why this demo

Personalization is where consumer skincare economics actually live — retention, willingness-to-pay, and data network effects all hang off ‘this score is about YOU’. This demo proves goodbadgray’s personalization is real math over real config vectors (weights × tilts × overlays, reproduced verbatim from the engine), not a marketing segment label. Showing the full breakdown table is deliberate: it demonstrates the system is auditable, which is the property that lets the same engine be sold B2B to retailers and clinicians who cannot ship black boxes to their customers.

every score, weight, tilt, overlay, verdict and gated ingredient on this page is verbatim output of the live gbg engine, harvested 2026-07-08. the page recomputes the weight algebra client-side and it matches the engine to machine precision.

personalization, in full daylight

same bottle. different skin. different answer.

the live engine analyzed one real sunscreen exactly once — 42 ingredients, one 14-metric score vector. below, six real profiles read that same analysis through their own weights. six different verdicts, and every multiplication shown.

8,595 ingredients scored by the live engine · 28 products analyzed · this page shows 6 of 8 harvested profile runs for this product — nothing here is mocked

the six faces

click a face — the full calculation opens below

the math, in daylight

score × weight = contribution · sum the column · that is the answer