Side by side.
Ingredient-versus-ingredient comparisons, cited and conservative. No winner declared: who should choose which, and why.
Same table for every reader. Same sources for every claim.
Lactic vs glycolic vs mandelic acid for sensitive skin
Lactic acid · Glycolic acid · Mandelic acid
Three alpha hydroxy acids at three speeds: glycolic is the smallest and fastest, lactic also hydrates, mandelic is the largest and slowest, which is why reactive skin often starts there.
Updated 16 September 2026
Niacinamide vs vitamin C for uneven tone
Niacinamide · Vitamin C
Niacinamide slows pigment transfer and is stable and forgiving. Vitamin C acts on the pigment-making enzyme and adds antioxidant support, but pure L-ascorbic acid is fussy. Many routines use both.
Updated 16 September 2026
Tranexamic acid vs kojic acid
Tranexamic acid · Kojic acid
Tranexamic acid calms the signals that switch pigment cells on and is stable at 2–5%. Kojic acid blocks the pigment enzyme directly but oxidises in the bottle and can sensitise, so the EU caps it at 1%.
Updated 16 September 2026
Potassium azeloyl diglycinate vs azelaic acid
Potassium azeloyl diglycinate · Azelaic acid
Azelaic acid has decades of evidence, mostly at 15–20% medicine strengths that can sting. Potassium azeloyl diglycinate is its water-soluble, glycine-linked cousin: gentler, easier to formulate, far less studied.
Updated 16 September 2026
Modern UV filters vs legacy filters
Modern filters (Tinosorb S, Tinosorb M, Uvinul A Plus) · Legacy filters (avobenzone, oxybenzone, octinoxate)
Tinosorb S, Tinosorb M and Uvinul A Plus are photostable, cover UVB through long UVA and barely penetrate skin. The legacy trio are US-approved but need stabilising or carry absorption and regulatory concerns.
Updated 16 September 2026
The Comparator opens soon.
Paste any product's INCI and see how it compares to The Constant on efficacy, irritation, delivery and evidence. One email when it opens. Nothing else.
By subscribing you agree to receive this email per our Privacy Notice.
Cited endpoints, not vibes.
Reported tolerability profile.
Whether the active gets where it needs to go.
Study type, sample, duration, source.