8:40 a.m., a factory conference room, February 2023

I still remember standing there with a pen in my hand and a coffee I'd forgotten about. Twelve trays of lip balm samples were lined up across the table. Every unit labeled SPF 30. Every tube printed cleanly, caps seated tight, product level consistent from tray to tray.

They looked perfect. That's exactly what bothered me.

Six years of reviewing cosmetic deliveries will do that to you. The more polished a batch looks, the harder I look at the parts that aren't visual.

How we got here

In December 2022, a mid-market retailer asked us to develop a tinted SPF lip balm line for their 2023 sun season. Launch date: May 1. Not negotiable. They wanted the product positioned against established names like Aquaphor Lip Repair Lip Balm SPF 30 and Essence Pout Protector SPF 20 Lip Balm — same general category of tinted, protective lip care — but with their own shelf identity. We were the sourcing and quality partner behind it. Our name went on nothing.

We quoted three contract manufacturers. We chose the second-fastest to develop, the middle quote. First production run: 2,500 units. If the test batch cleared, we'd release 50,000 more before summer.

Straightforward project on paper. Until it wasn't.

The first red flag was a verb tense

I spent about 20 minutes on the samples that morning. Tube printing: fine. Shade match: fine. Texture and glide: fine.

Then I read the carton. "SPF 15 Tinted Lip Balm."

Sorry — fifteen?

The brand's brief said SPF 30. The approved sell sheet said SPF 30. My own spec notes said SPF 30. Somewhere between the brand deck and the factory's work order, the number had shifted by half.

The project manager — genuinely helpful, not defensive — told me the formula team had built the default sunscreen system to SPF 15. "We thought that was the SPF level you wanted. You said 'SPF lip balm.' You didn't say more."

We said "SPF lip balm." They heard "SPF — any level." And they picked the easiest one to test.

Somebody, somewhere, meant something else

Here's the part I've had to sit with since: this wasn't a quality failure on their end. It was a communication failure on ours.

The brand's original brief said "tinted lip balm with SPF" — clean marketing language, completely useless from a formulation standpoint. A formulator needs a number, a test protocol, and a reference standard. "With SPF" gives them none of that. So the formula team reached for what they knew: their existing sunscreen base, already tested at SPF 15, ready to go, cheaper to produce.

And honestly? I understand the instinct. But under the FDA's OTC sunscreen monograph, a lip balm sold with an SPF claim in the U.S. is regulated as an over-the-counter drug. The label value must match the tested value. SPF 30 is SPF 30. There is no "close enough."

We had a product that physically worked and was legally mislabeled.

The second problem was quieter

Once I started pulling threads, the carton issue wasn't alone. Two of the four shades — the reds — had been formulated against the factory's own pigment stock, not the approved color reference we'd sent months earlier. Under the fluorescent ceiling lights in that conference room, the difference was basically invisible. I only caught it because I walked the tray over to the window and held a sample against natural daylight.

Not catastrophic. Not a safety issue. But it wasn't the shade the brand had signed off on. If it shipped, someone in the marketing department would notice the first time they opened a tube outdoors.

I sent the batch back

Not because any single issue was fatal on its own. Because all three together — wrong SPF level, wrong carton copy, off-spec pigment system — were the kind of thing that eats a launch.

I told the factory we'd redo it. They weren't thrilled. I wasn't either. But we both knew a redo at that stage was cheaper than a recall three months in.

The gap was ours to own

I want to be careful with how I say this, because it's easy to point at the factory and stop there. But the honest answer is we didn't have a formal specification confirmation process for lip balm fills. No signed document that said "this number means this, this shade reference is this physical chip, this test lab is this one." We ran on shared understanding and email threads. That works until it doesn't.

Looking back, I should have built the spec sheet into the RFQ from day one. I didn't, because "SPF lip balm" felt obvious at the time. It never was.

What we changed

The rework took about three weeks. The factory rebuilt the formula, re-ran the SPF test with an independent lab, and sent fresh samples. This time we had a signed spec sheet attached to every document in the chain — SPF level, test protocol, lab name, shade reference code, tube diameter, fill weight, carton copy. Every line a specific value. No adjectives.

Did it slow things down? Yes. Did it cost more? Also yes. Rework, re-testing, and a compressed timeline are not cheap. But the process we came out with is the one I now run on every SPF product we source.

Twelve checkpoints before any mold opens. Every one of them signed by both sides.

If you're sourcing SPF lip balm, do this

Write every number down. Not the number you meant — the number that goes on the label, the number that goes in the test report, and the number the formulator should hit. Those three are the same number. If they aren't, you have a problem, and you probably won't find it until you see the samples.

  • State the SPF level as an exact figure — no ranges, no "around," no "approximately"
  • Define shades against physical reference chips, not descriptions
  • Name the testing lab and the protocol before production starts
  • Put the carton copy into the spec sheet, verbatim — not just into the brand deck

A bit tedious. Much cheaper than three weeks of rework.

How it ended

The corrected batch landed in mid-April 2023, roughly ten days behind the original schedule. The retailer made their season window. Customers never saw the near-miss. But we knew. And the process came out sharper for it.

Since then I've rejected other batches. But never again because of a sentence somebody assumed the other side understood. That one, at least, is retired.