Who this checklist is for (and why I built it)

This is for anyone handling liquid foundation sourcing for a brand, an OEM order, a distributor restock, or a private-label line. If you're also sourcing a liquid foundation brush to pair with the formula, it applies to you too.

I started managing beauty sourcing orders in 2018. Over the first three years, our team wasted roughly $8,400 on rework, dead stock, and rush freight—all on liquid foundation orders. All of it was avoidable.

This checklist is what that money bought me. Seven steps. Order matters—skipping step 1 makes step 3 pointless.

1. Lock the end-user scene before you talk formulas

Most buyers start with the formula. I did too. Backwards.

A liquid foundation for mature skin doesn't get better because a supplier put "mature" on the label. It gets better because you specified the behavior: finer pigment dispersion, more emollient slip, better spreadability to avoid settling into fine lines, and typically a lower alcohol content than the general-population version.

Write these into the RFQ before you send it:

  • End-user profile (age band + typical skin condition)
  • Coverage preference (sheer, medium, buildable)
  • Finish (matte, satin, demi-matte)
  • Spreadability target: does it need a "clinging" finish or a "gliding" finish
  • The tool most users will pair it with—brush, sponge, or fingers

"Clinging" versus "gliding" matters for how the liquid foundation pairs with a brush. That one word changes the brush spec.

Where this falls apart: buyers write "mature-skin foundation" on the spec sheet and let the factory guess. The factory guesses in the direction of whatever's cheapest in their current mix. It rarely matches the brand promise.

We ate this in 2023. A 6,000-unit run, complaint category: creasing. Repackaging costs came straight out of margin.

2. Write the spec as a contract clause, not a wish list

A spec that can actually be manufactured against looks like this:

  • Viscosity—measured with the same instrument (e.g., Brookfield RVT, spindle 6, 24h and 72h after fill)
  • Specific gravity
  • Coverage grade—stated numerically, e.g., "grade 3 on a 1–5 scale tied to our reference"
  • Color match (ΔE stated numerically, under a named illuminant—D65 recommended)
  • pH range
  • Shelf life (unopened) and period after opening
  • Pump/dropper delivery consistency

Viscosity is the most frequently skipped. It's also the thing that flips during scale-up.

September 2022: sample read 3,800 cps. Production read 4,700. Visually similar—but on application the liquid foundation started dragging mid-spread. Reblend, retest, expedited ship. Damage: $3,800 and 11 days late.

Here's the part that still gets me. The most frustrating part of liquid foundation sourcing is that the same viscosity drift keeps showing up across vendors who all "fixed" it after the last run. You'd think written specs would prevent it. They don't—unless the spec is a number, not a word.

"Silky." "Smooth." "Long-wear." Those are words. Write numbers.

3. Test the liquid foundation and the brush as one system

This is the step most brands skip. Not the obvious skip. The one that feels like it's already covered but isn't.

Test data on the liquid alone tells you how the liquid behaves on a back-of-hand swatch. It says nothing about how it behaves when a brush for liquid foundation is doing the work—and that's how most end users apply liquid foundation makeup.

The test I run now, on every sample round:

  • Apply the same liquid foundation with three different brushes
  • Apply the same brush to three different batches
  • Check brush pickup—how much product stays on the brush rather than transferring to skin
  • Run the brush through a wash cycle and re-test

We ran this for the first time in Q1 2024. Discovery: our preferred liquid foundation brush left visible streaks on the oil-containing formula. Not on all formulas—on the one we'd already greenlit. 2,400 brush heads rejected. About $1,600 gone. (Ugh.)

Not ideal. But cheaper than returns in month four.

4. Clear compliance at sample stage, not at packaging sign-off

This one gates your launch date directly.

Per FTC advertising guidance (ftc.gov/business-guidance/advertising-marketing), beauty claims need to be:

  • Truthful and not misleading
  • Substantiated with evidence
  • Clear about endorsements and testimonials

In practical terms: if "16-hour wear" is going on the liquid foundation carton, the test data backing that claim needs to be in the supplier file at sampling—not pulled together after packaging is printed. Anything resembling anti-aging or clinical claims has to be dropped unless there's real substantiation behind it.

Fall 2023, we came within one signature of approving inner-box artwork. Compliance caught us. The wear claim we'd wanted couldn't be substantiated by the supplier. Dodged a bullet. One click away from a full reprint and a three-week launch delay.

My advice, and I've repeated it inside our team too many times: finalize the claim language before sampling. Factories will print what you ask them to print. Legal will then ask you where the evidence is.

5. Sample across the color grade, not just the lightest shade

Deeper shades behave differently. More pigment load, different rheology, sometimes a heavier skin feel.

In our own QC log, about 70% of the defects we've caught landed in the middle and darker half of the color grade, not the top. The light shade is usually the one that got polished first—because that's the one the factory shows you.

What to check every batch:

  • At least three random units per shade
  • Shade match under D65 and A illuminants
  • A spread test on the deepest shades—do they behave noticeably more powdery or lifting than the lighter shades in the same line?

One supplier in 2023 had beautiful light shades and shade 4 that separated at 34°C. We swapped the whole grade to a different vendor. Simple as that.

6. Price on landed cost, not FOB

People think expensive suppliers deliver better quality. The reality runs the other way—suppliers who can deliver consistent quality earn the right to charge more. The causation reverses. Looking at FOB only hides this.

The math I use:

Landed cost = (FOB unit price × quantity) + freight + insurance + duty + inland transport + rush/storage surcharges + (expected defect rate × unit price)

If you're doing domestic fulfillment in the U.S., factor in postal rates. First-Class Mail letters cost $0.73/oz as of January 2025 (source: usps.com/stamps). That line item is smaller than freight, but it's real, and it's not free. Prices re-index annually—check current figures; they've already moved once this year.

I ran this once comparing two suppliers. Supplier A quoted FOB $2.40. Supplier B quoted $2.75. Landed cost, including B's lower defect rate at scale, put B ahead. Don't hold me to the exact dollar split—the point is that A looked cheaper and wasn't.

7. Run a 1.5x fill pressure test before shipment

Low cost, high value, skipped constantly.

Fill 30–50 finished units from the final batch to 150% of nominal fill volume. Run a shipping simulation: 4 hours on a vibration table plus standard drop.

Watch for:

  • Pump delivery at 1.5x load
  • Dropper overflow at 1.5x load
  • Inner box staining from product migration (more common than it sounds)
  • Cap closure integrity

When we ran this last, 12 of 500 pumps leaked at 1.5x. 2.4% isn't terrible—until it's your shelf.

Common mistakes, so you don't repeat mine

  • Quoting before specifying the end-user scene. The factory optimizes for convenience, not fit.
  • Writing specs with adjectives. Silky, smooth, long-wear—those are wishes, not clauses.
  • Testing the liquid and not the brush. The end user judges both. The liquid foundation brush is part of the product.
  • Leaving claim language until after sampling. Compliance is a sampling-stage constraint, not a last-minute checkbox.
  • Approving the lightest shade and shipping. The trouble usually sits one and two grades down.
  • Bidding FOB-only. Landed cost is the only number that reflects what you'll actually spend.
  • Skipping the 1.5x fill test. Cheapest step in the whole checklist.

None of this is new. The best sourcing teams have been doing versions of this for years—just more formally. What's changed is the tooling: digital viscometers and colorimeters are now affordable at small-factory level, and suppliers are more willing to share data than they were in 2020 (that was a low point). Fundamentals haven't moved. Execution has—and that's the part that keeps improving.

If you remember one thing: work backwards from the end-user scene. Everything else is detail.