Methodology
How we score
Every audit runs 25+ checks against a single snapshot of your page and rolls them into one score out of 100. Here is exactly how that number is built — and the one part we keep private, so scores stay hard to game.
GEO — AI readiness · 40%
Can AI answer engines like ChatGPT, Claude, and Perplexity find, understand, and cite this page? Answerability, quotability, llms.txt, AI-crawler access, structured data, freshness.
SEO fundamentals · 35%
The on-page basics that still decide classic-search visibility: titles, descriptions, headings, canonicals, indexability, social tags, links.
Technical · 25%
The delivery plumbing: HTTPS and HSTS, security headers, mobile viewport, compression, page weight — what lets engines reach you at all.
What a single check earns
In short: pass earns full points, improve earns half, fail earns none.
Each check carries a weight reflecting how much it matters. A pass earns the full weight, an improve earns half, and a fail earns nothing. Informational findings and skipped checks are excluded entirely — they never drag a score down for something we couldn't or didn't measure.
How the overall score is blended
In short: each category is normalized to 0–100, then blended GEO 40 / SEO 35 / Technical 25.
Points earned in a category are divided by the points possible, giving each category a score out of 100. The overall score is the weighted blend of the three — GEO carries the most weight because being readable and citable by AI engines is what we believe decides the next decade of visibility. If a category couldn't run at all, the remaining categories are re-weighted rather than treating the gap as a zero. And when a page can't be read at all, nothing is scored — you'll see "not scored" rather than a misleading number.
Reading the bands
In short: 80+ is strong, 50–79 needs work, below 50 needs attention.
The bands are calibrated so that a well-maintained site lands in the 70s–80s with a clear, short fix list. A perfect 100 is rare by design — it means every weighted check on the page passed outright.
What we deliberately keep private
In short: the per-check weights and detection heuristics — so scores measure quality, not gaming.
We publish the categories, the category weights, and what every check looks for (each finding on a report explains itself and links to a full guide). We do not publish the exact weight of each individual check or the thresholds and heuristics behind pass, improve, and fail. Publishing those would invite optimizing for our scoreboard instead of for real visibility — the same reason search engines keep their ranking factors close.
What we promise instead: every audit is stamped with the engine version that produced it, scoring changes ship in versioned updates, and re-running the same unchanged page on the same engine version yields the same score.
See it in action on your own site.
Run a free audit →