Products + Devices / Object Files
Decide what the category has to prove before the brand enters the room.
An object gets taken apart the same way a claim does — the eight passes — with two extra columns that only hardware needs: how it fails, and what a year of owning it actually costs. A good product does not need to be medically necessary. A beautiful product does not get to borrow clinical authority from the packaging.
I am exactly the person beautiful packaging works on. There is a small crow somewhere in my brain that still believes the heavier jar may know something. That is why pleasure gets its own column here instead of being smuggled into the evidence.
Last updated: September 1, 2026 · Live category, product and device files linked below
Start here
Shop the job before you shop the brand.
Framework
Product Category Intelligence
Define what the category owns, what changes performance, how it fails, what a year costs, and which part of the price is buying pleasure.
Evidence
Beauty Evidence Has Levels
Ingredient research, mechanism, finished-product testing and marketing translation are different evidence objects.
Review method
Product Reality Checks
Judge the job, proof, fit, burden and pleasure without forcing every purchase into clinical necessity or frivolous waste.
Use the review method → · Open the Product Evidence Receipt →
Buy by category
The buying guides are live.
Cleanser
Choose by removal load
What it has to remove, how skin feels afterward, and whether the formula makes the rest of the routine easier.
Moisturizer
Choose by barrier job
Dryness level, vehicle, routine fit and whether the finished formula earns its repair language.
Sunscreen
Choose from the protection label outward
Protection first; then cast, eye behavior, tint, makeup compatibility, reapplication and recurring cost.
Vitamin C
Do not let 20% win by default
Exact form, stability, packaging, finished-product evidence, tolerability and whether the bottle survives ownership.
Retinoid
Settle the molecule before the percentage
Tretinoin, adapalene, retinal and retinol do not share the same regulatory lane or evidence file.
Pigment
Name the pigment problem first
Brightening is not one job. Match the serum to PIH, melasma, sun-related pigment or another defined concern.
Devices
The hardware still has to survive ownership.
Light
LED Masks and Panels
Wavelength, output, dose, schedule, eye protection, exact-device evidence and price inflation.
Current
Microcurrent
Temporary tone is not a portable facelift. Gel, protocol, evidence, and how long the result survives after the device turns off.
Heat
At-Home Radiofrequency
Warmth is not a collagen contract. Device-specific evidence, temperature controls, use schedule and safety cutoffs.
Ownership
Device Maintenance Math
Hardware plus gels, charging, replacement parts and minutes multiplied by the schedule you will actually keep.
Products + ingredients
The live files are already on the desk.
Daily protection
Sunscreen You Will Actually Use
The label matters. So do film, finish, amount, reapplication and whether the bottle survives real life.
Barrier
Ceramides
Barrier architecture, not generic moisture language. Ratios and finished-formula context still matter.
Multi-tasker
Niacinamide
Useful at sane concentrations; less useful when every product already contains it and a separate serum duplicates the job.
Acid
Azelaic Acid
Clinical receipts exist. Concentration, vehicle and the actual job still decide whether a particular bottle earns its place.
Open the full Ingredient & Mechanism Glossary → for the broader index rather than turning this page into a second glossary.
How this desk reads a product
Where an object file differs from a claim file.
A product review here runs the same eight passes as everything else on this site. Then it adds the two questions only hardware and bottles force: how does this thing fail, and would you buy it again at full price after a year of charging it, cleaning it, replacing the gel, and remembering to use it. The passes decide whether the claim holds. Those two decide whether you would live with it.
Clinical value, functional value and pleasure are kept in separate columns on purpose. A product can be a delight and a poor treatment. It can be evidence-backed and completely wrong for your face. Collapsing those into a single score is how shopping guides stop being useful.
A product should enter the site through a question, not a shopping slot.
Bring the exact claim — not the category — and let the desk separate signal from theater.
The Register of Unproven Claims
Claims on the box, examined
The box is allowed to be beautiful. These are the sentences on it that have been checked against the published evidence, and the dates that checking happened.
-
Daily sunscreen prevents visible skin ageing
Established
Randomised, in people, over four and a half years, measured on skin rather than in a survey. It holds.
-
Beef tallow as a skincare base
No adequate source found
Claims for acne, eczema and psoriasis are common on social video and largely uncited; a 2025 analysis found most promoting posts showed financial bias.
-
Polynucleotide injections regenerate the skin
Thinly supported
Nine studies, all rated low or moderate quality, covering 219 patients between them. Real signal, and nowhere near enough of it to conclude.
-
Topical NAD+ and NMN serums
Untested as claimed
No controlled human trial of topical NAD+ or NMN with skin endpoints has been published. The formulation literature exists because the molecule penetrates poorly.