A Lighting Spec Sheet Is a Hypothesis, Not a Promise

The batch I rejected two weeks before handover

We were two weeks out from a hotel handover when I rejected sixty plaster sconces.

The paint sheen was off. Roughly two Delta E from our master sample, which sounds like nothing until you put six of them on the same wall under 2700K downlights and one reads grey while the other five read warm. Two units out of sixty. That's all it takes.

The vendor's reply is the part I still remember: "That's within industry standard."

Maybe. But "industry standard" isn't a number I can hold anyone to, and that's the entire problem. I review somewhere around 300 submittals a year, and in 2024 I rejected roughly a fifth of first deliveries — not because suppliers are dishonest, but because a document was read as a promise when it was really a guess.

A spec sheet is not a promise. It's a hypothesis — and in commercial lighting, the distance between those two words is where projects quietly lose money.

Argument one: our vocabulary is broken, and we pretend it isn't

I get asked some version of this roughly once a month: what is an enclosed light fixture?

Ask a designer and you'll hear "it has glass or a diffuser so you can't touch the lamp." Ask a specifier and you'll hear something about damp or wet location ratings, dust ingress, and whether the diffuser is part of the listed assembly. Ask a facilities manager and you'll hear "the one I can hose down."

Three answers. One phrase. A change order sitting between them.

The one that actually matters is the listing. Under UL 1598 — the Standard for Luminaires — a fixture marked for a damp or wet location has been tested and marked as such. An open-frame fixture with a decorative glass shade is not automatically "enclosed" in the listing sense, even though it looks enclosed to anyone standing in the room. That distinction is the reason a corridor fixture that passed a visual review can fail an inspection, and it almost never appears on the drawing.

It's tempting to think the term is descriptive. It's regulatory. Those are not the same job.

Argument two: a collection name is not a specification

This is where the Visual Comfort conversation usually starts, and where it usually goes sideways.

If your schedule says "Visual Comfort Oslo chandelier," you have named a collection. You have not specified a finish, a crystal or alabaster grade, a canopy diameter, a lamping option, or a lead time. I've sat in a bid review with four quotes for "the Oslo" where the only thing they had in common was the silhouette in the rendering.

Same story with the Talia medium chandelier. Two vendors can both be technically correct and still deliver two different objects — different drop length, different diffusion, different warmth at the edges. Wall lights are worse, because now geometry has consequences: anything that projects more than 4 inches into a circulation path starts a conversation with the 2010 ADA Standards, and that number is not on most lighting schedules I review.

Two years ago I ran a test I should have run earlier. Same room, same camera, same exposure. A designer fixture on the left, a well-executed copy on the right. Nine people on our team, no labels. Seven picked the original. Four of those seven said it was because of how the metal caught light along the edge — a finishing difference, not a design difference.

Does that mean the copy was a mistake? No. For a pop-up, a shoot, a three-year lease build-out, I've recommended the cheaper version and I'd do it again. The failure isn't choosing the copy. The failure is choosing the copy and telling the client it's the same thing. That's a document problem pretending to be a taste problem.

Argument three: category words aren't SKUs

"Fish chandelier" is not a product. It's a shape family — organic silhouette, layered scales or glass leaflets, usually a statement piece over a bar or a stair.

Send that phrase to five suppliers and you'll get five different objects at five different price points, and none of them are wrong. That's not a market failure. That's a purchasing document failure.

Tube LED is the more expensive version of the same mistake, because it hides in the schedule as a commodity line item. A four-foot tube LED is interchangeable right up until you start asking about CRI, R9, flicker percentage, color consistency across a 40-unit run, and whether the specific model is on the DLC qualified products list. DLC technical requirements set floors for those characteristics — but "meets DLC" and "meets your project" are two different sentences, and only one of them survives commissioning.

People assume expensive fixtures create better rooms. The causation runs the other way: rooms that demand precision are where expensive fixtures get specified, and then judged harder. The price is a consequence of the scrutiny, not a substitute for it.

The pushback I expect

Three objections, and my honest answers.

• "This is just brand protection dressed up as quality control." No. In a Q1 2024 audit we returned a full batch from a name-brand manufacturer because the canopy diameter ran a quarter inch over spec. The brand didn't save them. The measurement did. I've rejected more expensive fixtures than cheap ones, because the expensive ones get specified in rooms where somebody is actually checking.
• "So you're saying cheap fixtures are bad." No. I'm saying unverified fixtures are risky. Those are different sentences and I'll defend the distinction all day. Budget-driven selections are legitimate on the overwhelming majority of projects — I make them constantly.
• "This only matters on big jobs." It matters more on small ones. A large project has a clerk of works, a commissioning agent, and a punch list. A 40-room boutique build-out has one person doing three jobs and a contractor who wants to be done.

What I check now, in this order

Three things. The order matters more than the list.

First, the listing, not the look. UL or ETL mark, location rating, and whether the diffuser is part of the listed assembly. If a fixture is going into a bathroom or a covered terrace and the submittal doesn't state the rating in writing, I don't approve it — I ask again.

Second, the sample, not the sheet. Physical sample under the actual project color temperature. Delta E is a number you can argue with; "slightly warmer in person" is an argument you'll lose at handover.

Third, the name plus three modifiers. I never approve "the Oslo." I approve the Oslo in a stated finish, a stated size, and a stated lamping option. If a vendor can't echo those three things back in writing, they haven't read the spec — they've read the picture.

None of this is glamorous. It's the least interesting work in the project (ugh, genuinely). But it's also the reason I sleep before handover instead of refreshing a tracking number at 1am.

I still think the spec sheet is the most useful document in commercial lighting. I just stopped treating it as a contract. It's a hypothesis. Verification is the only thing that turns it into a promise — and the fixtures that pass are rarely the ones that looked best on paper.