I still have the photo on my phone. March 2018. A 48-inch chandelier hovering over a 36-inch dining table. The client's text just said: "Is this what we discussed?"

It wasn't. And that was the second of five lighting mistakes I made between 2017 and 2023. The mistakes cost my company roughly $47,000 in returns, rush replacements, and discounts I offered to keep clients from walking. But the real cost was the six months I spent wondering if I'd chosen the wrong career.

Here's the thing: every one of these mistakes came from the same root cause. I was buying lighting like it was a commodity. It isn't. Lighting is a system of invisible variables—color temperature ranges, beam angles, spectral distributions—that don't show up on a product page. And nobody explains them until you've already made the mistake.

I'm writing this so you don't have to learn the way I did.

Mistake #1: Assuming "3500K" Means the Same Thing to Every Vendor

This is the one that started it. In late 2017, I was specifying under cabinet lighting for a 42-unit kitchen renovation in a mid-rise apartment building. Budget was tight, so I split the order: GE Enbrighten under cabinet lighting for 30 units (plug-in, easy install, solid reviews) and GE premium under cabinet lighting for the remaining 12 (hardwired, higher CRI, "matching" color temperature).

On paper, both were 3500K. On the wall, they weren't.

Two weeks after install, the property manager sent me a photo of two adjacent cabinets. One was warm white. The other was… beige. Not a huge difference in a photo, but against a white backsplash, it looked like two different manufacturers had been involved.

Here's what I didn't know: "3500K" is not a color. It's a bin—a range. Per ANSI C78.377, the standard tolerance for a nominal 3500K LED is ±145K. So one fixture could be 3355K and another 3645K, both labeled "3500K," and both technically compliant.

But that's the spec sheet talking. In practice, I've measured fixtures from different product lines that were nearly 400K apart. Under cabinet lighting is the worst place for it, because the light bounces off a white backsplash at close range. Every 100K feels like 200K.

The fix cost $6,800 in replacements plus a crew re-visit. I'm not 100% sure the mixed product lines were the whole story—there might have been a driver inconsistency too—but the color temperature gap was the visible symptom. The lesson wasn't "buy better products." It was "ask for the binning data before you order, not after."

Mistake #2: Sizing a Chandelier by Feel Instead of Formula

The 48-inch chandelier over the 36-inch table is the one I still get teased about. In my defense, the client's designer said "something dramatic." In reality, I should have known the formula:

Room width in feet × 2 to 2.5 = chandelier diameter in inches.

A 12-foot dining room gets a 24–30 inch chandelier. A 14-foot room gets 28–35 inches. My room was 13 feet. I ordered 48 inches. The designer loved it. The client didn't.

The deeper issue, though: that formula is for dining rooms. Chandelier lighting in a two-story foyer has a different rule—width in feet plus height in feet, converted to inches. Kitchens over an island: sum the island's length and width, convert to inches. Each context has its own math.

The $7,200 lesson: a custom resize to 32 inches, plus re-crating and shipping the original fixture back. The client kept the smaller version. She also kept a running joke about "the UFO phase" of her renovation.

I still use the formula. I also still call that client every year to make sure she's not still mad. She is, a little. Fair enough.

Mistake #3: Believing "Full Spectrum" Means "Good for Plants"

In 2019, I got pulled into a small indoor farming project. Not my usual lane—I mostly do commercial and industrial—but the client was a friend of a friend, and the order was $4,700 for grow lights. Easy money, I thought.

The plants grew. Tall, thin, and pale. "Leggy," in horticulture terms. The client's operations manager asked me what spectrum I'd specified. I said "full spectrum." He said, "That doesn't mean anything."

He was right.

Does blue light help plants grow? Yes—but the popular articles get the mechanism wrong. Blue light (400–500nm) drives photomorphogenesis: compact growth, thicker leaves, better branching. Red light (600–700nm) drives photosynthesis: total biomass. Most "full spectrum" LED grow lights have plenty of red and just enough blue to pass a marketing claim. The result is plants that grow fast but stretch.

"Full spectrum is a label, not a measurement," he told me. "Show me the SPD chart or we're guessing."

What I should have done: asked for the spectral power distribution (SPD) chart, not the product description. A good vegetative-stage spectrum has a blue-to-red ratio somewhere in the 1:3 to 1:5 range, with peaks around 450nm and 660nm. Take this with a grain of salt—the exact ratio depends on the crop—but most leafy greens seem to do fine in that range. Not "full spectrum." Specific.

The $4,700 order ended up as a partial refund plus a reorder from a specialty supplier. Total loss was closer to $4,700 once I factored in the rush shipping on the replacement. The friendship is a little more professional than it used to be.

Here's the counterintuitive part: conventional wisdom says "more blue equals better vegetative growth." In my experience, too much blue without sufficient red produces compact but slow-growing plants. Like most things in lighting, it's a ratio problem, not a "more is better" problem.

Mistake #4: Learning About Beam Angles on a Gallery's Dime

This is the one that almost ended my career.

In early 2021, I specified a spotlight package for a small art gallery. Eight adjustable spotlights, 15° beam angle, warm white, high CRI. The spec was clean. The installation was not.

On the wall, each fixture created a hot center with a visible ring around it—like a flashlight held too close. The gallery owner, who had been patient through two other delays, sent me an email with the subject line: "This is not what we agreed to."

Here's what I'd missed: beam angle labels aren't standardized across manufacturers. A "15°" from one vendor can behave like a "24°" from another. The fixtures I ordered had a beam angle that narrowed to about 8° at the center—great for highlighting a single vase, terrible for lighting a 24-inch painting.

I should have asked for the IES file. That's the photometric data file that shows actual light distribution, not just a number on a spec sheet. Every serious commercial lighting vendor has them. I didn't know to ask.

The rework cost $9,900: new fixtures, new trims, and a crew that had to re-aim and re-commission the system. The gallery owner kept me on the project. Barely. I spent the next month wondering if "lighting procurement specialist" was still the right job title for me.

That spotlight order nearly ended my career. Not because of the money—because I realized I'd been specifying products I didn't fully understand.

Mistake #5: Reading "Zigbee Compatible" as a Yes/No Question

The last one happened in 2022, on a 60-room boutique hotel project. The client wanted smart lighting controls—scene setting, scheduling, the works. I specified GE Cync-compatible fixtures and a third-party Zigbee controller. The product pages said "Zigbee compatible" on both sides. What could go wrong?

Everything, for about six weeks.

Zigbee isn't a single protocol. It's a family. Zigbee 3.0, Zigbee Home Automation 1.2, Zigbee Light Link—they share a radio, not necessarily a conversation. My controller was Zigbee 3.0. The fixtures were running an older profile. They could see each other on the network. They just couldn't agree on what a "dim" command meant.

The commissioning took three times as long as quoted. The client lost confidence in the system—and in me. I think the controller vendor genuinely believed their product was compatible. The word "compatible" just means different things to different people. We eventually swapped the controller for a Cync-native gateway ($14,200 for hardware and re-commissioning). The system works now. The client still asks, every time we talk: "So is the lighting going to behave this time?"

That's the cost that doesn't show up on the invoice: the credibility tax.

What I Do Now (The 10-Minute Version)

I don't have a perfect process. But I have a pre-order checklist that's caught 23 potential issues in the last 18 months. It's not glamorous. It works.

  1. CCT binning: Ask "what's the tolerance?" not "what's the color temperature?" If the vendor can't answer, that's your answer.
  2. Chandelier sizing: Use the formula for the room type. Don't trust "it looks proportional in the showroom."
  3. Grow lights: Request the SPD chart. Look for peaks at 450nm and 660nm. Ignore "full spectrum" as a standalone claim.
  4. Spotlights: Ask for the IES file. Read the beam angle at 50% intensity, not the marketing number.
  5. Zigbee: Ask "which profile?" and "which gateway has been tested with these fixtures?" The word "compatible" without a tested gateway is a marketing term, not an engineering one.

None of this is advanced. It's just the stuff I wish someone had put in front of me in 2017. These usually catch the biggest issues, though not always the subtle ones.

Here's what I've come to believe: most lighting problems aren't product problems. They're information problems. The product is usually fine. The gap is between what the spec sheet says and what the space needs. Closing that gap is the job.

I'd rather spend 10 minutes explaining spectral distributions to a client than deal with a $4,700 refund six months later. An informed client asks better questions and makes faster decisions. And honestly, they're more forgiving when something goes wrong—because they understand what "wrong" even means.

That's the part that took me $47,000 to learn. The rest was just paperwork.