Commercial Lighting Procurement: 4 Scenarios, 4 Different Answers (and Where the Replica Question Actually Lands)

There's No Single "Best" Lighting Purchase

I've been asked to "just pick good lighting" enough times to know the request is missing information. The right answer depends almost entirely on what kind of purchase you're making, and in my experience those fall into four fairly distinct buckets:

• Spec-driven projects — new construction or a renovation, with drawings, a specifier, and a GC.
• Replacement and maintenance — something on the ceiling died and you need the same thing again.
• Design-led refresh — a lobby, boardroom or client-facing space where the fixture is part of the look.
• Multi-site rollout — the same order, repeated across locations and across years.

Worth saying up front: most people I've worked with assume they're in bucket three. Most of the time they're in bucket two.

Quick context on who's writing. I'm the office administrator for a roughly 400-person company. Facilities and lighting ordering land on me — call it $180K a year across eight to eleven vendors — and I report to both operations and finance. I'm not a lighting designer. I'm the person who has to make the paperwork hold up and the install not go sideways. So take the aesthetic advice elsewhere. This is the procurement view.

Scenario 1: Spec-Driven Projects

If there's a specifier or an architect in the picture, your job as the buyer is mostly to not get in the way and to make sure the paper trail is complete before the electrician shows up. This is where a track lighting specification guide earns its keep.

What I ask for on every submittal package, every time:

• Photometric file (IES / LM-63 format), so the designer can actually run the layout instead of guessing
• An LM-79 test report for the specific fixture being shipped — not the platform, not the family
• LM-80 data plus a TM-21 projection if anyone is promising a lumen-maintenance number
• Listing certificate — UL 1598 for the luminaire, UL 8750 for the driver, or the ETL equivalent
• Dimming protocol and compatibility list. 0-10V, DALI, or line voltage. "Dimmable" on its own is not an answer.
• CCT, CRI, and TM-30 values if color consistency matters in the space

On track lighting OEM programs specifically, here's the thing people miss: OEM doesn't remove work, it transfers work. The box carries your name, so the submittal, the listing documentation, and the warranty channel are yours to administer. That's a perfectly good arrangement if the manufacturer hands you the technical file and holds the listing. It's a problem if all they'll hand you is a price list.

My slightly contrarian take: with OEM track, the cheapest program is rarely the problem. Inexpensive fixtures with a clean technical file are a reasonable buy. Inexpensive fixtures with no file are the problem — you can't submit them, you can't get them through a rebate review, and you can't prove anything to an inspector.

We tried to save around $9 a head on a track order in 2022 by going with a supplier who wouldn't publish a dimming compatibility list. They said it would "work with standard dimmers." It worked with roughly two-thirds of them. We paid an electrician to swap drivers and dimmers on 60 heads, and the fixture-level savings never showed up anywhere in the accounting.

One boundary: I can only speak to domestic, mainland-US commercial installs. If you're specifying for the EU or UK, IEC 60598, CE/UKCA marking, RoHS and REACH all enter the conversation, and I'd defer to someone who runs those projects regularly rather than pretend I know.

Scenario 2: Replacement and Maintenance

This is the most common scenario and the one people overthink. When a fixture dies in a tenant space, the winning move is usually to buy the identical thing, not to upgrade.

• Write down what's actually up there — model, CCT, wattage, driver, voltage, mounting, finish. Photograph the label. Most people can't, and that's the real problem.
• Match the CCT. A 4000K fixture sitting next to a 3000K fixture is visible from across the room, and it's the complaint you'll get.
• Verify lead time with a rep, in writing, before you promise anything to anyone. Website "in stock" indicators have cost me more credibility than any other single issue in this category.

Buying spares is the piece of advice that finance pushes back on hardest, and it's the one I'd defend the most. Ordering 40 fixtures? Order 43. Not because decent spec-grade product fails often — it usually doesn't — but because lead times on the second order are frequently nothing like the first, and discontinued lines are real. A box of three fixtures in the back room costs less than one after-hours call-out.

If the fixture is on an emergency or egress circuit, stop. Buy the listed, code-compliant replacement and don't value-engineer it. That isn't a cost conversation.

Scenario 3: Design-Led Refresh

Some spaces exist to be looked at — the lobby, the reception desk, the boardroom, sometimes a client-facing lounge. That's where spending real money on a piece with a design pedigree is defensible, and it's the scenario where a pendant light catalog is the right starting point rather than a price sheet.

Practical notes on buying from a catalog: request finish samples instead of trusting a screen, confirm cord and canopy lengths before ordering (a 6-foot drop in a 14-foot lobby is an extremely common mistake), check dimming compatibility against whatever system is already in the building, and ask for lead time per finish — some finishes run considerably longer than others.

The Replica Question, Handled Honestly

A lot of people searching for a Flos Arco floor lamp are actually searching for a flos floor lamp replica, because the original is a serious line item. I'm not going to moralize about that. Plenty of people buy a lookalike for their own living room and understand exactly what they're getting.

But in a commercial building — and I say this as the person who has to answer when something goes wrong — the argument against it isn't primarily ethical. It's continuity and paperwork:

• No submittal package. No LM-79 report, no photometric file, no listing certificate tied to the specific product. If an insurer or a fire marshal asks, "it's the same as the original" is not documentation.
• No reliable second order. In three years, when you need another one or a replacement component, the listing may be gone and the new batch may not match.
• No parts channel. A piece like the Arco has a marble base, a telescoping stem, and a specific reflector. On the original, you can usually source the component. On a copy, the answer is often "buy another one."
• Nobody to call. A warranty is only as good as the entity that honors it, and a marketplace listing isn't an entity.

Here's the counterintuitive part, and it echoes Scenario 2: for a design piece, the fixture is frequently the smallest number on the invoice. Rigging, junction box location, the dimmer, the lift, and the electrician's time often cost more than the lamp itself. Saving 40% on the fixture and then paying for a second installation two years later is not a saving.

One more boundary. I'm describing commercial spaces with inspectors and insurers in the picture. Your apartment is a different risk profile, and I'm not going to pretend otherwise.

Scenario 4: Multi-Site Rollouts

If you're buying the same fixture for several locations, or you'll be buying it again next year and the year after, you're in a total-cost-of-ownership situation whether or not anyone uses that phrase in the meeting.

What actually drives cost over a rollout:

• Order-level consistency across batches
• Labor per service call, which dwarfs the fixture price
• Part and SKU continuity 24 months out
• Who administers the warranty claim and how fast a replacement arrives

During our 2024 office consolidation, the scoring spreadsheet clearly favored the cheaper vendor. My gut said no — their rep took four or five days to answer even simple questions, every time. I went with the higher-scoring, more expensive option and spent the following two weeks second-guessing the call. It turned out that four-day email cadence was an accurate preview of their service posture. I just didn't know that yet. Sometimes the slow reply is the whole story.

Years earlier we did the opposite. The low bid came in about $22 per fixture under the mid bid on a 120-fixture order — roughly $2,600 saved on paper. Over the next six months we ran nine service calls across six sites for driver failures. At around $450 a call, plus two calls that had to be scheduled outside business hours, the savings were gone and then some. The fixture itself wasn't even bad. The support channel was.

If you ask me, that's the whole argument in one paragraph. The sticker price is the most visible number and the least predictive one.

Which Scenario Are You Actually In?

Five questions. Answer them honestly and the scenario usually picks itself.

1. Is there a specifier, architect or GC involved? Yes → Scenario 1. Get the technical file before you get a price.
2. Can you name the exact model, CCT and voltage of what's currently installed? No → Scenario 2, and your first task is documentation, not purchasing.
3. Will people outside the company see this space? Yes → Scenario 3. That's where design money is defensible.
4. Will you place this order more than once? Yes → Scenario 4. Buy on continuity, not unit price.
5. If it fails in year three, who signs the paperwork? If the answer is you, you're in a commercial scenario, and the cheapest option needs to be the one with a paper trail.

Most of these decisions get made on the quoted unit price because it's the only number that's easy to compare side by side. That's also why so many of them get made wrong. The questions that actually predict your cost are unglamorous: is there a technical file, will they answer in a day or a week, can you buy the same thing again in two years, and who pays for the lift when it fails.

I've been running this category since 2020, and I can say with some confidence that the lowest quote has cost us more, in the end, more often than it has saved us. Not every time. But often enough that I stopped treating price as the deciding factor.

Pricing and lead times move constantly. Anything I've referenced here comes from our own purchase records through early 2026 — verify current numbers and lead times with a rep before you commit to anything.