Evisentra Claim Readiness Rubric v1.0
How a claim and its evidence are judged. The method and rules are public and the reasoning is reproducible, so you can check (and challenge) it — while your evidence stays private.
Claim-readiness dimensions: 0 = missing · 1 = weak · 2 = partial · 3 = strong. Evidence items are graded separately on a 1–5 quality (DQR) scale.
What the 1–5 evidence grade means
| Grade | What it looks like |
|---|---|
| 1/5 | Self-declared or unverifiable — no standard, issuer, or scope you can check. |
| 2/5 | Relevant but weak — second-party or partial; named method missing or scope unclear. |
| 3/5 | Credible but incomplete — a recognised method, but second-party or not fully scope-matched to the claim. |
| 4/5 | Strong — recognised standard, dated, mostly scope-matched; minor gaps. |
| 5/5 | Independent, standard-based, current, and scope-matched to the exact claim. |
| Dimension | What we check |
|---|---|
| Claim specificity | A specific, named, provable attribute — not a vague 'eco / green / sustainable'. |
| Evidence fit | The evidence type actually addresses the claim (recyclability needs collection data, not an LCA). |
| Boundary clarity | System boundary / functional unit stated; any comparison is like-for-like. |
| Market-rule fit | Assesses fit with selected markets — for example US FTC + California AB 1305, EU consumer green-claim rules, and UK DMCC/CMA guidance. |
| Standard / method named | A recognised standard is cited (ISO 14067/14021/14026, EN 13432, ASTM, GHG Protocol). |
| Evidence independence | Third-party verified outranks second-party, which outranks self-declared. |
| Burden-shifting risk | A benefit claim does not hide a bigger impact elsewhere (low carbon vs high water / land). |
| Public / private evidence status | The basis is public and re-checkable, or private and not independently re-checkable by outside readers. |
| Uncertainty / gap disclosure | Uncertainty and missing data are disclosed, not hidden in a headline figure. |
| Safer wording available | A specific, qualified rewrite exists when the literal claim is risky. |
Who authored this
Authored by Evisentra (Nutavix LLC) from public regulatory guidance (US FTC Green Guides + California AB 1305, the EU Empowering Consumers Directive, the UK DMCC Act) and public LCA standards (ISO 14067/14021/14026, EN 13432, GHG Protocol). The rubric and rule mapping are public and reproducible — not a proprietary black box. Private client evidence remains private.
How to challenge it
Disagree with a dimension or a score? The rubric is public, and method challenges are welcome. Send a critique to [email protected]. Substantive method corrections may be reflected in rubric updates.
The 'standard / method named' dimension draws on a public list of recognized global standards — see standards & methods →
Method history
Two independent version layers: the rubric (the judgment dimensions, currently v1.0) and the rule-base/engine (the encoded rules and evidence requirements, currently preflight-v1.5). Reports state both. Versions on this page are generated from the same constants the engine uses, so they cannot drift from the reports.
v1.5 — July 2026: evidence-strength gate — a requirement now counts as fully Supported only with independent third-party evidence against a recognized standard; anything weaker (self-declared, second-party, or no named standard) is Supported with qualification and no longer counts toward the fully-supported coverage figure. The recognized-standards list behind evidence grading was extended to the certification programmes and test methods in the public standards registry (GRS, RCS, ISCC, ASTM D6400/D6866, EN 16640/17681, AS 5810, BioPreferred, FSC, PEFC, TOF). PPWR recyclability wording corrected to grades A/B/C (below C = technically non-recyclable); EN 13432 and EN 17681-1 catalogue links corrected; the “collection actually happens” criterion re-grounded in PPWR Article 6 with the 2024/825 reading labelled as an interpretation.
Symbols & imagery update — July 2026: symbol mentions (chasing-arrows/Mobius, resin codes, Seedling/BPI/OK-compost marks, eco logos and leaf imagery) now trigger the matching claim family in the free Claim Check, with symbol-specific pitfalls in the rule content; the paid engagement's artwork & symbols check is named explicitly. Detection only — no change to evidence requirements, scoring, or verdict logic; fingerprints unchanged.
Traceability update — July 2026: criterion IDs (EV-…), an assessment-basis section, a criteria crosswalk, defensible determination labels (Supported / Supported with qualification / Insufficient evidence / Contradicted / Not applicable), and a three-conclusions split added to every report. Presentation and traceability only — no verdict-logic change; content fingerprints unchanged.
v1.4 — July 2026: recyclable evidence requirements split — technical recyclability (composition/format) and actual market collection are now separate requirements, so a met technical requirement can no longer mask missing collection evidence.
v1.3 — July 2026: Canada row updated (the internationally-recognized-methodology requirement was removed by the Budget 2025 Implementation Act, Royal Assent March 26, 2026); PFAS state rows split (ME / MN / CA) and an EU PPWR Art. 7 food-contact PFAS row added; AB 1305 and EU 2024/825 summaries expanded with applicability detail.
v1.2 — July 2026: added bio-based, BPA-free, reusable/refillable, and source-reduction claim families; published the public rule index at /rule-base.
v1.1 — July 2026: added PFAS-free / restricted-substances and compostable claim families; added Canada market rules (Competition Act s.74.01(1)(b.1)/(b.2), Bill C-59 + Competition Bureau guidelines).
v1.0 — July 2026: first public release. Substantive corrections are logged here.
Content fingerprint
Every report carries a content fingerprint— the same claim, evidence, and method version always produce the same hash, so content changes are detectable. It is a change-detection check, not certification or authentication.
What QC checks before a report ships
- Every rubric dimension is scored — none skipped.
- Every conclusion cites a submitted document or a public rule.
- Gaps are recorded, never filled in with opinion.
- The specialist-opinion recommendation is set for the claim family.
- The report ID and content fingerprint are checked.