Lighting practice
Three Checks I Run Before Any Lighting Order Ships (Price Isn't One)
Three checks decide whether a lighting order ships or goes back to the vendor: the beam pattern on a live sample, the install conditions on site, and the total cost if it fails. Price is the number I look at last. Honestly, it's the least useful thing in the whole file.
In our 2024 review cycle, I sent back 31% of first-article samples. Almost none of them had a wrong datasheet. The datasheet was accurate. The product was not.
The short answer, in the order I actually check
Check 1 — beam pattern over brightness. A headlight that throws 20% more lumens but smears the cutoff line is a downgrade, not an upgrade. This one catches people the most often.
Check 2 — install conditions, not the box. Most of the failures I flag on track lighting have nothing to do with the fixture. Dead-end feeds, reversed polarity, adapters on the wrong track rating. Boring stuff. Expensive stuff.
Check 3 — the perception spec. CRI, R9, color consistency across a batch, and how the light actually looks on the thing you're lighting. This is the check nobody budgets for and everybody notices.
None of those three is price. That's deliberate.
Who's writing this, and why the list exists
Quality and brand compliance manager at an automotive lighting distributor. I review every SKU and every spec sheet before it reaches a customer — roughly 200+ items a year, across headlight bulbs, work lights, and the architectural fixtures we stock for showroom and shop builds. Four years of it.
In our Q1 2024 quality audit, 11 of the 38 new SKUs in the pipeline had documentation errors that would have reached the customer. Not the product. The documentation. Torque specs, beam distance at 10m, and a color temperature claim that was off by about 500K.
That audit cost us a week of work. It probably saved us a $22,000 recall. We've run it every quarter since.
Why I stopped leading with price
My first year, I did it the other way. Compared quotes, took the low bid, moved on. Then we shipped a batch of bulbs where the base collar was machined 1.2mm off spec. The bulbs seated fine. The connectors held. And roughly 15% of them pushed the beam cutoff up into oncoming traffic, which is how I ended up on the phone at 8 a.m. with a fleet account whose vans were failing inspection.
We replaced 400 bulbs. The unit price difference we'd "saved" was around $0.40 per bulb. Replacement, shipping, and labor ran way past $6,000. On a $4,800 order. I still kick myself over that one.
The cheapest line item on the quote is usually the most expensive one on the invoice. That's the entire total-cost argument in one sentence, and I've never found a cleaner way to put it.
Check 1 — the beam pattern is the product
On H4 HELLA headlights and any other halogen replacement, the spec I care about most isn't wattage or lumens. It's filament position relative to the reflector. H4 is a P43t base with a dual filament, and the tolerance on where that filament sits is tight. Millimetres matter. Get it wrong and the light scatters. It looks brighter from the driver's seat. It's worse on the road.
What we actually do to a headlight bulb sample:
- Verify the E-mark and the regulation reference — ECE R37 governs the bulb itself, R112 governs the lamp unit it goes into. Check the current revision; these get amended.
- Mount the sample in a reference housing and photograph the cutoff at 10m.
- Burn it for 30 minutes, then re-measure. Cheap bulbs drift when they heat up.
That third step catches the most problems. Photometry measured cold, then measured again at operating temperature. Some samples shift their cutoff noticeably once they're hot. Not a spec-sheet lie — a design shortcut.
Here's the counterintuitive part. A 100W off-road-only bulb in a stock H4 housing does not give you more usable light. It gives you a scattered wash that looks impressive against a garage wall and useless at speed, plus heat that degrades the reflector, plus a compliance problem. 55/60W at the right focal point beats 100W at the wrong one. Every time we've tested it.
A lot of the returns we process start with a catalog search for "hella headlights bulbs" and end with someone who bought the right part number for the wrong model year. Different problem, same fix: our PO now carries a fitment checkbox that a human has to sign. Not a system. A person.
Check 2 — the install is where track lighting fails
Most of the how-to-install-plug-in-track-lighting content online stops at "snap the adapter in and twist." That's the easy part. The failures I get called about are mechanical and procedural:
- Adapter rated for a different track series. They look compatible. They aren't, and the contacts don't fully seat.
- Dead-end feed with no end cap, or no strain relief on the cord.
- Polarity reversed at the outlet. Track runs are polarity-sensitive on plug-in sections, and you can't tell by looking at it.
- Total load above the track rating — or worse, above the circuit rating, because someone added two more sections six months later.
NEC Article 410 covers luminaires and track, including listing and installation requirements. Check your local edition and ask your AHJ before you build. Standards move; my reading of them is not the authority.
The single check that saves the most service calls: assemble 10 feet of track on a bench, plug it in, load it with the actual fixture count, and run it for an hour. Most of the time it's fine. When it isn't, you find out in a back room instead of a finished showroom with the client standing next to you.
Check 3 — perception is a measurable spec
A bar chandelier over a dealer hospitality lounge. A career spotlight on the featured vehicle in the showroom — the one fixture whose entire job is to make one object the center of attention. Two totally different products, one shared problem: the datasheet tells you the color temperature and tells you nothing about how it renders the thing you're lighting.
CRI is a starting point, not an answer. R9 is where cheap fixtures fall apart, and R9 is exactly what you need if you're lighting a red car, leather, wood veneer, or a person's face.
We ran a blind test in 2023 with 14 people from sales and service. Same fixture, two versions. One at 90 CRI with decent R9, one at 80 CRI — basically the standard bargain option. They looked at a painted fender and a leather seat cover. 11 of 14 called the higher-CRI version "more premium" without knowing which was which. The cost difference was $26 per fixture. On a 40-fixture showroom, that's about $1,000 for a measurable change in how customers read the space.
Never expected perception to be the easiest thing to quantify. Turns out it's usually the cheapest upgrade on the whole project.
ANSI/IES RP-2 gives recommended practice for merchandising areas if you need a starting point on illuminance. But the number that matters is the one your client sees, not the one on the plan.
Where this list breaks down
Three honest caveats.
First, the bench test and the burn-in are a distributor's process. If you're buying six fixtures for one shop, running a lab isn't economical. Buy from someone who already does it and ask for the results. That's what a distributor is for.
Second, under real time pressure this whole list compresses. Last October I had about 90 minutes to approve a substitute fixture before a showroom opening deadline. Normally I'd want a photometric file and a sample in hand. There was no time. I went with our existing vendor on track record alone. It worked out. It could have gone the other way, and I knew that when I signed off.
Third, the value-over-price argument holds for anything installed, inspected, or seen by a customer. It's weaker for consumables you'll replace in six months no matter what you buy. Don't over-apply it. Sometimes cheap and disposable is genuinely the right call.
And a time bound on all of this: the regulation references and pricing here reflect what I had in front of me in Q4 2024 and early 2025. Standards get amended, tariffs move, supplier quality drifts. Verify current revisions and current quotes before you sign.