What score should we aim for?

How the number is built

The score is the mean of the scored checks. A pass counts 1, a warning 0.5, a fail 0, and a check that does not apply to the page is left out of the average rather than counted against it. Green starts at 80, amber at 50, red is below that. Because every scored check weighs the same, a fail in Technical costs exactly what a fail in Content costs, and the way up is to work through the fails in whatever order is cheapest to ship.

Only two checks can warn: llms.txt and AI crawler access. Everything else passes or fails, so most of the score moves in whole steps.

Where 86 comes from

Three of the 21 checks look for the OKF bundle, Eleva's own Open Knowledge Format, served at /okf. It is our format, not an industry standard, so a site that does not publish one fails those three and can reach 86 at most with everything else green. That ceiling is fine. We treat 86 with no OKF bundle as a clean page, and we would rather you spent the effort on a fail in Rendering or Crawlability than on those three rows.

What the score is for

It is a readiness check. It tells you whether an engine that reaches this page can fetch it, parse it, and quote it, and it tells you the same thing again after a fix, since every run is kept and the before and after sit as two rows. It does not tell you whether the engine will choose this page over a competitor's; that is what the daily prompt tracking measures. A template that scored 90 can drop to 60 after a redesign nobody mentioned, which is why we re-audit key pages on a cadence rather than once.