistency, and change control — and a fixture is what happens to arrive along with it.

Here's the blindspot in this space. The obvious thing to look at is the brand name on the spec sheet. When it says "Bridgelux LED" or "Bridgelux COB," the thing I nearly missed is that it's telling you about one component, not the product. Bridgelux makes the emitter — the LED package itself. That's a component inside the fixture, not a finished luminaire. A COB (chip-on-board) assembly is a different construction from an array of individual SMD emitters, and done properly it tends to give you a single, tighter light source and better color consistency across a production run. But "Bridgelux inside" on a marketing sheet isn't the same as the right part number inside the tested configuration. Ask for the exact series and part number, and confirm it shows up in the LM-80 report and in the LM-79 data for that specific SKU.

Should mention something else: even a correct emitter can be undone by the driver. Emitter and driver are a matched pair, and mismatches show up as flicker, as poor behavior at the bottom of the dimming range, and in how long the system actually holds up.

The things that bite private label buyers are quieter than you'd expect. Silent component substitution is the big one — the factory swaps a driver vendor or an emitter bin and nobody tells you. Your listing still reads the same, your next container doesn't perform the same. Put a written change-notification clause in the PO. Then: who actually holds the listing. If your name is on the box, you need to know whose UL or ETL file covers the product and whether you're permitted to reference it. That's the gap I run into most in private label paperwork. You also want a locked BOM with traceable lot numbers — boring, and worth more than a fancier emitter — and compliance documentation per SKU, not per product family. "Our whole line is RoHS compliant" is not a document.

If you're selling smart bulbs, add these

Smart bulbs are their own compliance category, and this is where I see the most hand-waving from suppliers.

I should add that none of this is legal advice. If you're importing at volume, your customs broker or a compliance consultant will want the actual documents, not a summary. I've had one shipment held up three weeks because a supplier's compliance folder and their physical labeling didn't match. That one wasn't cheap.

How to figure out which one you are

Four questions. Answer them in order — the first "yes" you hit is your scenario, and if you hit more than one, work from the stricter checklist.

  1. Who holds the warranty obligation if it fails in year three? If it's you, and the product is in your own building, you're Scenario 1.
  2. Does anyone downstream ask you for a spec sheet, a test report, or a certificate? An inspector, a general contractor, a customer's QA person. If yes, Scenario 2 — your documentation has to survive someone else's scrutiny, not just yours.
  3. Does your company name or logo appear on the product, the box, or the listing? If yes, Scenario 3. No exceptions.
  4. At reorder, do you know whether you're getting the same bin and the same driver? If the honest answer is "I don't know," you're in Scenario 2 or 3 territory whether you think you are or not.

One thing that isn't on the list but matters: how concentrated your volume is. If 80% of your spend is a single panel, you can afford to be picky about that one and generic about everything else. If your spend is spread across fifteen SKUs, your real problem is SKU count, not supplier selection. I learned that one the slow way, over about eighteen months.

There's something oddly satisfying about finally getting this systematized. After three years of chasing invoices and arguing with finance about purchase orders that didn't match, having one page per SKU — the LM-80 report, the listing file number, a locked BOM, a photo of the actual label — is the payoff. Twenty minutes per SKU to set up. Saves me hours every quarter.

So, which one are you? Not which one you'd like to be. If you buy lighting for your own building, you can tolerate a slightly higher unit price and should optimize for spares and standardization. If someone else's name is on the invoice, optimize for documentation you can hand over without flinching. If your name is on the product, optimize for change control above everything else.

Get that wrong and you'll spend the savings twice.