Transparent by design
How we score a public website
This is a bounded public-evidence assessment, not a legal, security, accessibility, ranking or conversion certification.
Five observable dimensions, 100 points
The fixed rubric scores technical accessibility and reliability (20), search and content structure (25), core task and conversion path (25), trust and transparency (15), and experience and basic usability (15).
Score, coverage and risk are separate
- Pass: public evidence satisfies a check.
- Partial: public evidence partly supports it.
- Fail: public evidence shows a reproducible issue.
- Unassessed: a bounded GET-only audit cannot judge it reliably. It is excluded from the score and lowers coverage.
- Not applicable: a check has a documented public reason not to apply.
The score is earned points divided by assessed applicable points. Coverage is assessed applicable points divided by all applicable points. Below 45% coverage, the result is marked Provisional and has no performance band.
Risk flags
Possible public credentials are blockers. Missing core CTAs, legal pages, indexability evidence or primary public navigation can be high risks. These flags stay separate from the score so a high number cannot hide a public safety or trust issue.
Sampling and limits
We use safe, unauthenticated GET requests for the submitted page, robots.txt, sitemap.xml, reachable legal pages and a small set of safe navigation pages. We do not log in, submit forms, pay, publish, modify the target, bypass robots or infer private work.