Grow Light Sourcing: 8 Questions B2B Buyers Ask Before They Sign the PO

I run lighting procurement for a 140-person commercial greenhouse operation. Six years on the job, roughly $300,000 a year across fixtures, drivers, controls and spares, and somewhere north of 40 lighting tenders since 2019 — retrofits, two expansions, and one OEM project I'd run differently if I got a do-over.

What follows is the FAQ I wish somebody had handed me in year one. These are the questions buyers actually ask during sourcing. Two of them are questions most of us only think to ask after the PO is signed, which is exactly the problem.

• What's actually included in the price per fixture?
• Why do two fixtures with identical PPF numbers perform differently?
• Does the cheapest cost per µmol actually win?
• What compliance paperwork should I have on file before installation?
• OEM or private label?
• How do I tell a real manufacturer from a trading company?
• Is "LED vs HPS" still the right question in 2025?
• What goes wrong after install, and who pays for it?

1. What's actually included in the price per fixture?

The fixture price is usually 55–70% of what we end up spending. The rest is freight, duties, mounting hardware, cabling and connectors, drivers (sometimes quoted separately, and it isn't always obvious), controls integration, a spares float, and warranty logistics.

In a 2023 tender for 1,400 fixtures, one vendor came in at $218 per fixture with everything bundled. Another quoted $196. I nearly shortlisted the second one until I read the line items: no drivers, a different mounting bracket, and freight listed as "quoted separately at time of shipment." That cheaper quote needed about $11,400 in extra hardware and roughly $9,800 in freight I'd assumed was included. On paper it was 10% cheaper. Delivered, it was about 6% more expensive. The fine print did the damage.

Then there's the part nobody puts in the comparison spreadsheet: photons per dollar of electricity. Our blended rate in 2023 was around $0.14/kWh. To match the photon output of a 3.4 µmol/J fixture, a 3.0 µmol/J unit has to draw roughly 53W more at the same output. Over 4,000 run hours a year, that's about 212 kWh per fixture — call it $30 a year, per fixture. Across 1,400 fixtures, roughly $41,000 a year. That's arithmetic, not a promise: your tariff, your run hours and how you manage dimming decide the real number.

2. Why do two fixtures with identical PPF numbers perform differently?

Because PPF is a quantity, not a quality. Two fixtures can both claim 1,700 µmol/s and behave nothing alike over a canopy.

What actually moves the result:

• Spectrum. R:B ratio, far-red content, and how much energy lands outside the photosynthetic response curve. PPF tells you how much light exists. Spectrum is what determines how the crop grows.
• Distribution. Beam angle, lens design, and the uniformity map across your actual bay. A fixture with a beautiful center reading and 30% falloff at the edges is a different product from a flatter one.
• Tolerance. Does the datasheet say ±5% or ±10%? Across 1,400 fixtures, that's the difference between a lighting plan and a guess.
• Measurement conditions. Numbers should come from a full-fixture LM-79 test at a stated ambient temperature and drive current. Numbers calculated from LED bin datasheets are not the same thing.

Worth knowing: the terminology here is standardized under ANSI/ASABE S640, which is the reason PPF and YPF mean specific things rather than whatever a brochure wants them to mean.

I assumed "same PPF" meant "same result" back in 2021. Didn't verify. Ran a side-by-side in two identical compartments and the crop under the flatter-distribution fixture came in more even and hit our target weight about four days sooner. Same nominal PPF on both datasheets. That was a $12,000 lesson in reading the whole document instead of the headline number.

3. Does the cheapest cost per µmol actually win?

Sometimes. Over 4,000 annual run hours, usually not. The trouble is that cost per µmol is easy to flatter, and everyone in the market knows it.

Common ways a headline efficacy figure gets dressed up:

• Measured at nominal drive current and 25°C, while you'll be running it dimmed in a 28–32°C greenhouse.
• Calculated from LED bin datasheets instead of a full-fixture measurement, which quietly omits driver and optical losses — typically 8–12%.
• Tolerance quoted at the optimistic end of the band.
• Efficacy quoted without saying whether the driver is included.

So when I see 3.5 µmol/J on a one-page flyer and 3.3 µmol/J on an LM-79 report from an accredited lab, I know which one goes in the comparison sheet. Bottom line: a 3.5 that only exists on a bench at 25°C is worth less to me than a 3.3 I can hold the vendor to inside my own facility.

4. What compliance paperwork should I have on file before installation?

For North America, at minimum:

• ANSI/UL 8800 — the safety standard for horticultural lighting equipment. Not the same claim as "the driver is UL recognized."
• DLC listing — matters if you want utility rebates, and it forces a certain documentation discipline on the vendor.
• LM-79 test reports from an accredited lab, plus LM-80 data and TM-21 projections if anyone is promising output maintenance over time.
• IP rating under IEC 60529, and check that it's the fixture that's rated, not just the connector.

For Europe, add CE marking, the applicable ErP obligations, and RoHS. Nothing exotic — but all of it needs to be in the file before an insurer or an inspector asks, not after.

I asked one vendor point-blank whether a fixture was UL listed. They said yes. What they meant was the driver carried a UL recognized component mark and the fixture had a self-declared compliance mark. Not the same claim at all. We were using the same three words and meaning two different things. Now I ask for the certificate number and the file, in writing, before anyone gets shortlisted.

5. OEM or private label?

Private label means taking an existing catalog product, putting your brand and part number on it, maybe changing the housing color or connector. Low MOQ, fast, and you're mostly buying logistics and branding. OEM means changing the actual spec — spectrum, form factor, driver, dimming protocol, sometimes firmware. Higher MOQ, NRE or tooling costs, a longer timeline, and somebody has to own the certification.

I went back and forth on this for about three weeks in 2023. Private label would have shipped in nine weeks. The OEM route took roughly five months and around $34,000 in non-recurring engineering, but it matched our existing mounting rails and our 0–10V control system exactly. We chose OEM because the retrofit labor savings showed up on every single install, not just the first one. Your math depends on how many units you're genuinely committing to.

The question to answer before either path: who holds the certification when the fixture carries your name on the label? That single line changes the risk profile, and it usually lives somewhere in the terms and conditions.

One pattern I noticed while shortlisting for that project — manufacturers with real OEM capability tend to get technical very fast. When Valoya came up in our 2023 evaluation, the conversation went straight into spectrum targets, drive current and PPFD uniformity maps rather than a catalog PDF. That's usually a decent signal.

6. How do I tell a real manufacturer from a trading company?

Ask these five things and the answer usually reveals itself:

1. Who owns the factory — and can they show it on a live call (not a stock photo set)?
2. Which LED bins, and which driver brand? Vague answers here are a red flag.
3. Can they explain why the spectrum is what it is for your crop, or do they just say "full spectrum"?
4. Do they have full-fixture LM-79 reports, or only LED datasheets?
5. What happens on an RMA — who pays freight, and what's the replacement lead time in writing?

A trading company isn't automatically a deal-breaker. Some are excellent, and they'll tell you plainly that they source from a partner factory and hold the relationship for you. The problem is the ones claiming to be the factory who can't answer question two.

7. Is "LED vs HPS" still the right question in 2025?

The fundamentals haven't changed — daily light integral, uniformity, spectrum, canopy penetration. What changed is the execution and the economics around it.

Back in 2020 the question was mostly "can LED compete?" For a lot of applications now, that's settled, and the conversation has moved on to how you spec the fixture and how you control it. But I'd push back on anyone who says the LED-versus-HPS question is closed for every operation. If your power is cheap, if your capital budget is tight, or if you're leaning on lamp heat to carry part of your heating load, HPS can still pencil out. I've walked greenhouses where it does.

What was best practice in 2020 isn't automatically right in 2025 — largely because dimming, controls integration and per-fixture data have changed what "good" looks like. The physics didn't move. The toolkit did.

8. What goes wrong after install, and who pays for it?

Three things, in order of how often they bit us:

1. Warranty start date. Ours ran from shipment, not commissioning. We lost about seven weeks of coverage on phase one because the fixtures sat in a warehouse (unfortunately, for longer than anyone planned). Read that line twice.
2. RMA freight. A "5-year warranty" means very little if you're paying $40 a unit to ship a failed driver back across an ocean.
3. Spare parts window. Ask how long drivers and boards will be available. Two years is not long enough for a fixture you intend to run for ten.

Honestly, I've never fully understood why some manufacturers hold regional spare inventory and others don't. My best guess is it comes down to distributor relationships and warehouse cost more than anything technical. Either way, get the commitment in writing with a number attached to it.

And one more thing that belongs in the contract and usually isn't: the acceptance test. Define what "as specified" means before delivery — measured PPFD at defined points, in your facility, at your drive current. Otherwise "within tolerance" becomes a very elastic phrase.

My experience here comes from roughly 40 tenders, mostly 2 to 12 hectare greenhouses in Europe and North America. If you're running 20,000+ fixtures across multiple sites, your procurement reality is genuinely different from mine — you'll have leverage and legal review I don't have, and some of this advice will need adjusting.