Luminous efficacy — lumens out per watt in — is the headline number on every LED datasheet. It’s also the number most likely to be quoted in a way that flatters the product. Here’s how to read it, and what “good” looks like today.
Chip efficacy is not fixture efficacy
LED packages in lab conditions exceed 220 lm/W. But a fixture adds driver losses (3–8%), optical losses through diffusers and lenses (5–15%), and thermal derating at operating temperature. A “200 lm/W LED” routinely ships as a 150 lm/W luminaire. Only fixture-level efficacy, measured to LM-79 or EN 13032, belongs in a comparison.
Current benchmarks by category
| Category | Acceptable | Good | Premium |
|---|---|---|---|
| High bay | 140 lm/W | 160 lm/W | 180+ lm/W |
| Panels (back-lit) | 110 lm/W | 130 lm/W | 150+ lm/W |
| Panels (edge-lit, UGR<19) | 100 lm/W | 115 lm/W | 130+ lm/W |
| Linear (office) | 110 lm/W | 130 lm/W | 150+ lm/W |
| Floodlights | 120 lm/W | 140 lm/W | 160+ lm/W |
| Downlights | 90 lm/W | 105 lm/W | 120+ lm/W |
Notice the pattern: every optical refinement — microprisms for glare, deep reflectors for downlights, high-CRI phosphors — costs efficacy. A 130 lm/W UGR<19 panel may be a better product than a 150 lm/W bare-diffuser one.
When higher efficacy is worth paying for
Efficacy is a running-cost lever, so its value scales with hours and wattage. A warehouse running 200 W fixtures 6,000 hours a year saves real money on a 20 lm/W improvement; a hotel corridor downlight at 10 W saves cents. Rule of thumb: prioritize efficacy on anything over 100 W or over 4,000 annual hours, and prioritize light quality (CRI, glare, binning) everywhere people linger.
Watch for these on datasheets
- Efficacy quoted at 25 °C ambient — fine for offices, optimistic for a 45 °C ceiling void or a Gulf-region car park.
- Efficacy quoted at 4000K/CRI 80 while the project needs 3000K/CRI 90 — subtract 10–15%.
- “Up to” figures that apply only to the largest, coolest model in the family.