I've managed our frame purchasing budget for six years now — roughly $380,000 a year at a 46-person eyewear distributor, spread across 30-some independent optical accounts. When another buyer asks me how to source rectangular glasses, I don't answer right away. Not to be cagey. It's that rectangular frames sourcing isn't one decision. It's at least three, and each one has a different cost structure, a different MOQ reality, and a different place where the money leaks out.
What follows is the split we use internally. I didn't design it in a meeting room. I built it after two first orders died on the shelf and one reorder that looked like a 6% saving turned into a four-figure correction.
How the three scenarios separate
The dividing line isn't order size. It's whether you have your own sell-through data for the shape.
- Scenario A — first order. You're buying rectangular eyeglasses for the first time, or buying from a new supplier, and you don't yet know which widths and colors move.
- Scenario B — reorder and line extension. You know what sells. You're replenishing, adding colors, or widening an existing range.
- Scenario C — your own label. Your logo, your packaging, your mold, your compliance file.
Almost every buyer I've had this conversation with assumed they were in B and were actually in A — or assumed C because they once asked for a custom temple color. Getting it wrong is expensive in a specific way: you spend Scenario C money on a Scenario A problem, and then there's no budget left when the real C work starts.
Scenario A: first order, no history
Your dominant cost here isn't unit price. It's inventory you can't move, plus the cost of exiting a bad bet. A $14 frame that clears in ten weeks is cheaper than an $11 frame that sits in a drawer for eight months. I've run that arithmetic both directions and it isn't close.
Buy stock, not development
A supplier's existing stock program does two things your own development run can't: it ships in weeks rather than months, and it lets you order 30 pieces instead of 300. You'll pay more per frame — usually 15% to 30% more than a made-to-order run. That premium isn't a markup to argue down. It's the price of not committing capital to a shape you haven't tested. Budget it as research, because that's what it is.
Spend the budget on size coverage, not branding
This is where I'd push back on the usual advice. When money is tight, the instinct is to concentrate it — one model, one color, one width, 100 pieces. In rectangular frames that's backwards. Lens width and bridge width drive fit, fit drives returns, and year-one returns are the most expensive thing you'll handle.
Our first order was six SKUs at 30 pieces each — 48, 50, 52 and 54 mm lens widths in two colors. Landed cost was about $3,180, though I'd pull the invoice before quoting that figure anywhere. Eight weeks later we knew two of those widths were producing 71% of sales and the 54 mm had barely moved. That knowledge cost us less than the 4% discount another supplier offered on a single-SKU, 300-piece run.
Custom temples and logo work can wait. In year one they don't sell frames, they just make the SKU count smaller and the exit harder.
The rimless assumption that keeps costing people money
There's a durable belief in this category that rimless rectangle glasses should be cheaper than full frames, because there's less material — no acetate rim around the lens. It sounds like arithmetic. This was true a couple of decades ago, when frame material was the dominant cost line and mounting labor was cheap. That's changed.
Material savings on a rimless rectangle today are trivial. What you pay for instead is machining: drilling, bushings, mounting screws or cord, and fitting labor that has to align two lenses to a tolerance measured in fractions of a millimeter. The after-sales profile is worse, too. A rimless rectangle with a chipped lens edge or a loosened mount is a repair or a replacement, not an adjustment.
I'm not saying skip rimless — we stock it and it sells. I'm saying don't model it as the budget alternative to a full rectangle frame. In our returns log the rimless rectangle repair rate has run meaningfully higher per unit than acetate rectangles. If I remember correctly it was close to triple in 2024, though part of that was one mounting batch we should have caught at intake. Which is the recurring theme in this whole article.
Scenario B: reordering, and the mix is settled
Your objective changes here. You're no longer buying a product, you're buying consistency — and consistency is where buyers quietly lose money, because a quote sheet only shows price.
Two quotes from our 2024 files, same rectangle glasses frames, same nominal spec:
- Supplier A: $9.80 per frame, MOQ 300, 35-day lead time.
- Supplier B: $11.40 per frame, MOQ 120, 28-day lead time.
I went back and forth on this for the better part of two weeks. On paper, A wins by 16%. What that 16% actually buys you, though, is a five-month inventory commitment if the SKU moves 60 units a month — in a shape you might need to adjust before the last third clears. B's MOQ was two months at the same velocity. So the real comparison wasn't $9.80 against $11.40. It was $9.80 with a quarter of a year of cash parked in a rack, against $11.40 with the option to change course.
What settled it wasn't price at all. B shipped a golden sample and a batch measurement sheet with every order. A's third batch arrived visibly darker than the first, and we had no retained sample to prove it against — only an original invoice and a memory. Replacing that batch cost more than the annual saving the 300-piece MOQ produced.
Five minutes with a caliper at intake beats five days of returns processing. I know that reads like a slogan. It's the line item I point at whenever someone asks why we don't just take the cheapest quote.
If you're in B, the practical changes are boring. Keep a signed golden sample per SKU. Keep one unit from every incoming batch. Measure three things on arrival — lens width, bridge width, temple length — because those three catch most of what actually goes wrong. We inspect 100% of cartons on a new batch, then run AQL 2.5 at general inspection level II for subsequent deliveries. None of it is sophisticated. It's just cheaper than the alternative.
Scenario C: your own label
This is where the money genuinely moves to the front of the process, and where buyers who are comfortable in A and B tend to get hurt.
Tooling: the line that disappears
Tooling for a simple rectangle runs roughly $1,200 to $3,500 per style, depending on whether it's a modification of an existing mold or a new one. Spread across 3,000 pieces, that's $0.40 to $1.17 a unit — which is why a supplier will tell you tooling is basically free. It is free, at 3,000 pieces. At 600 pieces the same tooling is $2.00 to $5.80 a unit and it's the largest single number on your landed cost.
So don't fold tooling into a unit price you'll compare against other quotes. Once it's inside the number, you stop seeing it, and you stop asking who owns it. We paid for a mold that lived in a supplier's tool room for three years before anyone read the ownership clause properly. Nothing went wrong. That was luck, not process.
Compliance is the part you can't improvise
Rectangular eyeglasses for adults are a low-risk category, which people hear as no requirements. That's not the same thing. Requirements shift by market, and they attach to the label rather than to the price.
Frames sold into the EU are generally assessed against EN ISO 12870 (ophthalmic optics — spectacle frames), with metal components that contact skin tested for nickel release per EN 1811. Measurement conventions follow ISO 8624, which matters more than it sounds: if you and your supplier measure lens width differently, your golden sample is worthless. In the US, spectacle frames sit under FDA ophthalmic device regulations at 21 CFR Part 886; most non-prescription frames are Class I and exempt from premarket notification, which is not the same as exempt from labeling rules. Non-prescription sunglasses add ANSI Z80.3. And origin and material claims have to be substantiated — the FTC's advertising rules apply to eyewear the same way they apply to anything else.
The procurement consequence is simple. Ask for the test reports before the PO, not after the shipment. A supplier who can't produce a nickel release report on request is giving you information about the next three years, and it's worth having that information early.
The counterintuitive part
When the launch date is tight and the budget is already stretched, the instinct is to cut the pre-production sample round or the third-party testing to buy back two weeks. I did that exactly once. The two weeks came back as five, because 800 pieces arrived with a temple angle that matched nothing, and we had no signed pre-production sample to hold up next to them.
The pre-production round on that style would have cost about $180 in samples and freight. The rework ran north of $2,600 — I want to say $2,800, but I'd have to check the credit note. That's the whole argument for prevention over correction in a single line: the cheap step is the one at the front.
Which scenario are you actually in?
Four questions, in order. The first one you answer 'no' to is your scenario.
- Do you have 8 to 12 weeks of your own sell-through data on this shape? If not, you're in A, whatever your annual volume is. Importing 4,000 frames a year doesn't help you if you've never sold a rectangle in your market.
- Is expected volume for this line above roughly 500 pieces a year, with a settled width and color mix? Below that, the MOQ math in B never works — you pay the consistency premium for a SKU that doesn't earn it.
- Are you putting your own name, packaging, or a custom mold on the product? Any yes puts you in C, because you've taken on label liability and tooling ownership whether you planned to or not.
- Can your supplier produce a golden sample and batch measurements on request? If not, stay in A's buying pattern — small, frequent, stock-based — regardless of what your volume says. Volume without consistency just means buying the same mistake in larger quantities.
The awkward overlap is the one most buyers actually occupy: you've placed one order, it sold acceptably, and now you're placing order two. That's technically B, but buy it like A. Keep the size spread wide for one more cycle, avoid a 12-month MOQ, and only then move to the consistency-first model. I learned that one the expensive way — I locked a 300-piece MOQ on my second order because the first sold out, then watched two widths sit for seven months.
What doesn't change
Different budgets, different MOQs, different paperwork. Same failure mode in all three: the problem was visible when the carton was opened and nobody measured it.
Our current intake check has twelve points — the three measurements, color against the golden sample, hinge tension, temple tip fit, lens groove depth, screw torque, nose pad attachment, logo placement, and the paperwork. About twenty minutes per shipment once you've done it a few times.
Last year it flagged something on three of fourteen incoming shipments. Two were minor enough to accept with a note on the file. One went back. That single return paid for the checking time on the other thirteen several times over.
You can argue about which scenario you're in — that part is a genuine judgment call and reasonable people land differently. Measuring the box before it goes on a shelf isn't.