I'm not a cosmetic chemist. I'm the person who fills out the purchase order.

Officially: office administrator for a 42-person private-label beauty company. I manage all contract-manufacturing and packaging orders — roughly $370,000 a year across 11 vendors — and I report to both operations and finance. Somewhere along the way, powder foundation sourcing became my problem to solve.

So this isn't a formulation guide. If you want to argue about binder-to-pigment ratios, talk to a chemist. This is the procurement checklist I actually run before I commit to a new supplier or a new format — six steps, about an afternoon of real work if your spec is ready, plus however long sampling takes.

I learned most of it the expensive way. Which is the other reason to write it down.

Step 1: Write the spec before you talk to anyone

Pressed, loose, mineral, cream-to-powder. These are not interchangeable, and a supplier who is genuinely excellent at one of them is not automatically good at the rest. I've watched people assume otherwise and lose a quarter to it.

Pressed powder has to hold together in a pan and survive shipping. Loose powder lives or dies on the sifter and the milling. Cream-to-powder has to actually set on a face, not just in a lab dish. Mineral formats carry their own sourcing questions around the pigment and filler supply chain. I don't run those lines. But I've learned that if I don't specify upfront, I get quoted on whatever's easiest for the supplier to make.

Before the first call, get this on one page:

  • Format and sub-format
  • Coverage level (sheer, medium, full) and finish (matte, satin, luminous)
  • Target skin profile — including whether you're building a powder foundation for mature skin range or a general-market one
  • Shade count and the approximate spread
  • Packaging: pan size, compact type, sifter type
  • Whether a powder foundation brush ships with it

Checkpoint: if you can't fill that page in from memory, you're not ready to talk to suppliers. You'll get sold, not sourced.

Step 2: Get the shade reference system in writing

"We match your shade" means nothing until you know what they're matching against. Physical chips? Spectrophotometer readings? A proprietary numbered system you'll have to translate later?

"Industry standard color tolerance is Delta E < 2 for brand-critical colors. Delta E of 2–4 is noticeable to trained observers; above 4 is visible to most people." — Pantone Color Matching System guidelines

Ask what tolerance they work to, and get it into the quote as a number, not a promise. Then ask what happens when a batch falls outside it.

We went four rounds on a pressed shade in 2024 — three, actually, the fourth was a re-press after a cracking issue — before anyone realized the supplier was matching against a printed Pantone chip that had been sitting under a desk lamp for two years. Faded. Nobody's fault, exactly. Expensive anyway.

Step 3: Sample in finished packaging — and test it with the brush

This is the step people skip. Cheaper. Faster. Useless.

Lab bulk in a jar tells you about texture and nothing else. It doesn't tell you whether the sifter dispenses evenly, whether the pan cracks in transit, whether the compact hinge survives a handbag, or whether a cream-to-powder actually sets after two hours on an actual face. For loose powder especially, the delivery mechanism is the product.

Order retail-ready units. Then test them the way a customer would.

Which brings up the brush, which almost nobody plans for. Pressed and loose formats want different brush density. A tightly packed brush that works beautifully on pressed powder will grab a clump out of a loose sifter and dump it on someone's cheek. If a powder foundation brush is part of your kit or your recommendation, test the finished powder with that brush. Not a generic one from the office drawer.

Cost us a full shade relaunch once. We tested with the wrong brush and shipped a loose powder that looked patchy in the field.

Step 4: Ask the mature-skin question like a buyer, not a chemist

I'm not a formulator, so I can't evaluate whether a specific emollient system reduces the appearance of fine lines on textured skin. That's genuinely outside what I can judge, and I stopped pretending otherwise after a supplier walked me through a technical deck I had no business questioning.

What I can do is ask for evidence that's actually checkable:

  • Wear-test photos on the real demographic, at 4 and 8 hours — not on a 24-year-old model
  • Panel size and how the panel was recruited
  • Whether the testing was on textured or mature skin specifically
  • How the product behaves over a moisturizer or primer, which is where mature-skin cream to powder foundation usually falls apart

And here's the flag: if the answer to "how does it perform on mature skin" is "it's anti-aging," that's a marketing answer, not a formulation answer. Per FTC business guidance on advertising, claims have to be truthful, not misleading, and substantiated with evidence. "Anti-aging" is a claim somebody has to defend with a file. Ask to see the file.

Step 5: Run the paperwork check before the sample even lands

Do this in parallel with sampling. It's the cheapest filter you have, and it kills bad suppliers before they cost you anything but an email.

Ask for: certificate of analysis per batch, full INCI ingredient list, allergen statement, safety data sheet, batch and lot traceability, stability testing summary. If it's a mineral powder foundation, ask specifically about the pigment and mineral supply chain and whether they can document origin.

Then ask the boring question nobody asks: can you issue a proper invoice?

In 2023 I found a mineral powder house quoting about 18% under our regular supplier. I ordered 2,000 units. They sent a handwritten receipt. Finance rejected the expense report outright, and I ate roughly $1,900 out of the department budget to make it go away. Now I verify invoicing capability before I verify anything else. It's a five-minute question and it has saved me four separate times since.

One more thing on documentation, because it comes up constantly: FDA does not approve cosmetics or their ingredients before they go to market. Color additives are the exception. So when a supplier leads with "FDA-approved formula," that tells you something — either they don't know the rules or they're counting on you not knowing them either. Either way, that's your answer.

Step 6: Lock MOQ, lead time, and the failure plan

Quotes are the easy part. Ask these before you sign:

  1. MOQ per shade, not per order. Six shades at high MOQ each is a different commitment than one SKU.
  2. First-order lead time versus reorder lead time. These are usually very different numbers.
  3. If a batch is off-spec, who pays for the re-run, and who pays for the delay?
  4. What happens if a shade fails wear testing after production?
  5. Do they hold inventory or make to order — and does that change in Q4?

The vendor who told me "we're not the right fit for that format — here's who is" earned my trust for everything they did take on. I don't need a supplier who does everything. I need one who tells me the truth, because I'm the one who has to explain the result to operations and finance. A specialist who knows their limits is worth more to me than a generalist who overpromises.

Things that go wrong anyway

None of the above makes you immune. These are the ones I still catch myself on:

  • Accepting a quote without a written spec. If it isn't on paper, it isn't the deal. It's a mood.
  • Treating lab samples as production samples. They are not the same object. Different equipment, different batch sizes, different failure modes.
  • Letting the supplier write your label claims unattended. They write their words. You carry the liability.
  • Testing on one face. One person's skin tells you about one person's skin.
  • Skipping the invoicing check because you're excited about the product. This is the one that costs you money out of your own budget.

I can only speak to our situation — a small operation with a predictable shade range and 6 to 10 SKUs live at a time. If you're launching 40 shades across two formats simultaneously, the sequencing here won't hold up and you'll need a different plan. But the six steps themselves have held up across every format we've added, including the ones we eventually dropped. Which is its own kind of useful.