Let me start with one number: $480. That's what it cost us to replace a single burned-out lamp in a chandelier last year. The lamp itself was $34. The other $446 was the service call, the lift rental, the two hours of labor, and the “while we're up there” inspection fee that appears on every invoice when maintenance touches a fixture twelve feet above a hotel lobby floor. The lamp didn't die of old age. It failed seven months after we installed it.

I'm the procurement manager for a 200-person hospitality group, and I've managed our lighting budget for six years — about $420,000 in cumulative spend, tracked in a system I built after getting burned on hidden fees twice. So when I say we kept buying lamps the wrong way, I'm not guessing. I've got color-coded spreadsheets that say so.

The Problem We Thought We Had

When the failed-lamp reports started piling up, maintenance said what facility teams always say: “The bulbs are junk.” So we did the obvious thing. We switched brands. Paid more per unit. Felt responsible. Six months later, the replacements started failing too.

My initial approach to this whole problem was completely wrong. I thought we had a quality problem, and the fix was spending more on lamps. But when I pulled the full purchase history and compared it against the fixture inventory, the data didn't implicate one brand. It implicated our buying process.

If you've ever tried to reconcile three vendors, two properties, and fourteen lamp SKUs in a single spreadsheet, you know the feeling. The units all matched on paper. The failures didn't.

What Was Actually Going On

Warranty Was a PDF Nobody Read

The first surprise from the audit: most of the failed lamps were still under factory warranty. We threw away something like $3,000 in claimable replacements over two years because nobody had the warranty terms in front of them at the moment of failure.

Take the Feit Electric LED bulbs warranty — the largest line in our purchase log. The terms weren't buried. The claim path was straightforward: proof of purchase, a photo of the failed lamp, and a form. But our purchase orders never captured warranty periods, and the maintenance crew wasn't going to dig through vendor PDFs with guests checking in.

It's tempting to think warranty claims aren't worth the paperwork. Fine. Then price the paperwork. At our blended labor rate of $85 per service visit, filing a claim instead of replacing a lamp was worth roughly $70 each time. We skipped it because the claim process wasn't part of our system. That's not a vendor problem. That's a spec-review problem. Period.

We Matched on Price, Not on Fixture

The second pattern was more embarrassing. The seagrass chandelier in the lobby and the Uttermost chandelier in the conference room both use decorative filament lamps with small bases. Gorgeous fixtures. But they trap heat — enclosed glass, tight fittings, almost no airflow. The original spec called for an enclosed-fixture-rated lamp. We didn't order that version.

The replacement lamps we bought weren't rated for enclosed fixtures. The spec sheet said so, buried in the installation notes. Nobody read that far. The result: median failure around seven months for lamps designed to last years.

There's an engineering reason for this. According to the U.S. Department of Energy (energy.gov), excess heat at the LED junction is a leading cause of premature failure and light degradation. Put another way: an LED lamp in a sealed brass fixture is running a marathon in a sauna. The lamp isn't failing. The environment is killing it.

We Ignored Color Quality Until the Whole Fixture Looked Wrong

The third mistake is silent. When we re-lamped one chandelier with budget-friendly alternatives, the numbers all matched: same wattage, same lumen output, same color temperature. But the wall behind the fixture looked flat. The wood tones looked muddy. In a hospitality space, that's not a nuance — that's the product.

The spec we'd ignored was CRI, or Color Rendering Index — defined by the International Commission on Illumination (CIE) as a 0–100 scale measuring how accurately colors appear under a light source compared to natural daylight. The original lamps were CRI 90. The replacements were CRI 80. Seeing them side by side in the same fixture made me realize the spec sheet wasn't marketing. It was the product.

By the time we noticed, four mismatched lamps made the other ten look wrong. So we re-lamped the entire fixture. Fourteen bulbs, two hours of labor, about $600 total. For a spec that was visible in the datasheet all along.

We Put Smart Bulbs in Applications That Didn't Need Them

And then there's smart lighting — the shiny-object corner of our purchasing history. We standardized on a smart WiFi bulb from Feit Electric for common areas, and honestly, it was great in the right spots: entry sconces, exterior fixtures, front desk lamps. Scheduling cut idle burn, dimming set the mood, and the app took ten minutes to configure.

But one client wanted smart bulbs in every guestroom ceiling can. Each can runs maybe 40 minutes a day. The total cost of ownership math was brutal: a $12 smart bulb plus configuration time versus a $3 standard lamp plus a $15 timer on the circuit. The timer wins in year one, and by year three it's not a contest. Nobody likes that answer. The spreadsheet doesn't care.

The Real Cost of Getting It Wrong

In our 2023 audit, we counted 412 failed lamps across all properties over eleven months. Sixty percent failed before the midpoint of their rated life. Thirty-four percent were replaced while still under warranty. The combined cost — replacement parts, labor, emergency visits — came to roughly $11,400.

I want to be precise here: these are our numbers, not industry averages. I've maintained the cost tracking system since 2019, so I trust the figures. The exact warranty recovery amount, I'd have to check the claim log — maybe $200. It was almost nothing.

Here's what the audit made clear: 87% of those failures were predictable before purchase. The fixture constraints were documented. The warranty terms existed. The CRI was in the datasheet. We just didn't read them, because we were busy comparing unit prices. We weren't buying lamps. We were buying warranties, heat tolerances, and color quality. We just didn't know it.

What We Changed (and You Can Copy)

We didn't switch to premium everything. We changed the buying process. Five rules, in the order they matter:

  1. Warranty is a specification, not a checkbox. Every lighting line item now includes the warranty period and a one-sentence claim procedure from the vendor. If a vendor can't state it clearly, we don't buy. The Feit Electric LED bulbs warranty made this easy — clear terms, clear claim path. The discipline was on our side.
  2. Confirm the fixture before ordering. Is it enclosed? What base size? Maximum wattage? Do we need dimming? The maintenance lead signs off on a fixture list before any lighting order goes out. It's a boring process. It saved us about $7,000 in year one.
  3. Select CRI by application. CRI 90+ for guest-facing and decorative fixtures. CRI 80 for storage and mechanical areas. On the lobby's seagrass chandelier, we standardized on 2700K with a CRI 92 rating — though I might be mixing that Kelvin figure up with a different project. Don't quote me on the exact color temperature. Quote me on the rule: check it before you buy it.
  4. Deploy smart bulbs where schedules do the work. Entries, exteriors, front desk, common stairwells. Not ceiling cans that run 40 minutes a day. Feit Electric's smart WiFi bulb is now a standard spec for specific applications — not everywhere, and not nowhere.
  5. Think about heat before you install LED strip lights. The question we get most often is “where to put LED strip lights?” Our internal answer: in coves and under cabinets with open airflow, not inside sealed aluminum channels that trap heat. And verify the power supply is rated for continuous duty — voltage drop is real, and the far end of a low-end strip will be visibly dimmer. Then everyone gets to read the fine print.

One more thing worth saying: none of this requires being a lighting engineer. It requires reading the four lines of a spec sheet that matter and asking the vendor one honest question: “What happens when this fails?” The vendors with a clear answer are the ones we still order from.

Here's my conclusion after six years and about 4,200 line items: lights don't fail. Buying decisions do. The most expensive lamp you'll ever buy is the one you purchase twice — once as a product, once as labor.

Read the warranty. Check the fixture. Compare CRI. Do the math.

And start a spreadsheet. It'll judge you. But it'll also save you.