ScanVeer

Scoring Registry Changelog

Every change to how ScanVeer weighs an ingredient, dated, with its evidence reason. If a score moved, the entry that moved it is on this page.
The rule behind every entry: ScanVeer scores exposure-realistic risk, not intrinsic hazard, and every weight must be derivable from its evidence tier — E1 established harm at real-world exposure (weight 4–5) · E2 strong evidence, threshold disputed (3–4) · E3 mixed human evidence (2–3) · E4 animal-only / in-vitro / mechanism-only (1) · E5 no credible evidence of harm at use levels (0–1). A weight outside its tier's range fails our build. Weights change only through this ledger — never quietly.
2026-08-30

Telling us what is in your fragrance no longer costs a product points

disclosure is not a penaltyno verdict moved

What was wrong: European law requires a maker to print certain fragrance ingredients by name — linalool, limonene, citronellol and others — once they are present above a very small amount. It is not a confession. It is the law making the label tell you what is inside the word "parfum". We were counting it twice: once for the fragrance, and again for each name that spelled out what the fragrance contained. The effect was backwards. Across 500 real products we checked, those that told you nothing beyond "parfum" averaged 51.8 and sat in Good, while those that named their allergens averaged 49.4 and sat in Poor. We were quietly rewarding the products that told you less.

What changed: when a label says parfum or fragrance, the named allergens beside it are treated as that disclosure rather than as separate problems, and they no longer add to the score. They stay on the product page, listed by name under their own heading, because if you react to linalool you need to see the word linalool — you just are not charged for the maker having printed it. 114 products moved, every one of them upward.

What did NOT change: no product's rating moved. Excellent, Good, Poor and Bad hold exactly the same products as before. Nothing was re-rated up or down; a penalty that should never have applied was removed.

Two things we deliberately left alone: ingredients that are banned in the EU, such as Lilial and HICC, still carry their full weight even though they are also fragrance substances — a banned ingredient is not a disclosure. And where a label names an allergen without ever saying parfum, the named ingredient still counts, because there it is not describing a fragrance elsewhere on the label; it is the fragrance.

Still open: undisclosed fragrance is present in most cosmetics, so it still influences a large share of cosmetic scores. Whether it should carry that much weight is a separate question we have not answered yet. This change fixes the part that was plainly unfair — punishing the honest label — not the larger one.

Checked against the same 2,000 real barcoded products (500 cosmetics) frozen on 2026-08-29. Products moved: 114, all upward. Products changing rating: 0. Requirement: EU Regulation 1223/2009, Annex III.
2026-08-29

Cosmetic scores now tell similar products apart — no product was re-rated

resolution, not re-ratingno verdict moved

What was wrong: our cosmetic score only looked at the most serious concern in a product and stopped there. Everything milder was found, listed for you on the product page, and then quietly left out of the number. The result was that very different products got the same score. Across 500 real products we checked, 264 of them — more than half — scored exactly 52. A product with one moderate concern and nothing else looked identical to one carrying that same concern plus five milder ones.

What changed: milder ingredients now count toward the score instead of being ignored. On the same 500 products, the number of different scores we produce went from 15 to 27, and the pile-up on 52 dropped from 264 products to 26.

What did NOT change: no product was re-rated. Nothing moved between Excellent, Good, Poor or Bad — we checked every one of the 500 and the counts in each band are identical before and after. If a product read Good yesterday it reads Good today. This was about telling products apart, not about judging them differently.

A limit we have not solved: 161 of those products now sit together at 50, the bottom of the Good band. They are not all the same product — every one of them carries extra milder ingredients beyond its main concern, some two of them and some as many as nine, and today we still show them all the same number. We hold them at 50 on purpose: letting them fall further would have moved 90 products from Good to Poor, and re-rating products was not what this change was for. The honest reason the room ran out is that our "one moderate concern" starting point sits only two points above the Good boundary. Whether a product with one moderate concern should be called Good at all is a separate question, and it is still open. When we answer it, it will get its own entry here.

Checked against 2,000 real barcoded products (500 cosmetics) frozen on 2026-08-29, not hand-picked examples. Before/after counts: distinct scores 15 → 27; products changing band 0.
2026-08-25

Cellulose gum (CMC) — tier E3 → E5, weight 2 → 1 · Cyclamate added to supplements

CMC re-tiered on new evidencesupplement coverage

Reason: EFSA’s re-evaluation of the cellulose family (verified today) found no safety concern at reported uses and no ADI needed — a risk assessment that sets the tier under rubric §7. The small human emulsifier trial that motivated the old tier is noted in the panel as an open research question. Cyclamate joins the supplement registry at its aligned weight so a cyclamate-sweetened supplement is flagged and capped like every other non-nutritive sweetener.

Source: EFSA Journal re-evaluation of celluloses (PMC)
2026-08-25

The full ledger correction — 109 weights re-derived from their evidence tiers

106 weights DOWN3 weights UPrange rule fully armed

Reason: the evidence-tier audit found weights set by intuition years ago that the attached evidence never supported. After the four highest-impact cases were ruled individually (BHA/BHT, saccharin, aspartame, erythritol), the remaining 109 keys are corrected in one pass: every weight now sits at the nearest bound of its evidence tier's allowed range. Notable: minor synthetic dyes, caramel colors, food-context parabens, gallates, TBHQ, polysorbates, EDTA and flavor nucleotides move DOWN; formaldehyde-releasing preservatives (established human sensitizers) move UP. From this entry forward the build fails on any weight outside its tier — there is no list of exceptions left.

Every key (109) — old weight → new weight, with tier
MapKeyOldTierNew
cosmeticbutylparaben5E41
additivee9524E41
cosmeticbenzophenone4E41
cosmeticmethylparaben4E41
cosmeticparaben4E41
cosmeticpropylparaben4E41
namedcyclamate4E41
additivee1043E41
additivee1313E41
additivee150d3E41
additivee1513E41
additivee2143E41
additivee2153E41
additivee2183E41
additivee2193E41
additivee3193E41
additivee6213E41
additivee927a3E41
cosmeticci 191403E41
cosmeticcocamide dea3E41
cosmeticethanolamine3E41
cosmetichydroxyapatite [nano]3E41
cosmeticnano-hydroxyapatite3E41
cosmeticperfluoro3E41
cosmeticphenoxyethanol3E41
cosmeticphthalate5E33
cosmeticpolytetrafluoroethylene3E41
cosmeticptfe3E41
cosmeticresorcinol3E41
cosmeticretinyl palmitate3E41
cosmetictriclosan5E33
cosmetictriethanolamine3E41
foodConcernaroma artificial3E41
foodConcernartificial flavor3E41
foodConcernarôme artificiel3E41
foodConcernblue 13E41
foodConcerncarrageenan3E41
foodConcernmonosodium glutamate3E41
namedazodicarbonamide3E41
namedblue 13E41
namedcaramel color3E41
namedcaramel colour3E41
namedethylparaben3E41
namedmethylparaben3E41
namedmonosodium glutamate3E41
namedtbhq3E41
namedtert-butylhydroquinone3E41
suppblue 13E41
suppcarrageenan3E41
additivee1244E33
additivee1322E41
additivee1332E41
additivee133b2E41
additivee150c2E41
additivee1742E41
additivee2114E33
additivee3102E41
additivee3112E41
additivee3122E41
additivee3852E51
additivee4072E41
additivee4322E41
additivee4332E41
additivee4342E41
additivee4352E41
additivee4362E41
additivee4772E41
additivee6222E41
additivee6232E41
additivee6272E51
additivee6312E51
additivee6352E51
cosmetic2-bromo-2-nitropropane3E14
cosmeticbronopol3E14
cosmeticci 159852E41
cosmeticcyclopentasiloxane2E41
cosmeticmineral oil2E51
cosmeticoxybenzone4E33
cosmeticpeg-2E41
cosmeticpetrolatum2E51
cosmeticpolyacrylamide2E41
cosmeticpropylene glycol2E51
cosmeticsiloxane2E41
cosmeticsodium hydroxymethylglycinate3E14
cosmeticsodium laureth sulfate2E51
cosmetictriclocarban4E33
foodConcernartificial color4E33
foodConcernartificial colour4E33
foodConcerncaramel color2E41
foodConcernmodified corn starch2E51
foodConcernmodified starch2E51
foodConcernmono- and diglycerides2E41
foodConcernmonoglycerides2E41
foodConcernpotassium sorbate2E51
namedblue 22E41
namedcarrageenan2E41
nameddisodium guanylate2E51
nameddisodium inosinate2E51
namededta2E51
namedmonopotassium glutamate2E41
namedpolysorbate2E41
namedpropyl gallate2E41
namedribonucleotide2E51
suppacesulfame4E33
suppartificial color4E33
suppartificial colour4E33
suppartificial flavor2E41
suppartificial flavour2E41
suppsucralose4E33
2026-08-25

Non-nutritive sweeteners — the never-green cap now covers ALL products

category cap extended

Reason: the beverage cap shipped hours earlier; this extends it. The WHO guideline we cite advises against non-nutritive sweetener use generally — not only in drinks — so a sugar-free gum or dessert sweetened with aspartame or sucralose now also tops out at 49 (Poor). Sugar alcohols (erythritol, xylitol) are a different class and are not affected by this cap; their own evidence-tiered weights apply.

Source: WHO non-sugar sweeteners guideline (2023)
2026-08-25

Artificially sweetened beverages — capped at 49 (never green)

beverage category cap

Reason: a drink whose entire formula is water plus non-nutritive sweeteners should never present as "Good." The WHO advises against non-sugar sweeteners for weight control, and the Nutri-Score 2023 beverage algorithm itself lets only water score well. A beverage containing any non-nutritive sweetener now caps at 49 (Poor). This is a category rule — the individual sweetener weights stay evidence-tiered exactly as ruled the same day (aspartame E4/weight 1).

Sources: WHO non-sugar sweeteners guideline (2023) · Nutri-Score 2023 beverage algorithm
2026-08-25

Sucralose — text-detection weight 4 → 2 · Acesulfame K — 4 → 3

detection-path alignment

Reason: a defect, not a re-evaluation: the same sweetener carried different weights depending on whether it was detected from a database tag (correct, evidence-tiered) or from the printed ingredient text (stale, higher). A product could score differently based on which database answered. Text detection now matches the tag weights exactly, and a test forces the two paths to stay identical forever.

2026-08-25

Rubric ruling: a hazard classification is not an evidence tier

binding rubric §7

Reason: risk assessments at real-world exposure (JECFA, EFSA, FDA, SCCS) now SET evidence tiers; hazard identifications (IARC, NTP, Prop 65) can raise a tier at most one step and only with a human-relevant mechanism; species-specific mechanisms (rat forestomach, alpha-2u-globulin, rodent peroxisome proliferation) are excluded; when bodies disagree, the copy must say so; and emerging human evidence can move a weight UP ahead of regulators. The four entries below apply this ruling.

2026-08-25

BHA — weight 4 → 2 (tier E3) · BHT — weight 4 → 1 (tier E4)

BHA 4 → 2BHT 4 → 1family split

Reason: the two antioxidants were weighted as one family; the evidence differs. BHA’s carcinogen listing (NTP) rests on rat forestomach tumors — an organ humans do not have — while reviews remain active: mixed evidence, weight 2, no longer triggers the high-risk cap. BHT has an established EFSA acceptable daily intake: the risk assessor sets the tier, weight 1. Both panels state the disagreement between hazard bodies and risk assessors in plain language.

Sources: NTP Report on Carcinogens (BHA profile) · EFSA Journal 2012 BHT re-evaluation
2026-08-25

Saccharin — weight 3-4 → 1 (tier E5)

weight 4 → 1tier E5

Reason: the worked example of the new ruling. The historical cancer concern came from a rat-specific bladder mechanism that does not operate in humans, and NTP removed saccharin from its Report on Carcinogens in 2000. Our old weight was a 25-year-stale hazard echo. The WHO’s class advisory on non-nutritive sweeteners remains cited.

Sources: NTP RoC delisting (2000) · WHO non-sugar sweeteners guideline (2023)
2026-08-25

Aspartame — weight 4 → 1 (tier E4)

weight 4 → 1tier E4

Reason: IARC’s Group 2B ("possibly carcinogenic") is a hazard flag on limited evidence. The same day, JECFA — the body that assesses risk at real-world intake — re-affirmed the 40 mg/kg/day acceptable intake (roughly 9–14 diet drinks). Hazard asks "could it ever, at any dose?"; risk asks "does it, at doses people actually consume?" — and the risk answer remains no at typical consumption. Both findings now appear together in the evidence panel.

Source: WHO — Aspartame hazard and risk assessment results (IARC + JECFA, 2023)
2026-08-25

Erythritol — weight 4 → 3 (tier E3, rule-6 hold)

weight 4 → 3emerging human evidence

Reason: two human studies — a >4,000-person association with cardiovascular events (Nature Medicine 2023) and an interventional study showing 30 g acutely enhanced platelet reactivity in every healthy volunteer tested (ATVB 2024) — hold erythritol at the TOP of the mixed-evidence tier even though regulators still classify it GRAS. This is the rubric’s rule 6 working as intended: evidence can move weights up ahead of regulators. The old weight 4 (a hard score cap) overstated today’s evidence; weight 3 keeps a prominent orange flag.

Sources: Nature Medicine 2023 (NHLBI summary) · ATVB 2024 brief report (PMC)
2026-08-25

Commonly-avoided groups: sulfates, synthetic dyes, talc, denatured alcohol, non-nutritive sweeteners

non-scoring badges6 of max 8 groups

Reason: five new groups join aluminum in the commonly-avoided system — neutral, informational badges for ingredients many people choose to avoid. They never change a score, and allergens are deliberately excluded: for an allergic user an allergen is a safety fact handled by the separate allergen-alert system, never a lifestyle preference. Denatured alcohol enters as a weight-0, information-only entry.

2026-08-25

Antiperspirant aluminum salts — weight 3 → 1 (tier E5)

weight 3 → 1 tier E5 — no credible evidence of harm at use levels tagged: commonly avoided

Reason: our previous weight treated antiperspirant aluminum as a moderate hazard. The evidence does not support that: the National Cancer Institute finds no scientific link between antiperspirant aluminum and breast cancer, and measured skin absorption is very low. The score now reflects the evidence; the widely-held preference to avoid aluminum is served by a separate non-scoring tag instead of an inflated hazard weight.

New in-app evidence copy: “No established link to breast cancer or Alzheimer’s. Skin absorption is very low (roughly 0.0005-0.01%). Can irritate sensitive or freshly shaved skin. Wash off before a mammogram — it causes imaging artifacts. Long-term cumulative low-dose exposure has not been definitively studied.”

Keys: aluminum/aluminium chlorohydrate · aluminum/aluminium zirconium · aluminum/aluminium chloride  ·  Sources: NCI Antiperspirants/Deodorants & Breast Cancer fact sheet · Alzheimer’s Association myths page · aluminium-26 dermal-absorption tracer studies (Flarend 2001; de Ligt follow-up, PMC) · NCI Mammograms fact sheet