What we measure
A HowLegit audit does not assess whether an organisation is "good" or "bad". It assesses whether an organisation's public claims are substantiated by evidence.
We examine what an organisation says publicly: on its website, in reports, through social channels, in leadership communications. We check whether those statements are backed by verifiable evidence, referenced against established standards, and consistent across channels.
The grading system
Every public-facing claim identified during an audit is graded individually:
| Grade | Meaning | What it takes |
|---|---|---|
| PASS | Substantiated | Claim is supported by verifiable evidence and consistent with referenced standards |
| FLAG | Unsubstantiated | Claim lacks sufficient evidence, is partially supported, or uses imprecise language |
| FAIL | Contradicted | Claim is directly contradicted by available evidence or misrepresents the situation |
Non-compensatory scoring. Strong claims in one area cannot compensate for weak claims in another. A checkpoint with any FAIL claim cannot score above a threshold, regardless of how many PASS claims exist alongside it. This prevents an organisation from making many true claims to dilute the impact of false ones.
Claim classification
Before grading, every claim is classified by type. Different types carry different burdens of proof:
| Classification | Example | Burden of proof |
|---|---|---|
| Absolute | "We are carbon neutral" | Highest. Requires direct, verifiable proof. No substantiation = FAIL. |
| Aspirational | "Working toward net zero by 2045" | Medium. Requires evidence of a credible pathway: interim targets, baseline data, methodology. |
| Contextual | Technical claims within qualifying context | Assessed in situ. Would a reasonable reader of this page understand the claim's scope? |
Classification is stated in the report commentary so legal reviewers can follow the reasoning. The distinction between "we are carbon neutral" (absolute) and "we are committed to carbon neutrality" (aspirational) is material. They carry fundamentally different evidentiary requirements.
The three dimensions
Audits assess claims across three independent dimensions, each with its own grade:
Environmental (E): 6 assessment points
Message integrity and specificity. Evidence integration. Comparative fairness. Visual and symbolic framing. Customer journey consistency. Future framing and strategic alignment.
Social (S): 6 assessment points
Specificity and representation accuracy. Evidence policy and risk disclosure. Remedy visibility. Visual and cultural integrity. Cross-channel consistency. Salience, materiality, and outcome relevance.
Commercial (C): 7 assessment points
Claim clarity and consumer interpretation. Evidence and substantiation. Comparative and superiority claims. Outcome promises and refund guarantees. Pricing and scarcity representation. Platform fit and disclosure prominence. Tone consistency and global risk fit.
19 assessment points across all three dimensions. Each follows established frameworks including ISO 14021, Fair Trading Act (NZ), ESRS/CSRD, GRI, B Corp V7, and ISSA 5000.
Grade bands
Dimension scores map to letter grades using fixed bands. Boundary values belong to the higher band.
| Score | Grade | Score | Grade | Score | Grade |
|---|---|---|---|---|---|
| 90-100% | A | 60-66% | B- | 40-46% | C- |
| 80-89% | A- | 53-59% | C+ | 33-39% | D+ |
| 73-79% | B+ | 47-52% | C | 27-32% | D |
| 67-72% | B | 20-26% | D- |
Confidence bands
Every grade is accompanied by a confidence band indicating how much evidence the grade rests on:
| Distinct claims identified | Confidence | Meaning |
|---|---|---|
| 15+ | High | Grade based on an extensive claim landscape |
| 6-14 | Medium | Grade directionally reliable |
| 1-5 | Low | Grade based on a thin claim base |
| 0 | -- | Insufficient data for a letter grade |
Confidence bands are based on the number of distinct public-facing claims identified and assessed, not the number of assessment points (which are fixed). This captures how much evidence the grade actually rests on.
The audit process
- Evidence collection Automated scan of all public channels: website pages, PDF reports, social media profiles, press releases. Supplemented by adversarial research into what the organisation is not saying.
- Claim identification Every public-facing statement is catalogued, classified (absolute, aspirational, or contextual), and mapped to the relevant assessment point.
- Grading Each claim is assessed against the relevant standard. Individual claim grades roll up to assessment points, which roll up to dimension scores using non-compensatory aggregation.
- Human review An auditor reviews the analysis, checking standards application, data inputs, and potential bias. Grade rationale is documented with two-sided reasoning: why the grade is not lower, and what brought it down.
- Publication and right of reply Findings are published with full evidence citations. The graded entity receives free access to their findings and can submit counter-evidence, comment, and request a regrade.
Right of reply
Every graded entity receives free access to their findings at app.howlegit.com. This includes:
- Full findings with evidence citations
- Ability to submit counter-evidence and comments
- Request a regrade based on new evidence
- Track remediation progress
If submitted evidence changes the assessment, the grade is updated and the full history is preserved. The right of reply is free and unconditional. No payment is required to respond to findings.
Standards referenced
Assessments are benchmarked against established frameworks relevant to the jurisdiction and sector:
A HowLegit audit is a messaging integrity diagnostic, not a legal compliance opinion. Grades reflect the substantiation of public claims against referenced standards at the time of assessment. They do not constitute legal advice.