Five kinds of number we would not buy blind

Every product on this page is one we would recommend. Two of them we already have, in other posts, in positive terms.

That is the point. The interesting failures in skincare marketing are not on bad products from bad brands. They are on decent products from serious brands, and they follow a small number of repeating shapes. Once you can see the shape, you can read a page you have never seen before in about a minute and know which parts to keep.

So this is not a list of things to avoid. It is five shapes of number, with the product where we found each one.

One: a figure measured at the moment it flatters most

Where we found it: Torriden DIVE IN Toner

The page prints an improvement in skin smoothness and a reduction in sebum, both to two decimal places, both footnoted with a subject count. That footnote is more disclosure than most brands offer and the brand deserves credit for it.

Read the rest of the footnote. Both figures are average results immediately after use.

A watery toner floods the surface with water and glycols. For as long as that layer is wet, it swells the outer cells and fills in surface relief, so a smoothness reading taken at that moment records a hydrated surface rather than refined texture. Nothing separates the two. The sebum figure has the same problem from the other direction: a reading taken minutes after flooding the face with water records displaced and diluted surface oil, not a change in how much oil the skin makes. The formula carries nothing that suppresses or absorbs sebum. No niacinamide, no zinc, no silica, no clay.

The measurement happened. What it measured is not what the number is being used to say.

The tell: immediately after use, at first application, right after a single use. Whenever a timing phrase sits next to a percentage, read the timing first.

Two: a scoping footnote you cannot see

Where we found it: Aestura Atobarrier 365 Cream

The visible page says the capsule technology keeps ceramide sitting in the gaps of the barrier long enough to help it recover, and pairs that sentence with a figure for how much more remains after washing.

Follow the figure into the structured data behind the page and it is scoped to a raw material characteristic, confirmed by a test run outside skin. The block that carries that qualification never mentions barrier gaps or recovery at all.

So the correction that would let you discount the number, from a result on a face to a property of an ingredient in a dish, is not fine print under the claim. It is absent from the page you actually read.

The same page counts its own safety testing two different ways depending on where you look, listing three tests in the visible copy and eight in the data underneath. Neither of these changes what the cream is, and we still think it is the best barrier cream we have read. It changes how much the front of the jar is worth.

The tell: a mechanism sentence and a number placed side by side, where the number's own definition is somewhere else.

Three: a disclosed concentration that turns out to be a rounding error

Where we found it: Wellage Real Hyaluronic Soothing Cream

This one is unusual because the brand does disclose. Six hyaluronate species with figures attached, which almost nobody in the category does, plus a purity claim of one hundred percent.

Then you read the figures. They run from one part per million down to two ten thousandths of a part per billion. Products built around hyaluronic acid as a working ingredient generally use it somewhere between a tenth of a percent and two percent. The disclosed levels here are several orders of magnitude below that.

Meanwhile the purity number describes the raw material that went into the vat, not how much of it is in the tube. Purity and dose are different questions and only one of them is being answered.

None of this makes it a poor cream. It is a fragrance-free light gel cream with a wrinkle improvement registration, and glycerin and butylene glycol near the front are doing the hydrating. The hyaluronic acid on the front of the box is not.

The tell: a purity figure, or a form count, standing in for a concentration. If a brand gives you a number, check which question it answers.

Four: a claim its own ingredient list argues with

Where we found it: TONYMOLY Ceramide Mochi Toner

The marketing block on this product's page names four key ingredients, and hyaluronic acid is one of them. The statutory ingredient list on the same page, for both sizes, contains no hyaluronate of any spelling at any of its thirty four positions.

We checked that three ways, searching the ingredient block for the Korean forms of hyaluronic acid, hyaluronate and hyaluronic. Zero hits inside the list. The word appears exactly once on the whole page, in the marketing zone that makes the claim.

The likeliest explanation is not deception. Two other generations of this product do carry a hyaluronate, so this reads as copy written for an earlier formula and never updated. The effect on a shopper is the same either way: someone choosing this toner for hyaluronic acid gets none.

The rest of the page has a second version of the same problem. It describes the scent as a subtle citrus. Lemon peel oil is genuinely on the list, and so are lavandin, scented geranium, mexican juniper and lavender, all four printed above the lemon. A fragrance reactive buyer screening for citrus would miss an entire lavender family aroma system.

The tell: a claim in the marketing block that the statutory ingredient list does not carry. The list is the legally binding half of the page, and it is the cheaper thing to check.

Five: a percentage with nothing to compare it to

Where we found it: Anua PDRN Hyaluronic Acid 100 Moisturizing Cream

Three figures on this page: a collagen improvement percentage, a plumping percentage and a hydration window measured in days.

For each one, ask the question that makes a percentage mean something. Improved compared to what? Untreated skin, a previous formula, a competitor, and a vehicle without the active are four different claims with four different meanings, and none of them is named. No method, no panel, no instrument and no institution accompanies any of the three.

Underneath the numbers is a competent lightweight moisturizer with salmon derived sodium DNA, a hyaluronate complex, a light silicone base and no synthetic fragrance. It absorbs fast, which is the thing users report most consistently and the thing the base is built to do. That is the honest reason to buy it, and it is available without any of the three figures.

The tell: a comparative word with no comparator. Up to, more, faster, better. If the sentence does not say what it beat, it did not say anything.

What we do trust

It is worth naming the other side, because the pattern is just as consistent.

Compositional absences. No fragrance, no parabens, no organic filters. These are settled by the ingredient list itself, which is why they are the most reliable claims in the category.

Regulatory designations. A Korean functional cosmetic registration for wrinkle improvement or barrier recovery means a submission cleared review. It does not tell you the effect size, and it is still a checkable fact rather than an adjective.

Disclosed concentrations, in either direction. A brand printing five parts per million is telling you something true and useful, even when the number is small. Especially when the number is small.

Usage instructions. Brands overreach in marketing and are usually accurate about how to use their own product, because that is the part they actually tested.

Common questions

Does a bad number mean a bad product?

No, and treating it that way would throw out most of the shelf. A brand exaggerating a figure does not change the formula underneath it. What it changes is which reason you buy the product for, and that is worth getting right.

Why do so many of these numbers have two decimal places?

Precision reads as rigor. A figure like 68.3 implies an instrument and a protocol even when neither is named, whereas a round number invites a question. The decimals are usually the least informative part of the claim.

What is the fastest single check?

Look for the comparator. Most failing numbers on a skincare page fail because they never say what they improved on. That one question takes five seconds and it removes more claims than any other.

Are Korean brands worse about this than others?

Not in our reading, and the five pages above are all Korean because that is the shelf we read most closely rather than because the problem is local. The same five shapes turn up on Western and Japanese pages too. What Korean labelling does more often than either is print an actual concentration, which is the thing that makes checking possible in the first place.

If a claim on something you own does not sit right and you cannot tell why, send us the wording and we will look for the method behind it.