Procurement lead handling cosmetic-grade HA orders for 8 years. I've personally made (and documented) 12 significant sourcing mistakes, totaling roughly $18,600 in wasted budget. Now I maintain our team's HA intake checklist so nobody here repeats my errors.

If you're sourcing hyaluronic acid for a moisturizer line, an eye cream SKU, or a serum launch—this checklist is for you. Seven steps. Each one has a checkpoint. I'll flag the ones I learned the expensive way.

Step 1: Lock the molecular weight range before you contact anyone

My first big mistake, back in 2019. I ordered 1.8M Da HA for a serum line because "higher molecular weight = better, right?" What I got was a film-forming, tacky raw material that wouldn't behave in our water-light base. 300 kg. $4,200 down the drain.

What I learned after: in a hyaluronic acid moisturizer, high MW is fine—it sits on the surface and forms a nice film. In a serum or a hyaluronic acid eye cream, you usually want lower MW for penetration, or it just sits there. Same raw material, opposite needs.

Checkpoint: Put the range on the RFQ. Not "cosmetic-grade HA sodium salt." Write "HA sodium salt, MW range 800k–1.2M Da, cosmetic grade, COA required per batch."

Step 2: Make the supplier document "cosmetic grade"

Verbal assurance doesn't count. You need a COA with every batch—not later, not "when we can find it."

I don't have hard data on industry-wide fake-documentation rates, but based on our 5 years of receiving records, my sense is roughly 1 in 5 first deliveries has some mismatch between the COA and what actually shows up. High enough to check every time, low enough that suppliers think you're paranoid—until they don't.

Checkpoint: ISO 22716, GMP (if they ship to EU or US), COA with batch number, MSDS. If a supplier hesitates when you ask for the COA alongside the shipment, that's a signal.

Step 3: Ask for application samples in your actual finished format

We paid twice to learn this one.

A sample tested fine in the supplier's generic water base but failed when we ran it through our own cream base. Then the next sample—this one for a face serum—looked perfect in the lab but separated inside our eye cream tube within a month.

Never expected the biggest gap between "sample approved" and "production rejected" to come down to which base the sample sat in. Turns out HA behaves noticeably differently between a cream and a water-light serum. Generic test samples are a cost transfer—you pay at production.

Checkpoint: Request samples suspended in the base that matches your actual finished format. If the supplier can't tell you what their sample base was, push back.

Step 4: Test compatibility in your formula, not in isolation

This is where a lot of people get burned.

Hyaluronic acid and retinol: these two can fight in an emulsion if you get the order of addition wrong. Retinol wants to be dissolved in the oil phase first, then HA introduced. If you dump HA into a formed retinol system, you'll get clumping and pH drift more often than not.

Niacinamide vs hyaluronic acid: usually fine. Both are water-soluble. But at high concentrations, combined viscosity can spike. We once tried 2% niacinamide with 1.5% HA in a serum. It poured like honey.

Preservative systems: some HA raw materials carry residual ethanol or reaction intermediates that interact with certain preservatives. Test in your full system, not in distilled water.

Checkpoint: 8-week stability test at least, in your actual formula, at use concentration. Powder specs don't equal formula specs.

Step 5: Confirm batch consistency across three production lots

One supplier's first shipment was excellent. Second lot, slight viscosity drift. Third lot, foreign particles. Not malicious—just inconsistent process control.

We've caught 47 potential errors using this checklist in the past 18 months. Most of them stalled right here.

Checkpoint: Buy three separate lots. Test each. Molecular weight by GPC, viscosity, pH, microbial limits. If all three pass, you have real data. One lot is not data.

Step 6: Lock commercial terms in writing, with numbers

Everything in writing. MOQ, lead time, payment terms, penalties for delay.

This step's pricing reference is accurate as of Q1 2025. Raw material pricing moves—verify current quotes before you budget.

Checkpoint: MOQ aligned with your actual annual volume (a lot of suppliers set MOQs for small brands that don't make sense). Lead-time penalties in the contract. Payment terms tied to verified production performance, not sales promises.

Step 7: Find out what the supplier won't do, before you need it

The best supplier we've worked with told us on the first call: "We don't do custom MW fractionation. If that's what you need, here are two labs that do."

They got every order after that.

I've come to think that a supplier who says "that's not our strength" is more trustworthy than one who says yes to everything. Say-yes-to-everything almost always means do-everything-half-well. You want a partner who'll point you elsewhere when the fit is wrong—because the alternative is a partner who'll keep saying yes while missing your specs. Granted, if the answer is no on your core category, walk away.

Checkpoint: Ask them directly: "What do you not do?" If they pause for more than ten seconds, dig. Clear boundaries usually come with clear delivery standards.

Common mistakes that aren't in the seven steps

  • Choosing a supplier on price alone. I did. The cheapest one had the slowest lead time, the worst COAs, and cost us more in rework.
  • Skipping retainer samples. Even after Step 4, pull another sample during the first month of production. Formulas drift.
  • Treating the sales rep as your technical contact. Get the technical manager's direct line before you need to ask a molecular weight question at 4 p.m. on a Friday.
  • Agreeing to MOQ on a call. Never say yes to a quantity before you see the written quote.

Granted, about a third of these were only obvious in hindsight. A checklist wouldn't have saved me—it would've just made the mistake cheaper. Which is the whole point. So here it is.