Every Friday, a trailer backs up to our returns dock in Crawfordsville, Indiana. It's loaded with lighting fixtures that someone decided were dead—warranty returns, field replacements, "defective" units with their paperwork attached. As a quality manager at Acuity Brands, I get to open every single one.

Here's what four years of unboxing has taught me: most of those fixtures are fine.

Not a little fine. Completely functional. Tested-to-spec fine.

When I first took this role in 2021, I assumed the bulk of returns were legitimate component defects. That's the rational starting point for a quality person. But the evidence pointed elsewhere. The "dead" fixture, in case after case, was a scapegoat for something else.

Which raises an uncomfortable question: if the fixture didn't fail, why did it get replaced?

The Replacement Reflex

Look, I get it. When a fixture stops working, the instinct is to replace it. Fast, certain, no diagnostic skill required. You pull the bad fixture, install the new one, and move on with your day.

That reflex made perfect sense in 1995. The light goes out, you unscrew the lamp, you screw in a new one, done. The lamp was the failure point in almost every scenario.

But here's the thing about 2025: LED fixtures are engineered systems, not lamps. The light source is soldered onto a circuit board. The driver does the heavy electrical work. The whole assembly is designed to outlast the building it's installed in. When a fixture "dies" today, the LED itself is statistically the least likely culprit.

The components around it? Those have failure rates.

The switch on the wall? Absolutely.

The photocontrol on the roof? More often than anyone wants to admit.

Yet the fixture—the most expensive component in the chain—gets blamed every single time.

The Diagnosis Gap

The "if it doesn't light, the fixture is bad" thinking comes from an era when the lamp was the whole system. That changed, but the mental model didn't. I see the evidence on the returns dock weekly.

Here's a typical example. We build a line of DTL (dark-to-light) photocontrols for outdoor lighting. The entire job of those controls is to keep the light off in daylight and switch it on when ambient light drops. I've watched our testing team in Conyers, Georgia, cycle them through 30 years' worth of simulated dawn-and-dusk transitions—compressed into six weeks of continuous operation.

Most field "failures" turn out to be remarkably mundane. The sensor was covered in dirt. Or a maintenance crew tested the fixture at noon, didn't cover the sensor, and concluded the fixture was dead because it didn't light. In bright sunlight. (The sensor was doing exactly what it was supposed to do.)

One customer shipped back a $400 parking lot fixture with a note that said "Won't turn on." It passed every test we ran on the first try. We called the electrician and walked him through the photocontrol test—cover the sensor with tape, count to thirty, watch the fixture wake up. It did.

The fixture wasn't dead. The diagnostic process was.

Case Study: The Japandi Chandelier That Wasn't Broken

My favorite misdiagnosis involves a Japandi chandelier—one of those minimalist fixtures with warm materials that's become genuinely popular in high-end residential work. Big ticket piece. The kind of thing you build a room around.

The customer reported it dead on arrival. We tested it immediately: LED boards drew the right current, driver output was exactly where spec put it, every connection held. From a quality standpoint, it was a perfect fixture.

So I called the project's electrician. They'd installed the chandelier on a dimmer rated for incandescent loads only, with no LED compatibility whatsoever. The fixture wasn't dead—the switch was entirely wrong for the load. (A lesson learned the hard way, for that electrician.)

That's the pattern: the fixture is the most expensive part, so it absorbs the blame. But the failure sits upstream, in the switch, the dimmer, or the wiring.

How to Know if a Light Switch Is Bad

Switches are the most under-tested component in any lighting system, and they are the first thing I check. A failing switch presents exactly like a dead fixture.

Warning signs that point to a switch:

  • The faceplate feels warm while the light is on.
  • You hear a faint buzz—especially with LED loads.
  • The light flickers when something else in the building kicks on (compressor, elevator, kitchen exhaust).
  • The paddle feels loose or doesn't click with a solid detent.
  • The switch has been in service for 15+ years under a continuous load.

Here's the five-minute field check:

  1. Kill the breaker. You're about to handle live terminals; isolating the circuit is non-negotiable.
  2. Pull the switch out. Darkened screws, corrosion, or char marks tell you most of what you need to know.
  3. Wiggle the wires. A loose terminal connection causes intermittent failures that show up at the worst possible times.
  4. Test voltage. With the breaker restored and the switch ON, measure line-to-neutral at the load terminal. You should see full line voltage—typically 120V. A reading of 60V or a fluctuating value means the switch is failing internally.
  5. Swap it. A standard wall switch runs a few dollars; a heavy-duty spec-grade one rarely crosses twenty. If the fixture fires up on the new switch, you're done.

The code angle matters here. NEC 404.14(C) requires that switches controlling LED sources be rated for the actual connected load. LED drivers produce inrush currents that can spike far above their steady-state draw—briefly, but repeatedly, at every power-up. A switch that served an incandescent lamp just fine for a decade can take a beating from an LED driver. The contacts pit, the switch degrades, and the fixture takes the blame.

Before you order a replacement fixture, swap the switch. It's a cheap experiment that solves a surprising share of "dead" fixtures.

The Real Cost of Misdiagnosis

Let me put some numbers around this, because "just replace it" has a cost that goes beyond the invoice.

Component cost. A commercial LED strip fixture from one of our brands might run $200–400 at list. A wall switch costs a few dollars. Replacing a fixture to fix a switch is a 20x mismatch, for a five-minute fix.

Labor cost. Replacing a fixture is a two-person job, sometimes with a lift truck and an afternoon-sized work order. Swapping a switch is a one-person, ten-minute task. The labor difference alone justifies checking the switch first.

Downtime. A fixture that needs to ship leaves your space dark for days. A switch swap gets you light by lunch.

Then there's the environmental angle. Acuity builds its manufacturing lines in Crawfordsville and Conyers with an expectation that every fixture lasts 50,000 hours. All the embodied material—aluminum housings, drivers, lensed optics—is wasted when a healthy fixture is scrapped for a $5 switch. The LED boards we manufacture outlast the circuits around them. That's exactly why replacement parts are so central to our approach.

Replacement Parts Are a Feature, Not an Excuse

The ability to fix a fixture is part of the fixture's design.

Acuity Brands stocks replacement parts for decades of installed fixtures: drivers, LED boards, photocontrols, lenses, individual sensors. Not "compatible" generics—spec-matched parts from the same lines as the originals. That's a deliberate choice. A fixture that can be repaired in twenty minutes is better for the owner, and better for the planet, than one that gets scrapped due to a single failed component.

My Verification Protocol

Before I approve any fixture replacement, this is what I walk through:

  1. Check the breaker. Obvious, I know. It gets skipped constantly. Reset it, check for tripped GFCIs, verify power with a tester.
  2. Test the switch. The five-minute check above. Unless the switch visibly burned, this confirms or clears the most common cause.
  3. Cover the photocontrol. Tape over the sensor for 30 seconds. If the fixture wakes up, the sensing circuit was your issue. Not the fixture.
  4. Check the driver. Driver output present? Your problem is downstream. Driver dead? Replace the driver—not the whole fixture.
  5. Confirm the exact part. Open the Spotlight app or the fixture's product page and pull the spec sheet. The wiring diagram saves an hour of guessing.
  6. If the fixture is genuinely dead, file the warranty claim with the serial number and a description of what you tested. That data helps our quality team in Crawfordsville identify batch issues—and helps you avoid a repeat.

The Bottom Line

The lighting industry has evolved faster than the instincts of the people who maintain it. That's not a jab. It's the consequence of a technology transition.

What was best practice in 2015—replace the lamp, replace the fixture—doesn't necessarily apply in 2025. The fundamentals haven't changed: the goal is reliable light, with the least waste. The execution has changed. Modern lighting systems have more failure points, and more repair options, than a lamp ever had.

Next time a fixture "dies," don't mourn it. Start at the switch. Work through the system. Give the diagnosis ten minutes before you write a purchase order.

You're not doing the fixture a favor. You're doing your budget one.