I handle procurement for a 60-person eyewear operation — three storefronts and an online shop. Roughly $300K a year, spread across 11 vendors. Rectangle frames are the single biggest line item in that budget, which means I've had plenty of chances to get this wrong.

For the past two years I've been running the same comparison across suppliers: identical spec sheet, identical order volume, identical QC checklist. Two quotes keep coming back. One lands around $2.10 FOB. The other around $4.20. On paper, they describe the same product.

They don't deliver the same product. Here's how I break the comparison down, and what actually decides which quote I take.

The four things I compare

  • Frame geometry — true rectangle vs. "rectangle-ish"
  • Sizing — women's vs. men's, and whether the label means anything anymore
  • Black — gloss, matte, and batch-to-batch color drift
  • Landed cost — the invoice number vs. the cost per sellable unit

Then a short section on stock vs. OEM at the end, because that decision cuts across all four of the above.

Dimension 1: "Rectangle" is a range, not a spec

This is where most of the money hides.

A true rectangle frame has corners close to 90° and a straight top bar. A "soft rectangle" rounds those corners 6–10mm and gives the top bar a slight bow. Both get written up as rectangular sunglasses on a spec sheet. They are not the same product, and they don't sell the same way.

The $2.10 supplier delivered something in between. Corners rounded. Top bar with a gentle curve. On a shelf under store lighting, it reads as a wayfarer hybrid — not a rectangle. Customers called it out. "These aren't the ones in the photo."

When I put a $2.10 frame and a $4.20 frame side by side under the same light — same spec sheet, same SKU name — I finally understood why the return rate on the cheap batch was three times higher. It wasn't quality. It was geometry.

The $4.20 supplier shipped actual rectangle geometry, and their spec sheet listed corner radius and top-bar angle. Two numbers. That's all it took for me to know what I was buying before the container left port.

Takeaway: ask for corner radius and top-bar angle in writing. If a supplier can't put numbers on it, "rectangle" is a marketing word, not a specification.

Dimension 2: women's vs. men's rectangle sizing has largely converged

Used to be simple. Womens rectangle sunglasses ran 52–54mm lens width. Rectangle sunglasses mens ran 56–58mm. That was the rule for years, and a lot of suppliers still ship to it.

The market moved. Women's rectangle frames are running wider now — 54–56mm, oversized coverage. Men's rectangle has narrowed a notch to roughly 54–56mm. The gap collapsed from about 6mm to about 2mm.

Which means the gender label is nearly useless as a sizing instruction. It's a merchandising label now, not a measurement.

The $2.10 supplier ships by label. Women's batch arrives at 52mm. That fits like a narrow men's frame, and it looks small against the face our customer base is buying for. The $4.20 supplier sends a size chart with per-SKU lens width, bridge, and temple length, and I order against measurements instead of labels.

Takeaway: buy from a measurement chart — lens width, bridge, temple length, total frame width. If a supplier only quotes "women's" and "men's," you're buying their assumption, not your spec. Same applies to men rectangular sunglasses — measure it, don't trust the tag.

Dimension 3: black is not one color

Black rectangle sunglasses are our highest-volume SKU and the easiest one to mess up, honestly. "Black" covers at least four finishes:

  • Gloss black — high shine, shows every fingerprint
  • Matte black — flat, scratches show up fast
  • Rubberized or painted black — soft-touch, wears at the hinge
  • "Black" that's actually a very dark tortoise, or a warm black that reads brown under warm store lighting

Batch one from the cheaper supplier came gloss. Batch two, semi-matte. Batch three looked black in the warehouse and dark brown on the shelf. Not technically defective — just inconsistent. Customers don't file that as a defect. They just don't buy it.

The $4.20 supplier ships a physical color chip per batch and pins the finish per SKU on the spec sheet. Boring. Effective. There's a reason I stopped arguing about the price gap after that.

This is also where the compliance side lives. If your listing says "UV400," "polarized," or "eco-friendly acetate," that's a substantiation problem, not a marketing one.

Per FTC advertising guidelines (ftc.gov), claims must be truthful, not misleading, and substantiated with evidence. A supplier saying "yes, UV400" over email is not evidence. Ask for the test report.

Same logic applies to a "recyclable" frame claim — the FTC Green Guides require the claim to hold up where the customer actually lives. If the supplier can't document it, don't put it on the listing.

Takeaway: per-batch color chip in writing, finish pinned per SKU, and test documentation for any performance claim you publish. No documents, no claim.

Dimension 4: the landed number, not the FOB number

Now the one that actually decides the order.

Run 1,200 units of black rectangle.

Cheap quote: 1,200 × $2.10 = $2,520 FOB. QC reject rate came in around 9% on the matte black SKU — paint rub at the hinge, inconsistent finish. That's 108 units I can't sell at full price. Replacement units, return handling, and sorting labor pushed the effective per-sellable-unit cost to roughly $3.15.

Expensive quote: 1,200 × $4.20 = $5,040. Reject rate about 1.5%. Effective cost around $4.35.

Gap went from $2.10 to about $1.20. Still a real gap — but way smaller than the invoice suggests.

And on the matte black SKU specifically, it flipped. The cheap supplier's reject rate on matte black pushed effective cost above the "expensive" quote once I counted sorting labor. That one surprised me. I'd assumed the premium was always just a premium.

Takeaway: compute cost per sellable unit, not per ordered unit. Reject rate, sorting labor, replacement freight. The invoice is the start of the math, not the end.

Stock vs. OEM: is the SKU proven yet?

Quick version:

Stock gets you 10–15 days, no logo, fixed sizing, low commitment. Good for testing whether a rectangle style actually moves. OEM gets you 30–45 days, custom logo, custom lens width, MOQ around 300–600. Good once a SKU has proven itself.

This is where the sourcing side has shifted most between 2021 and 2025. Back then, stock basically meant whatever was left in a warehouse and OEM meant 90 days. Stock quality has climbed, OEM lead times have compressed. What was best practice in 2020 doesn't map cleanly onto the way this category works now.

Which quote I take

Scenario-based, because there's no universal answer:

  • Heavy return exposure, published claims, brand at stake — the verified supplier. The effective cost gap is small enough that return rate and listing liability decide it.
  • Testing a new rectangle style, pop-up or seasonal — stock, accept some geometry drift, inspect the first 50 units before listing.
  • Wholesale to other retailers — OEM with per-batch color chips and a written size chart. Your buyers will measure the frames.
  • High-volume matte black — run the landed-cost math per batch. The cheap quote flips on this SKU more often than you'd think.

There's something satisfying about finally having this on one page. After two years of post-mortems and a shelf full of frames I couldn't move, the whole thing comes down to four numbers: corner radius, lens width, finish per batch, and cost per sellable unit.

Everything else is the sales call.