There is no single right way to source rectangle sunglasses
I sign off on incoming eyewear before it reaches a customer — roughly 180 SKUs a year, spanning women's frames, men's frames, and the black rectangle sunglasses that a few of our accounts reorder by the pallet. When buyers ask me what to look for when sourcing rectangle sunglasses, the honest answer is that a brief that protects one buyer is a waste of money for another. Your situation decides the checklist, not the other way around.
Three things determine almost everything about how you should source: how much of your budget sits in tooling versus product, how brand-critical your frame color and logo placement are, and what a bad batch actually costs you once returns, credit notes and freight are counted. Those three variables sort nearly every buyer into one of three scenarios.
- Scenario A — a first run. You're testing whether the shape sells at all.
- Scenario B — you already have a winner and you're scaling it with brand requirements attached.
- Scenario C — unbranded stock bought to resell on price.
Same product category. Three completely different briefs.
Scenario A: You're testing the category for the first time
You're planning somewhere in the range of 300 to 1,500 pairs, two or three shapes, mostly plain black, and you want to know which one sells before you commit real money. The temptation here is to treat the first order as a branding exercise — custom temple logo, a Pantone frame color, a printed pouch.
Skip all of it. Rectangle shapes are among the most common molds in any eyewear factory's catalog, so on an existing mold you're paying for product, not novelty. Tooling is the single largest fixed cost in a first run, and it's the one cost you can defer until you have evidence you need it.
Where the money should go instead is in two places: hinge quality, and a documented transmittance test on the lens.
Here's the part that runs against what most first-time buyers assume. The failure that hurts a launch is not the frame color or the logo — those are the things buyers stare at, so they get checked. It's a hinge that loosens. Frame color you notice at receiving. Hinge failure you notice six weeks later, in the reviews.
Ask the factory for the hinge test method and the cycle count it's quoting against — not just the word "metal core." A supplier that can't name its test method is telling you something.
We ran a first order of about 1,200 pairs in 2023 with a custom temple logo. The logo was perfect. Eleven percent came back from the retailer for loose hinges — that was roughly $6,000 in credit notes, or closer to $7,000 once you count return freight. The logo cost more than the hinge upgrade would have, and the logo didn't generate a single return. Or rather, it generated zero returns; the hinge generated all of them.
One more thing in this scenario: don't assume women's rectangle sunglasses and men's rectangle sunglasses are the same frame with a different label. They aren't. A 2mm difference in temple length is the difference between "comfortable" and "slides down my nose all day," and that's a fit problem your customer will blame on you, not on the factory.
Scenario B: You already sell a rectangle and you're scaling it
Now you're at 5,000-plus units a run, you have a brand color, a logo on the temple, maybe polarized lenses. The math flips. Tooling is amortized across volume, so per-unit consistency is where the money belongs.
This is the scenario where color standards start to matter, and it's worth knowing the actual tolerance language before you write a PO.
Industry standard color tolerance is Delta E < 2 for brand-critical colors. A Delta E of 2–4 is noticeable to trained observers; above 4 is visible to most people. Reference: Pantone Color Matching System guidelines.
So define your master standard, approve it physically, and require each production lot to be matched against it — not against the last lot. Matching lot-to-lot compounds drift. Matching to a retained master doesn't.
Now the counterintuitive part, and it's the one I'd argue hardest for. If your hero SKU is black rectangle sunglasses, Delta E is rarely your problem. You can't really miss black. What goes wrong instead is gloss level and geometry.
Two vendors can both ship a frame that measures as black, and one comes in matte-ish and the other in a hard gloss. Side by side in a display case they read as two different products from two different companies. That's not a color failure — it's an unspecified finish. Write the gloss level into the spec sheet and keep a reference pair on the shelf.
The geometry issues are the ones that actually get returned: temple tip alignment, lens-to-frame seating, and symmetry. A 1.5mm asymmetry is invisible on a table and obvious on somebody's face. I've stopped trusting a table check entirely — we now photograph every approved pair straight-on and measure centerline symmetry before signing off.
If you carry both men's and women's rectangle sunglasses, resist scaling one into the other. The bridge width and temple length should be specified separately, even if the frame styling is shared.
Scenario C: You're buying unbranded stock to resell on price
No branding, mixed pallets, your customer is a discount retailer, and your margin per pair is thin enough that it's measured in cents. The standard play is to collect six quotes and take the lowest number.
The counterintuitive advice: the lowest FOB is frequently not the lowest landed cost once credit notes, rework and return freight are on the books. And the lever that moves landed cost isn't another round of price negotiation — it's the defect acceptance criteria.
ISO 2859-1 sampling procedures are the usual basis for a written AQL. An AQL of 2.5 on major defects is a common starting point for consumer goods inspection. Whatever number you agree, put it in the PO — along with what happens if the lot fails.
Most buyers spend three rounds of emails shaving a few percent off the unit price and never once discuss what happens when 4 percent of a lot is unusable. The price gap between quote two and quote five is usually small. The gap in what happens when a lot fails is 100 percent of the lot.
One trap specific to this scenario: black is often treated as the safe SKU in a mixed buy, on the theory that black hides defects. It hides color variation. It does not hide surface finish variation — scuffs, polish marks and flow lines show up as gloss inconsistency on black more than they do on a light tortoise. If you're buying black in volume without a written finish spec, you're buying whatever the line was running that day.
Low-MOQ runs on existing molds are possible in most cases, but the minimum depends entirely on the factory's current schedule and its tooling. Get the number in writing rather than assuming it, because it changes month to month.
Which scenario are you actually in?
Four questions, asked in order.
- Can you name your two best-selling lens widths right now, without looking them up? If you can't, you're in Scenario A regardless of how much you're spending.
- Would a visible color shift force you to discount the batch? If yes, you're in B, and you need a retained master standard and a finish spec.
- Is your margin per pair smaller than the cost of processing one return? That's C. Stop negotiating price and start negotiating the replacement clause.
- Is a single SKU carrying a large share of your revenue? Then you're in B for that SKU even if the rest of your catalog is A or C. Hero products get hero specs.
You can legitimately be in two scenarios at once — testing three new shapes while scaling your proven rectangle is normal. What isn't normal is writing one brief for the whole catalog. That's how a first-run sku ends up carrying custom tooling it doesn't need, and how a proven sku ends up on a spec sheet that says "black."
If you take one practical step from this: write your brief in three lines and send the same three lines to every factory you quote, so the comparison is real. Line one is quantity and whether tooling is new or existing. Line two is the one or two specs you won't move on — for A, the hinge; for B, gloss level plus a color standard; for C, AQL and the replacement clause. Line three is what happens if the lot fails.
Everything above reflects what I was seeing through Q1 2025. Frame costs, coating availability and standard revisions move faster than most spec sheets get updated, so verify current terms and current standard versions before you budget against any of it.
And my experience here is mostly mid-volume programs — call it 1,000 to 20,000 pairs per run, mid-tier materials. If you're working at 100,000 units or in luxury acetate, your constraints are different from mine, and the tooling math in particular may not transfer at all.
