Scoring Methodology
This page explains, in plain language, what goes into a ScanVeer score, what a score means, when scores change, and the rules that keep them independent. The detailed public walkthrough with worked examples is at How We Score; the dated ledger of every methodology and evidence-weight change is the Registry Changelog.
What a score is
A ScanVeer score is a number from 0 to 100 representing ScanVeer's assessment of a product relative to other products in its category, computed by the same documented method for every product. Scores are grouped into tiers (for food, cosmetics and supplements: Excellent 75–100, Good 50–74, Poor 25–49, Bad 0–24; for household products: Great, Good, Safe, Bad on a hazard-based scale). Some product classes deliberately get no number: fragrances, whose formulas are sealed, are shown as a disclosure profile instead of a score.
What ScanVeer considers
- Ingredient information — the declared ingredient list, its order, and which ingredients match ScanVeer's reviewed concern registry.
- Nutrition information — for foods and drinks, the declared nutrition panel (energy, sugars, saturated fat, sodium, fibre, protein and related values), per serving where the label gives one.
- Scientific evidence — published assessments from regulators and scientific bodies (for example EFSA, the U.S. FDA, IARC, NTP, WHO, the EU Scientific Committee on Consumer Safety, NIH ODS) and peer-reviewed research, each recorded in an evidence registry with its source.
- Evidence strength — how strong and how direct the evidence is. Each concern carries an evidence tier; a regulator's formal restriction weighs more than a single animal study, which weighs more than an association in one observational paper.
- Scientific consensus — whether major bodies agree. Where they disagree (for example a hazard classification from one body and a "within acceptable intake" conclusion from another), ScanVeer shows both rather than picking a side silently.
- Formulation — the whole product, not one ingredient in isolation: how many concerns, how severe, whether a single high-severity concern caps the score.
- Exposure and serving context — leave-on versus rinse-off for cosmetics, dose and daily-intake reference points for additives, serving size for foods, concentration where a label states it.
- Intended population — whether an assessment concerns the general population, children, pregnancy or another group, recorded with each evidence entry.
- Product category — foods, cosmetics, supplements, over-the-counter drugs, fragrances and household products are scored by different rules suited to how they are used.
The precise weights, thresholds and constants are ScanVeer's proprietary method. Where they have been published (see How We Score) they are published deliberately; they are not exposed elsewhere.
What a score is not
A score is not a laboratory result, a safety certification, a medical judgement or a personal recommendation. See the Health Disclaimer. Where ScanVeer lacks enough data to score honestly, it shows a "provisional" state and asks for the label instead of guessing.
When scores change
Scores may change, in either direction, when:
- new research becomes available and an evidence entry is added, upgraded or downgraded;
- a product is reformulated and the database record or a label photo is updated;
- data is corrected — by a community submission, a reviewer, or a brand correction ticket that passed review;
- the methodology is improved — every such change is dated in the Registry Changelog with its reason.
Because of this, a score you saw last month may differ from the score today, and a shared screenshot may not match the live result.
Independence
Brands cannot purchase a better scientific score. No payment, advertising, sponsorship, affiliate commission, certification fee or commercial relationship influences any score, tier, flag, evidence summary or recommendation. The scoring engine has no knowledge of links, tags, commissions or customers. Correction requests from brands go through the same human review as anyone else's and never change a score automatically — see the Brand Corrections Policy.
Evidence registry and product-facing language
Every material ingredient concern that affects a score references an entry in ScanVeer's evidence registry, which records the ingredient, the concern, the wording shown to users, the evidence strength, the source organisation and document, the exposure context, known limitations and the review date. Product-facing wording comes from these reviewed entries. Where the registry is still being extended to older entries, the app labels the evidence status honestly ("unreviewed is not the same as safe").
Data sources
Product data: Open Food Facts, Open Beauty Facts, Open Products Facts, NIH Dietary Supplement Label Database, USDA FoodData Central, U.S. FDA openFDA, UPCitemdb, and products added by ScanVeer users (reviewed before publication). Evidence: the bodies named above and peer-reviewed literature, each cited inside the app.