A score with
something behind it.

How the live baseline works

Version 2.2 parses public HTML from the homepage and up to five representative linked pages (about, leadership, contact, investors or offering, customers), respecting robots.txt. When an isolated rendered browser is configured, the homepage is also loaded at 390px mobile and 1366px desktop to check sideways scrolling, navigation and the mobile menu (one safe click on an identified menu control), keyboard focus and reach (Tab only), contact destinations (never submitting forms) and every axe-core rule violation; content seen only after JavaScript runs is labelled as rendered evidence. Each check comes from a versioned catalog and records one status: verified, requirement not met, not observed in sample, not tested, blocked/error or not applicable. Only verified (one point) and requirement-not-met (zero) results are scored; absence in a sample is never scored as a failure. Category averages are weighted using the model below. Untested categories are excluded and remaining weights are normalized. The headline is an observed-signal pass rate, not a readiness rating. An overall readiness verdict is issued only when at least 60% of applicable checks were scored and at least 6 of 8 categories have scored evidence; otherwise the report states that evidence is insufficient. Unknowns are never converted into failures.

Scored-evidence coverage divides scored checks, and completed coverage divides scored plus not-observed checks, by the checks applicable to your company type in the catalog; reports also show attempted, blocked and tested-weight figures, so reports with different scopes are not directly comparable. JavaScript-rendered content, hidden tools and pages outside the sample may be missed. Industry is recorded for review; industry-specific scoring and peer benchmarks are not yet applied.

What needs additional verification

Core Web Vitals and speed require performance tests and available field data. Accessibility needs automated and manual testing. Information accuracy, freshness, newsletter delivery, CRM connectivity and internal processes require review or authorization. Visible software signatures only suggest a tool may be present.

Standards that can evolve

We use foundational search practices, clear factual company content and WCAG 2.2 AA as assessment references. No special AI file guarantees visibility. Scoring versions should evolve as standards and verified evidence change.

Google Business Profile & local discoverability

An optional, separately scored diagnostic (catalog GBP-01 to GBP-14). We automatically search using your website and inferred company name, with trading name and location as a fallback, and show candidates for you to confirm — never auto-selecting — and check identity, status, website link, name, phone and address (compared separately), hours and review presence. Category, description and photo quality are reported as observed signals only and remain unverified. Online-only businesses are not penalised; unresolved eligibility stays unscored. Public data includes at most five review samples, so review history, velocity and response rates require owner-connected access. We do not measure local rank. Google-returned content is excluded from exports pending a terms review.