Lighting practice
LED Lighting Procurement: A Scenario-Based Guide to Getting TCO Right
Why There's No Single 'Best' LED Lighting Choice
I've been managing facilities and procurement budgets for a mid-sized commercial furniture manufacturer for the past seven years. Our annual lighting spend — bulbs, fixtures, drivers, emergency units — runs somewhere around $34,000, give or take. And in that time, I've learned one thing the hard way: asking 'which LED is best' is the wrong question.
When I compared our Q1 2023 retrofit costs against Q4 2024's — same building, similar square footage — the difference wasn't the product. It was the scenario. We'd been buying the same class of fixture for a showroom, a warehouse, and a grow rack, treating them as interchangeable. They're not. The TCO spread across those three applications was roughly 30%, and most of it came from ignoring application-specific requirements.
So here's how I now break it down. Three scenarios. Different priorities. Different TCO traps.
Scenario A: Task and Under-Cabinet Lighting — The CRI Trap
This is your showroom, your inspection bench, your customer-facing counter. For us, it's the finish-sampling area where clients compare veneer samples against paint chips. If the light shifts the color, the sale dies.
What matters most: CRI (color rendering index) and beam consistency.
Most buyers look at lumens and price. That's the iceberg's tip. Here's what I learned after replacing 46 under-cabinet fixtures in 2023: the $18-per-fixture option from a general supplier had a CRI of 82. The $27 ledvance under cabinet lighting spec'd at CRI 90+. We tested both on the same veneer samples and the color shift was visible to three out of five people in the room.
That $9-per-unit gap looked like $414 in savings on the PO. Then our quality manager flagged that clients were second-guessing color approvals, adding roughly 20 minutes per sample review. Across 90+ reviews in a quarter, that's 30 hours of design time — and at our loaded labor rate, $1,350 in soft costs.
"Saved $414 upfront. Burned $1,350 in rework time. The 'cheap' fixture wasn't cheap."
TCO items to model: unit cost + driver replacement interval (cheap fixtures typically use non-serviceable drivers; at 12 hrs/day, plan on swapping the whole fixture at year 3–4) + labor for install + any downstream operational impact. For task lighting, if CRI drops below 90 and color is customer-critical, add "rework risk" as a line item.
Scenario B: Smart / Networked Lighting — The Zigbee Mesh Question
This is where most of our clients (we supply furniture to commercial offices) are heading. They want occupancy sensing, daylight harvesting, zone control. That means Zigbee, and Zigbee means you have to think about network topology, not just fixtures.
What matters most: Zigbee routing and range — and I don't mean the marketing range on the box.
In mid-2024 we quoted a 14,000 sq ft office fit-out. Vendor A's proposal used a straightforward ledvance sylvania smart+ smart home lighting gateway with 38 devices. Vendor B used a different ecosystem at roughly 12% lower device cost. The catch: Vendor B's spec sheet listed a 'range zigbee' figure of 150 ft line-of-sight. Vendor A's didn't list range at all — they listed device count per coordinator and mesh depth.
We almost went with B. I'm glad we didn't. A mesh network's range isn't a spec you read; it's a function of how many mains-powered devices are in the mesh. Every hardwired fixture acts as a Zigbee router, extending the network. Battery devices don't. Vendor B's plan had 11 routers in a space where the far corner was 90 ft from the nearest. Best case: intermittent control. Worst case: a zone that needs a manual reboot every time the cleaning crew unplugs something.
The TCO hidden line: mesh troubleshooting. Our facilities team logged roughly 6 hours per month of 'smart system' calls in the first year of a different building's system. That's 72 hours annually — call it $3,600 in tech time — for a network that should have been designed with adequate routing density from day one.
Practical check: before signing off on any Zigbee spec, ask the vendor for a device map showing router density. If more than 20% of the floor area isn't within two hops of a router, push back.
Scenario C: Horticultural / Grow Lighting — The PAR vs Lumen Distinction
This one's niche, but it comes up if you're in food service, pharmaceutical R&D, or education (ag programs). We have a small client who runs an indoor herb operation and asked us to source lighting for a 600 sq ft space.
What matters most: PAR efficacy (µmol/J), spectrum match, and photoperiod control compatibility.
When people search "what is the best led grow light," they often get answers driven by lumens or wattage. Both are misleading. A 200W fixture at 2.7 µmol/J delivers 540 µmol/s of usable light. A 240W fixture at 2.2 µmol/J delivers 528 µmol/s. The cheaper unit looks brighter on a lumen meter but grows less. And over a 12-week cycle with 18-hour photoperiods, the efficiency gap compounds — we modeled roughly 8% more electricity per usable photon on the lower-efficacy unit.
The P&L approach I used: take total fixture cost + driver replacement at 50,000 hrs + electricity at your actual rate × annual run hours + (if applicable) HVAC load from heat output. On a 12-fixture order, the higher-efficacy option came in ~$480 more upfront. Electricity offset was ~$310/year at our rate. Payback under two years. That's before yield differences.
How to Tell Which Scenario You're Actually In
Most procurement mistakes happen because someone applies Scenario A thinking to a Scenario B problem. Here's the quick filter I use before gathering quotes:
- Does the light affect a decision someone else makes? (Color-critical? Inspection? Customer-facing?) → Scenario A. Prioritize CRI and serviceability over unit price.
- Will the lights be networked? → Scenario B. Require a router-density map and budget 5–10% of system cost for first-year commissioning and troubleshooting.
- Is the light a production input, not overhead? → Scenario C. Ignore lumens. Compare µmol/J and spectrum. Model electricity over 10,000+ hours, not 1,000.
And if you're buying across all three — which we do — keep separate TCO models. My spreadsheet has three tabs and one rule: never benchmark a warehouse fixture against a showroom fixture just because they have the same base type. That's how you end up paying $1,200 to save $400.
I built the first version of that spreadsheet in 2022 after a missed-spec order cost us $2,800 in express replacement. It's not glamorous, but neither is explaining to the CFO why the 'value' option failed inspection twice.