36 real websites, scored by the same 28 checks

Each site is fetched once, as a browser would, and the returned HTML is checked against 28 fixed rules across performance readiness, structure, trust signals, discoverability and mobile handling. The score is simply the proportion that passed. There is no model and no opinion in it.

Because the rules are fixed, two people auditing the same page get the same score. That is what makes these numbers checkable rather than merely stated — and it is why every finding in a report quotes the evidence it was based on, so you can search for it in your own source.

Scores drift as sites are rebuilt, which is why the measurement date ships alongside the data. A score here is a reading taken on 2026-07-31, not a permanent property of the site.

A high score does not mean a site is beautiful, useful or successful, and a low one does not mean the opposite. It measures what the page hands a visitor and a crawler on arrival — nothing more.