Why I Stopped Buying Blue Light Glasses Based on the 'Blocking Percentage' on the Box
I haven't trusted a "blue light blocking percentage" on a product box since September 2021. If you're sourcing blue light filter glasses for a retail client, you probably shouldn't either.
I'm a procurement lead handling anti blue light glasses orders for importer and optical retail accounts—seven years now. In that time I've personally made (and documented) more sourcing mistakes than I'd like to admit, and roughly $14,000 of my company's budget went to fixing them. Most of those mistakes were in the first two years. The most expensive one taught me a lesson I apply to every order today: the number printed on the box tells you almost nothing about what the lens actually does.
Here's why I think the percentage claim is the least useful spec in the entire category, and what I check instead.
The percentage is a marketing format, not a standard
From the outside, "90% blue light blocking" sounds precise—almost scientific. The reality is that there is no single, agreed convention for what that number measures.
Does it mean 90% of everything between 380nm and 500nm? 90% at one peak wavelength? Weighted for perceived hazard? Averaged across the visible blue range? Measured at lens center or near the edge? With a clear tint or a brown one? Each of those produces a genuinely different result, and a supplier can lean on any of them.
Most buyers focus on the biggest number on the sample card and completely miss the test method behind it. And that's exactly where a bad order starts.
The order that cost me $3,400 in one afternoon
I want to say it was 2,000 units, though I might be misremembering the exact figure—the redo stung more than the count. The supplier's spec sheet proudly listed "92% blue light blocking at 415–455nm." The sample looked right. Clear blue light glasses for a men's line, clean AR coating, decent weight. I approved it, we cut a PO, and the goods shipped.
Six weeks later, we sent a random pull to a third-party lab before releasing them to the retailer. The averaged transmission came back around 68% across that range—with a sharp cliff below 420nm. The "92%" was a peak number at one narrow slice of the spectrum, presented as if it were the whole story.
We rejected the bulk, ate the redo, ate a two-week delay, and ate a very awkward call with the buyer. That's when I learned that "blocking %" is a headline, not a spec.
What actually decides performance in a blue light lens
Three things matter more than any single percentage, and none of them fit comfortably on a product box:
- Wavelength-specific transmission data. I want a curve, not a number. Ideally 380–500nm in 5 or 10nm steps, with the test method named. A supplier who can send me that in under 24 hours usually turns out to be a supplier worth keeping.
- Lens construction. Is the blue light management in-mass (the lens material itself) or a coating? In-mass tends to hold up better over time and through edging. Coatings can be excellent too, but you need to know which you're buying because the failure modes are completely different.
- Visible color cast. A truly clear blue light glasses lens that claims very aggressive blocking is telling you one of those two claims is doing less than advertised. If the sample looks noticeably yellow or brown, your retail buyer will get returns from men who expected a clear lens. That's a real cost, and it never shows up on a spec sheet.
Prescription blue light glasses add a fourth variable: the lens substrate and coating have to survive surfacing and edging without losing the effect. I've seen coated plano lenses test fine and then show measurable drop-off after prescription processing because the coating wasn't formulated for that workflow. Ask for a post-processing sample, not just a plano.
"But my retail buyers want the high number on the box"
Fair. I get that the number sells. Retail shelves, Amazon listings, and marketing decks all want a big percentage. My position isn't that you should never show a percentage—it's that you should buy on verifiable data and use the percentage as marketing language, not as your acceptance criterion.
There's a legal dimension here too. Per FTC advertising guidelines (ftc.gov), health and safety claims—including vision-related claims—must be truthful and substantiated with competent and reliable scientific evidence. If a supplier can't produce wavelength-specific test data, you're the one carrying that exposure, not them. The number on the box is a marketing asset you generate from real test data, not a substitute for it.
When this advice matters less
I don't think every buyer needs the full lab packet on every order. If you're sourcing promotional giveaway glasses—branded event swag, not sold to consumers as eye-comfort products—a basic claim is usually enough, and going deeper just slows you down. If your end customer is a discount chain treating blue blockers as an impulse add-on, cost and delivery matter far more than spectral precision, and that's a legitimate call.
But if you're selling prescription blue light glasses to opticians, or building a branded anti blue light glasses line for men where the claim is part of the pitch, the test report isn't optional. It's the product.
What my sourcing checklist looks like now
Every blue light glasses order—clear, tinted, plano or prescription—runs through the same five gates before a PO goes out:
- Wavelength-specific transmission curve, with test method named and date stamped.
- Written confirmation of in-mass vs. coated construction.
- Color cast described in plain language, not adjectives.
- Pre-production sample verified against the curve, and a post-processing sample for prescription runs.
- Claim language reviewed against what the data can actually support before anything hits a box.
None of that is glamorous. But it's the difference between a line that repeats and one that gets returned.
So no, I don't think there's a "best" blocking percentage to source against. There's only data you can verify and data you can't. Pick the suppliers who make the first kind easy to get.