01
The wrong question
Pancake begins with “which foundation?”. That question presumes a full-face layer before anyone has asked what the face actually needs covered. Almost every cake outcome is a diagnosis failure, not a product failure.
Insights
Cake is a
structural failure
02
Layer arithmetic
Primer, base, concealer, powder, bronzer, setting spray is six films. Each edge is a place where the stack can lift, crease or go grey. Cutting one layer improves the finish more reliably than upgrading any single product in it.
03
Water leaves the film
Dry rooms, altitude and dehydrated skin pull water out of a base from underneath. Opaque formulas have the least flex, so they crack first — and powder over them accelerates it.
04
Sebum travels
Oil moves the base sideways. If you have ruled out midday intervention, the only honest answer is a thinner film plus blotting — blotting removes oil without adding a layer, which is the one free move in the system.
05
Placement reads as skin
The eye reads placed coverage as flawless skin and reads an even full layer as makeup. This is why strategic concealer outperforms full base at the same coverage appetite.
06
Desire is not the problem
Wanting makeup is not what causes cake. Spending that appetite on opacity does. Spend it on cream colour, brow structure, lash and lip instead and the appetite is satisfied without thickness.
07
Match before layers
On deep, olive, red-leaning and deep-neutral bands, a short shade range is how cake starts: people add product to force a mismatch. Widen the map first. Opacity never fixes undertone.
08
Claims are not architecture
SPF, 'treatment', peptides and barrier language on a foundation do not thin the film. If the active is not named and dosed, you are buying mood. Architecture still decides whether it cakes.
Transparent scoring
Pancake risk starts at 30.
Finish is 100 − risk.
Every variable below is live in the edit. Positive deltas raise pancake risk and lower the skin-like finish score. Negative deltas do the opposite. No black box — if a number moves, you can name the weight.
Coverage appetite
(coverage − 30) × 0.45
Raises risk when
You ask for more than light evening (coverage over ~30).
Finish effect
High coverage is the largest single drag on a skin-like finish.
Layer count
max(0, layers − 3) × 7 − max(0, 3 − layers) × 5
Raises risk when
More than three opacity films on the face.
Finish effect
Each extra edge is a place the finish can lift, crease or go grey.
Dehydration stress
dryStress × (coverage > 45 ? 1.2 : 0.55); dryStress = skin dry 10 / normal 3 + climate altitude 12 / dry 8
Raises risk when
Dry skin or thin air, especially under medium-plus coverage.
Finish effect
Water leaving the film is what turns opacity into cake.
Sebum movement
oilStress × (maintenance ≥ 2 ? 0.5 : 1); oilStress = oily 12 / combination 7
Raises risk when
Oily or combination skin with no planned touch-up.
Finish effect
Oil moves the base sideways; thin films and blotting protect finish.
Maintenance tolerance
maintenance ≥ 2 → −6; else 8 − maintenance × 4
Raises risk when
You refuse midday intervention (maintenance 0–1).
Finish effect
A slightly richer finish only survives if you will intervene.
Time on the face
≤5 min → −8; ≥25 min → +7; else 0
Raises risk when
Long routines (≥25 min) that tempt undiagnosed layers.
Finish effect
Minutes are a complexity budget, not a virtue.
Desire
desire ≥ 3 → +9; desire 0 → −10; else +2
Raises risk when
High ritual appetite spent on opacity instead of definition.
Finish effect
Desire is allowed — finish stays honest when appetite buys colour, not cake.
Anti-pancake / wear-less goals
−9 when escape-pancake or wear-less is selected
Raises risk when
—
Finish effect
Reorders scoring toward skin-like pathways first.
Reactivity
+5 when sensitivity ≥ 2
Raises risk when
Reactive skin plus heavy removal of thick films.
Finish effect
Fewer films protect both barrier and finish.
Texture visibility
+6 when texture or large pores is a concern
Raises risk when
Opaque product will settle into texture.
Finish effect
Diffusion and placement beat opacity on visible texture.
Outdoor exposure
+4 when outdoors ≥ 2
Raises risk when
Heat, sweat and UV shorten the honest life of a thick base.
Finish effect
Outdoor load pushes the kit toward thin SPF hybrids, not cake.
Mature / fine-line context
+5 when fine lines concern; +4 when mature-skin goal
Raises risk when
Powder-heavy stacks on textured, mature skin.
Finish effect
Emollient thin films read as skin; powder cakes in lines.
Deep depth + undertone pressure
+4 when depth ≥ 8; +3 when olive / red-leaning / deep-neutral undertone
Raises risk when
Short shade ranges force over-application to force a match.
Finish effect
Wrong undertone reads grey; people add layers to compensate. Architecture says match first, layers never.
Rosacea-prone architecture
+6 when rosacea goal or high sensitivity + redness concern
Raises risk when
Full-face opacity as the reflex answer to redness.
Finish effect
Corrector + sheer base outperforms a mask on reactive redness.
Performance / sweat load
+5 when performance-sport goal; +3 when outdoors ≥ 3 with shine goal
Raises risk when
Heavy base under sweat without a blotting plan.
Finish effect
Sport finish is thin, grippy where needed, and honest about reapplication.
HD / photography load
+4 when hd-photo or on-camera path goals emphasize event + texture
Raises risk when
Opacity under hard light; powder flashback; glitter highlight.
Finish effect
Lenses punish layers before they punish colour. Diffusion is the finish.
Type fit weights
From a neutral 50
Coverage match
−|type.coverage − profile.coverage| × 0.42
Closest delivery to your appetite wins.
Sebum behaviour
type.oil × 6 (oily) or × 3.5 (combination)
How the texture holds on sebum.
Dehydration behaviour
type.dry × 6 (altitude) or × 4.5 (dry)
Film flexibility against water loss.
Layer weight
−layerWeight × 7 if anti-pancake else × 4
Opacity load is always costly; costlier when escaping cake.
Upkeep demanded
−upkeep × 5 if maintenance = 0
No touch-ups means high-upkeep objects lose.
Time cost
−max(0, minutes − 2) × 5 if timeBudget ≤ 5
Slow steps drop out of short routines.
Wear-less / multi-use / goals
+8 to +22 on pathway-aligned types; −6 to −8 on conflicts
Goals reweight lanes, they do not invent miracles.
Filters & budget
mineral +7; silicone-averse −7 on silicone-leaning lanes; lean colour −3
Preferences are hard constraints when you set them.
Depth / undertone (base lane)
+6 when depth ≥ 7 and type is sheer-to-medium; −8 full coverage with depth ≥ 8 and no wide-range signal
Deep bands need placement and range, not more opacity.
Claim-literacy caution
−4 on full-foundation when SPF-hybrid goal is absent but treatment-hybrid language is the only lure
Do not buy thickness for a skincare claim.
Claim literacy
Named. Dosed. Tested.
When not to buy.
Makeup that borrows skincare language still has to earn the bag as architecture. These cards are education — never a safety ranking, medical clearance, or brand attack. If a claim cannot survive a follow-up question, it does not drive the purchase.
SPF on makeup
UV numbers only count when the film is continuous and reapplied. A tint is not a beach plan.
Treatment / actives
A percentage on a box is not a prescription. If the active is not named and dosed, it is marketing perfume.
Hybrid makeup-skincare
Hybrid means one object does two jobs — not that either job is maximal. Trade-offs are the point.
Barrier / calm claims
Calm language is not a diagnosis. Fragrance-free and short lists matter more than the word 'soothing'.
Peptides · retinoids · acids
If you already run a leave-on active at night, a foundation with a whisper of the same molecule is not stacking science — it is stacking irritation risk.
spf
Tinted mineral SPF 30–50 in a foundation or tint
A tinted SPF earns the bag when UV is a real job that day and you will lay a continuous film. It loses the bag the moment you use it as a sheer glow with no reapply plan.
Named
Active filters listed by INCI (e.g. zinc oxide, titanium dioxide) with a stated SPF number.
Dosed
SPF 30 is the floor for intentional UV days; SPF 50 only earns the bag if you can apply enough film.
Tested
Broad-spectrum testing on the finished formula, not a sister cream. Look for the number on the unit you buy.
When not to buy
When you will sheer it to a stain, blot half of it off by noon, or treat it as all-day beach cover without reapplication. Half a film is half the SPF.
spf
Chemical or hybrid SPF inside a skin tint
Hybrid SPF is honest when it replaces a second step you would skip. It is vanity when it is a second SPF on top of a full mineral fluid you already wear.
Named
Each UV filter named; 'broad spectrum' stated; PA or PPD if the market uses it.
Dosed
Same rule as any SPF — about a 1/4 teaspoon for the face, which most makeup applications never hit.
Tested
Photostability of the filter blend in that specific base, not a brochure claim.
When not to buy
Sensitive or fragrance-reactive days when the hybrid also carries scent; or when you already wear a dedicated SPF and the tint's SPF is just noise on the label.
spf
SPF setting powder / brush-on SPF
Brush-on SPF earns a pocket slot for reapplication over makeup. It loses the bag the day you cancel real SPF because the powder 'has SPF too'.
Named
Active filter and SPF number on the powder itself.
Dosed
Powders rarely deposit a continuous film. Treat stated SPF as a top-up signal, never the primary strategy.
Tested
In-vivo SPF on powder application method — rare; assume the number is optimistic unless the brand publishes method.
When not to buy
As your only UV product on a high-UV day. Powder SPF is a midday assist, not architecture.
treatment
'Treatment' foundation with niacinamide
Named and mid-list niacinamide is a nice-to-have. Unnamed 'brightening treatment' is a mood board. Buy the texture; do not pay prestige for a ghost percentage.
Named
Niacinamide on the INCI, ideally with a stated percentage.
Dosed
Skincare efficacy bands often cite ~2–5%. A foundation that lists it last is not a treatment dose.
Tested
Clinicals on the finished makeup film are almost never published. Borrow evidence only from the brand's leave-on data if concentrations match.
When not to buy
When you already use a properly dosed niacinamide serum and the foundation is thicker than you need. You are buying film, not more science.
treatment
Hyaluronic acid / 'hydrating treatment' base
Hydrating bases earn the bag when the film stays flexible. They lose it when 'hyaluronic' is the only argument for a base that still cakes by 3pm.
Named
Specific HA weights or sodium hyaluronate on the list — not just 'hydrating complex'.
Dosed
HA needs water and a humectant context. In a powdery or high-alcohol film it can feel tight, not plump.
Tested
TEWL or hydration measurements on the finished product, ideally 8+ hours — almost never shown for makeup.
When not to buy
On already-dehydrated altitude skin without a water-based prep underneath. HA in a dry film is not a drink of water.
actives
Peptide or 'firming' makeup
Peptides in foundation are a dinner-party claim. Pay for the finish and the shade map; do not tip for a molecule that never made it to a dose.
Named
Specific peptide names (e.g. matrixyl-class) — not 'peptide complex' as poetry.
Dosed
Effective leave-on peptide products state use level; makeup almost never does. Assume sub-therapeutic.
Tested
Independent firming data on a makeup film does not exist in consumer reality. Treat as texture marketing.
When not to buy
When the price jump is justified only by peptide language and the shade range or wear fails your actual face.
actives
Retinol or bakuchiol in a tint or primer
Retinoid makeup is a vice dressed as vanity. If you want the active, put it in the night routine where dose and recovery are honest.
Named
Retinol / retinal / bakuchiol named with a percentage.
Dosed
Even 0.1% retinol in a leave-on is a commitment. In makeup, contact time and occlusion change the risk profile.
Tested
Stability of retinol in a pigmented emulsion is hard; demand packaging that blocks light and air if you care.
When not to buy
On reactive, rosacea-prone, or pregnancy-caution days; or when you already run a night retinoid. Doubling actives is not sophistication.
actives
Salicylic / 'blemish treatment' concealer or powder
A dosed BHA concealer can earn a spot on a breakout day. A 'clarifying' powder with no percentage is just powder with better copy.
Named
Salicylic acid (BHA) with a percentage — OTC acne monographs often sit at 0.5–2%.
Dosed
Spot products can carry useful BHA; full-face powders with a whisper of it rarely clear anything.
Tested
Acne-claim products in regulated markets need monograph or drug-facts style labeling. Cosmetics wording is weaker.
When not to buy
All over dry or barrier-compromised skin; or as a substitute for a leave-on you already tolerate. Also skip if fragrance is along for the ride.
hybrid
Serum skin tint — makeup + skincare in one
Serum tints earn the bag on anti-pancake days. They lose it the moment you layer a second full base on top 'just in case'.
Named
Clear jobs: tone evening + stated skincare benefit (SPF, humectants) with ingredients to match.
Dosed
Coverage usually 15–30%. If you need 60, a hybrid will not become a foundation by wishing.
Tested
Wear tests and oxidation notes matter more than serum poetry. Ask how it looks at hour six, not in the tube.
When not to buy
When you need full event opacity, or when the 'serum' finish is pure silicone slip that slides off oily zones by lunch.
hybrid
BB / CC cream as skincare hybrid
BB/CC earns the bag when the range fits your depth and the SPF is real. It loses the bag as a default for everyone — defaults erase undertones.
Named
What the letters actually do in that brand — tone correct, SPF, moisture — named on pack.
Dosed
Often medium coverage with SPF 15–30. Check shade range: short ranges are the hidden cost.
Tested
Shade accuracy across depth bands; oxidation; transfer. Hybrid fame does not equal deep-range honesty.
When not to buy
Deep or olive undertones when the range is six fair-to-medium bottles. Hybrid is not progress if you cannot match.
hybrid
Multi-use balm that 'treats' as it colours
A good balm is architecture: one object, two jobs. A balm that sells itself as a facial is a scent story with pigment.
Named
Emollients and pigments listed; any 'treatment' actives named — not just 'nourishing oils'.
Dosed
Balms are emollient-first. They are not a leave-on acid or vitamin C serum with a blush on top.
Tested
Wear, migration into fine lines, and fragrance load — the practical tests that decide if it stays on the face.
When not to buy
Fragrance-sensitive or rosacea mid-flare days when essential oils are the 'botanical treatment'. Also skip if you need matte control.
barrier
'Soothing' / 'barrier-supporting' complexion
Soothing makeup earns the bag when the formula is short, fragrance-aware, and thin. It loses the bag when 'calm' is the font choice on a scented full-coverage bottle.
Named
Ingredients with a track record (centella, panthenol, colloidal oatmeal) actually listed — and fragrance status clear.
Dosed
Barrier claims without a fragrance-free statement are incomplete. Irritants undo the calm language.
Tested
Repeat-insult patch data is rare for makeup. Sensitive-skin marketing is not the same as dermatologist clearance.
When not to buy
During an active flare when you need bare skin or pharmacy derm-care, not a new pigmented product. Also skip if EO or heavy perfume is on the list.
barrier
'Clean' / toxin-free makeup as skin safety
Clean is a shopping filter, not a halo. The desk honours your filters; it will not pretend a label made the layers thinner.
Named
A published exclusion list you can read — not vibes. Note what is still allowed.
Dosed
N/A — 'clean' is not a dose. It is a filter set. Treat it as preference, never as medical risk reduction.
Tested
Third-party screens vary wildly. Ask what was tested; absence of a bogeyman ingredient is not proof of tolerance.
When not to buy
When 'clean' is the only reason to pay prestige prices for a short shade range or a thick film that cakes. Architecture still wins.
hybrid
Colour product (blush, bronzer) with SPF
SPF blush is a charming lie at face-application rates. Earn the bag on pigment and blend; outsource UV to a real film.
Named
SPF number and filters on that specific stick or cream.
Dosed
You will never apply cheek colour at sunscreen thickness. The SPF is almost always decorative.
Tested
If they only tested the base without pigment, the number does not transfer.
When not to buy
Anytime you would count it toward daily UV strategy. Buy the colour for the colour.
The house position
A finish has to earn
its place in the bag.
No toxin scores. No single “clean” product myth. No purge culture. The desk scores product types against a profile, names the weights, names what it is trading away, and tells you when the honest answer is fewer objects rather than better ones.
Run your edit