Methodology

Scores should explain themselves.

A score is a compressed summary, not a verdict on a business.

Our rules

• Every score shows the observed value, source, points earned, and points available for each check.
• Unknown and not-applicable checks are shown but excluded from the score denominator. Evidence coverage is reported separately.
• Recommendations are ordered by the largest disclosed point gaps.
• Estimates expose formulas and assumptions.
• Every browser rubric has a version so methodology changes are visible.
• Provider-backed results remain disabled when trustworthy live data is unavailable.

Rendered website checks

Live website audits load the public page in mobile Chromium using a disclosed network and CPU profile. They collect rendering timings, request and transfer totals, rendered content and conversion signals, and automated axe-core accessibility findings. Lab measurements vary between runs and are not presented as Google PageSpeed Insights or Google field data.

Language and intent

Service, location, and action language is matched by concept, not only by literal substrings. An equivalent phrase can establish clarity while the result still reports whether the exact target wording appeared. Editorial character ranges and heading counts are treated as notices, not invented Google ranking rules.

Google PageSpeed Insights uses its own Lighthouse environment and may also show Chrome User Experience Report field data from real visitors. We link to it beside each result so the two sources can be compared without pretending they are identical.