The short answer: 30 to 36 inches
How high above a table should a chandelier be? The bottom of the fixture should sit 30 to 36 inches above the tabletop. In an 8-foot room, that usually lands the fixture about 6.5 to 7 feet off the floor. For ceilings above 9 feet, add roughly 3 inches of hanging height per extra foot of ceiling. Dining tables typically want 30–34 inches; kitchen islands can go a touch higher — 36 inches is common — because people stand near them.
That part is simple. The expensive part is everything around that number: the fixture's listing, the control system it talks to, and who actually picks up the phone when the spec sheet says "compatible" and the job site disagrees.
Why I stopped trusting generic lighting advice
I'm a procurement manager at a 190-person commercial contractor. I've owned our lighting line item — roughly $310,000 annually across new builds and retrofits — for six years. I've negotiated with more than 25 vendors and tracked every order in a cost spreadsheet I've rebuilt twice (the first version, honestly, was a mess).
Early on, I treated lighting like any other line item: three quotes, pick the middle one, move on. That worked until a Q2 2023 retrofit where the "value" outdoor fixtures we bought needed a control gateway that wasn't in the quote. The gateway alone was $2,800, plus $1,400 in integration labor. The vendor's spec sheet said "Zigbee compatible" in bold. What it didn't say — until we called — was that it only spoke Zigbee through a proprietary hub. A 12% cost overrun hiding in a footnote. (Note to self: read the footnotes before the bold text.)
Outdoor Zigbee: where the real compatibility questions live
Outdoor Zigbee controls are a case study in why "compatible" is a four-letter word in procurement. Zigbee is a mesh protocol — every device can act as a repeater — which is great for coverage and terrible for accountability when something drops offline.
Three things I now verify before any outdoor Zigbee order:
- Mesh depth. A single gateway in a parking lot won't reach a fixture 400 feet away. Ask for a site plan with radio range estimates, not marketing copy.
- Firmware update path. Cheaper controllers sometimes require a proprietary dongle for firmware updates. That dongle becomes a single point of failure when the vendor discontinues it. This happened to us in 2024 with a different brand's bridge.
- Lumen output at 2700K–4000K. "Up to 8,000 lumens" is usually a 5000K number. Warmer color temperatures — the ones most commercial exteriors actually want — often run 15–20% lower. If your photometric plan was built on the headline number, you'll be adding fixtures.
In my experience, the vendors who answer those three questions in writing, within a day, are the same vendors who don't vanish when a fixture fails in month 14.
GE Lighting, Nela Park, and the rep locator shortcut
GE Lighting's Nela Park campus in Cleveland sounds like trivia until you're specifying a large outdoor project. It's been the company's lighting headquarters and research center since 1912 — one of the first planned industrial research parks in the U.S. What that matters for procurement is less about history and more about supply chain: a brand with a century-old R&D footprint usually has documentation, test data, and a support network that survives product line changes.
That's the theory. The practical version: use the GE Lighting solutions rep locator before you need it. I've made the mistake of hunting for a rep after a fixture failure at 4 p.m. on a Friday (not recommended). Locating your regional rep during the spec phase gets you their direct line, a backup contact, and early product change notices. GE's transition toward the Savant/Cync ecosystem is a good example — reps had transition timelines months before distributors posted them.
I'd argue the rep relationship is usually worth more than a 3–5% quote difference on a mid-size project. The rep who flags a discontinued SKU two months early has saved me more money than any bidding war, and I say that as someone whose job is literally running those bid comparisons.
When to admit you're out of your depth
Here's the part that took me too long to learn: a procurement manager should not be the last line of defense on photometric design. I can read a spec sheet and spot a hidden gateway fee. I cannot tell you whether a proposed layout meets IES illuminance standards for a loading dock, and I shouldn't pretend to.
The best lighting rep I've worked with — a regional outfit, not a big brand — told me on our second call: "We're strong on industrial high-bay, but for your sports field lighting, talk to these two firms." That cost him a small commission and earned him every industrial order I've placed since. The vendor who says "that's not our strength" is telling you something about everything they do claim.
What I'd do differently: bring a lighting designer or rep into the conversation at the schematic phase, not the bid phase. The cost of an early review is usually $0–$500. The cost of re-specifying fixtures after a photometric report fails is a different number entirely — ours was $11,000 in change orders on a single project.
Where this breaks down
None of this applies if you're buying one decorative chandelier for a residential dining room. Hang it 30 inches above the table, use the existing junction box, and skip the vendor analysis — the stakes don't justify the spreadsheet.
It also breaks down on very small commercial jobs (under $5,000 in fixtures) where the admin cost of three quotes and a rep call outweighs the savings. In those cases, buy from a distributor with a return policy and move on.
But if your lighting spend is north of $50,000 annually, the chandelier-height question is the easy part. The expensive ones are the specs you didn't know to ask about.
Rule of thumb from six years of invoices: the quoted price is usually accurate. The scope behind it rarely is.