Published draft methodology

How NORM v1.1 is designed
to assess public claims.

NORM v1.1 is a published draft under implementation and validation. Zero NORM v1.1 assessments have been published. Legacy v8 test runs are excluded from public coverage, scoring, ranking, and market statistics.

Consumer reviews follow the separate consumer-review-1.0 editorial policy, adopted October 13, 2026. NORM’s AI-assisted review is final for that route; it does not constitute independent human scientific review or v1.1 validation. Read the AG1 consumer review.

Core principle
One question, six diagnostic areas

Does the evidence a brand shows match the claims it makes? The v1.1 draft defines six diagnostic areas. No v1.1 assessment will be published until implementation, coverage, range, reviewer, arithmetic, and validation gates pass.

EC

Evidence Coverage

How many of the brand’s material claims have eligible evidence publicly presented by the brand?

ST

Source Traceability

Can the brand’s cited sources be identified, resolved, and connected to specific claims?

PM

Product Match

Does the cited evidence study the marketed product, formulation, ingredient, or delivery method being claimed?

UC

Use-Condition Match

Do population, dose, duration, and relevant conditions in the cited evidence match how the claim is presented to consumers?

RM

Result Match

Do the study design, measured outcome, and reported results in the cited evidence support the specific claim?

CP

Claim Proportionality

Does the marketing language remain proportionate to the evidence’s findings, qualifications, uncertainty, and limitations?

Under the draft method, each material claim would be assessed individually. Aggregate scoring remains provisional until validation is complete.

Provenance
Every future finding must have a paper trail

The v1.1 draft requires seven record layers. These are publication requirements. Calibration tests do not satisfy them and are not public assessments.

L1 · ASSESSMENT MANIFEST

The assessed entity, declared scope, capture time, method version, pipeline version, schema version, and publication status.

L2 · PAGE CAPTURES

Each reviewed page, its canonical address, capture method, capture time, response status, and content fingerprint.

L3 · SOURCE ASSETS

The preserved source material used for capture or review, with stable identifiers and integrity hashes where the storage system supports them.

L4 · REVIEW RUN

The tool versions, automated assistance, reviewer status, adjudication state, execution time, and validation results for the run.

L5 · CLAIM RECORDS

Each material claim as quoted, with page location, capture time, claim class, audience, materiality, and extraction status.

L6 · EVIDENCE RECORDS

Each cited source with identifiers, study design, population, intervention, comparator, outcomes, funding, correction status, and provenance.

L7 · FINDINGS AND CORRECTIONS

The claim to evidence match, finding, limitations, reviewer status, reason codes, challenges, adjudication, and material correction history.

The target system preserves versioned captures and correction history. Retention and immutability controls remain subject to implementation verification.

Intended users
One draft method, three use cases
👤
Consumers

Inspect a public record and its limitations. NORM does not recommend purchases or determine whether a product works.

🏢
Brands & Retailers

Identify language and provenance gaps through private workflow tools. Payment does not purchase a public finding.

🤖
AI Agents & Platforms

Use a future versioned record only after checking method, coverage, confidence, and review status. No v1.1 record is currently published.

For the complete published draft, including scope, proposed rules, limitations, and validation status:

Read the full whitepaper →

Zero NORM v1.1 assessments or scores have been published. Legacy v8 test runs are excluded from public coverage and comparison.

NORM is a private claim accountability platform, not a government regulator, certification body, medical authority, or legal opinion.

Consumer persuasion review · version 1.0 · October 13, 2026

We check the pitch and the path to purchase.

Every new Norm company consumer assessment must now account for four areas before editorial completion. This is a bounded source-backed review, separate from scientific support and the v1.1 draft scoring method.

Influencers and incentives

Sample actual statements, claims and disclosures. Identify verified sponsor, affiliate, advisory and ownership relationships, without inventing fees or assuming every recommendation is paid.

Website and offers

Check desktop/mobile claims, evidence access, testimonials, badges, price framing, selected plans, gifts and urgency. Record the observed market and session; do not assume one visitor’s selection is universal.

Advertising

Inspect a defined sample of creative or accessible transcripts and record which words, images, audio and disclosures were actually checked. Connect the promise to the product and evidence. State inaccessible or uninspected channels.

Buying and leaving

Follow the public path through plan choice and checkout where possible. Compare upfront and recurring cost, consent, renewal, refunds and cancellation terms. Reading a policy does not test the real cancellation experience.

Each finding separates what we observed, how it could influence a choice and what a buyer can do with that information. Retain the source, date, product version, quotation or media location, and material limits. Consider ordinary explanations as well as possible pressure. No invented manipulation score, motive, ad spend or measured effect.

All 16 checklist items must be accounted for. Unchecked work prevents completion. Partial access, a documented absence in checked sources or a justified scope limit stays visible and permits only bounded completion. Earlier reviews retain their original scope until a new dated review is performed. A persuasive presentation never changes the underlying scientific evidence.

This requirement is adopted for Norm’s editorial workflow and implemented in its local completion validator. Hosted intake enforcement and automated collection at scale remain to be implemented. Completion does not imply independent human review, legal compliance, safety certification or validation of the method. See the AG1 example · Publication policy.

© 2026 BHVD Inc.  ·  NORM public beta  ·  Published draft methodology