The short version

Here’s the short version: I run project orders for a lighting sales agency, and in late 2022 a plant café project cost us about $4,300 because I replaced specified components with cheaper lookalikes. The original spec had a mirror chandelier for atmosphere, cylinder downlights for the order counter, grow bars for a plant wall, and Acuity Brands DTL dark to light photocontrols for exterior lighting. It was a good specification. The trouble started when I let unit price make engineering decisions.

The lesson is not “buy premium or suffer.” It’s that the cheapest number on a quote is not the lowest-cost decision when you ignore what the light actually has to do. Since that project, I’ve kept a pre-order checklist that has caught 14 similar mistakes. This is the story of the ones that got away.

Why I keep a list of my own mistakes

I’ve been handling commercial lighting orders for eight years. I say that not as a credential—it’s a confession. In that time I’ve personally made and documented 11 significant mistakes, totaling roughly $28,000 in wasted budget between replacement parts, labor, and apologies.

Let me be precise about what I mean by “my mistakes”: I approved products I had not tested. I accepted substitutions without checking the details. I signed submittals that said “or equal” when nothing about the part was equal. The mistakes were not the manufacturer’s fault. In each case, the specified products worked exactly as promised.

The project: a plant café with conflicting priorities

A boutique café on the west side of Indianapolis asked us to quote the lighting for their new space: a mirror chandelier above the main seating area, a row of cylinder downlights along the counter, a plant wall with integrated grow lighting, and exterior wall sconces that needed dusk-to-dawn control.

The owner said: “The quote is higher than we hoped. Can you make it work?”

That’s a fair question. My mistake was answering it with the cheapest version of every line instead of an honest conversation about tradeoffs. Not zero. Not huge. Just real, and wrong.

Mistake one: the cylinder downlight that “saved” $600

The counter lighting was straightforward: 12 commercial LED cylinder downlights at 3000K. We specified an Acuity fixture because it had published photometric data, consistent color, and a known driver.

The owner needed the number down, and in my head, the downlights were the easiest place to cut. I found an imported cylinder downlight that looked identical in the spec sheet photos. Same aperture. Same trim. Same claimed color temperature. The price was about $50 less per fixture. Across a dozen, that looked like $600 in savings.

In our showroom, side by side with the original, the substitute looked close. In the café—above warm wood and glass pastry cases—the difference was not subtle. The original made the wood look like wood. The substitute made the counter feel flat, almost green. Both labels said 3000K and 90 CRI. The CRI number is an average; it does not tell you how a fixture renders saturated reds. The substitute’s datasheet did not mention red rendering at all.

Then the drivers started failing. The first one died around six months, the next a few months later. When we tried to use the warranty, the supplier’s product line had changed, and the remaining fixtures no longer matched.

We replaced all 12 units with the original specification. The re-installation labor, service calls, and new fixtures totaled about $2,900. A $600 saving became a $2,900 loss because I optimized the wrong number.

When our distributor checked stock on the original fixtures, the response came back from the Acuity Brands Lighting facility in Crawfordsville, Indiana: normal lead time, fixtures available. There had never been a supply problem. I had invented one to justify a bad decision.

Mistake two: substituting the photocontrols

The exterior lights were designed to follow daylight. In the spec, that meant Acuity Brands DTL dark to light photocontrols. The electrical contractor, looking to trim his bid, substituted a less expensive photocontrol instead.

I signed the submittal. That part is on me.

Three weeks after opening, the owner sent me a photo: a bright late-October sunrise, and the café’s outdoor lights still burning. The first service visit found one failed photocontrol. The second replaced all six outdoor photocontrols with the specified Acuity units.

Here is the detail that still irritates me: the twist-lock socket is standardized under ANSI C136.10, so the budget units fit perfectly. Fit was never the issue. Performance and reliability were.

The apparent saving was $126. The service cost was $490, plus a goodwill credit to keep the account. We were using the same word—“photocontrol”—and meaning different things. I meant a tested component. The contractor meant a part that screwed into the same socket. Those are not the same specification.

Mistake three: “how long does a grow light need to be on?”

This one did not need a fixture order. It needed research, and I skipped it.

The café’s plant wall mixed tropical foliage with culinary herbs. The owner asked me directly during a walkthrough: “How long does a grow light need to be on?”

I answered like a person who sells general lighting and did not want to admit the limits of his expertise. I said something like, “You can run it during opening hours. That should be fine.”

It was not fine. The herbs got leggy. The leaves washed out. The grow-light manufacturer’s crop chart called for 14 hours on and 10 hours off for that particular mix, at a specific intensity at the canopy. Some plants wanted a shorter day. A few wanted longer. The answer was not “whenever the café is open.”

The useful answer is: it depends on the crop’s daily light integral (DLI), then the photoperiod, then the intensity at the canopy. For many leafy greens and culinary herbs, 14 to 16 hours is a reasonable starting point, but you confirm it with the fixture manufacturer’s data instead of guessing. We fixed the wall with a simple timer and adjusted light levels. The replacement plants and a service call cost roughly $550—less than the other mistakes, but more than it costs to simply say “I’ll check the horticultural data and get back to you.”

The mirror chandelier, meanwhile, reminded me that decorative light and task light are different tools. It sparkled and bounced light around the room, but it did not add useful illumination to a single task. That was the cylinder downlights’ job. I cut the downlights first because they looked boring next to a mirror chandelier. Boring is often the thing actually doing the work.

The checklist that has caught 14 mistakes since

After that project, I turned the embarrassment into a checklist. It is not complicated, but it works:

  • If a fixture is substituted, demand published LM-79 and LM-80 test data and check its DLC listing where applicable. A spec sheet without a date is not a spec.
  • Name the photocontrol manufacturer and model in every submittal. “Photocontrol” is a category, not a specification.
  • If a client asks “how long” or “how much,” the answer starts with “let me check,” not with a guess.
  • Write down the cost of being wrong before you celebrate the cost of being cheap.

The list has caught 14 potential errors in the 18 months since. It has paid for itself many times over.

Where the advice ends

Look, I am not saying budget products are always wrong. They have a place when the application is forgiving, the client understands the tradeoff, and failure will not hurt their business. What I object to is accidental risk—choosing a lookalike because it looks equivalent on a spreadsheet when you never checked the details that matter.

I am also aware that time matters. I wrote this from project notes dated late 2022 and updated them in Q1 2025. Product lines change. Manufacturers change what they build and where they build it. Lead time and stock information goes stale. Verify those details with a current distributor or rep before building a schedule around my numbers.

The core lesson has not gone stale: do not let the price tag make engineering decisions. I did, and it cost $4,300 plus a chunk of credibility. The checklist costs nothing. Start one before you need it.