How the SEOuch score works
Every audit runs the same checks in the same order and scores them the same way. This page documents the weights so a score means the same thing on every site and so you can disagree with us on specifics.
Category weights
| Category | Weight | What it covers |
|---|---|---|
| Technical & crawlability | 20% | Status codes, redirects, HTTPS, robots.txt, meta robots, canonical tags, sitemap, hreflang, viewport, language and charset. |
| On-page | 20% | Title and description quality, heading structure, image alt text, link profile, social tags, sampled broken links. |
| Content & structure | 10% | Text extractable without JavaScript, thin-content signals, answers under question headings, lists and tables, visible dates. |
| Performance | 20% | Core Web Vitals from the Chrome UX Report where available, plus Lighthouse lab scores and the top opportunities. |
| Security | 10% | Certificate validity, HSTS, Content-Security-Policy, X-Content-Type-Options, frame controls, referrer policy, mixed content. |
| AI visibility | 20% | AI crawler access rules, fetch-as-bot comparison, snippet eligibility, structured data for entities and authorship, citability signals, llms.txt and related files (informational). |
How a check is scored
Each check is one of pass, notice, warning or error, with a point value that reflects how much the issue usually matters. A category score is the points earned divided by the points available in that category. The overall score is the weighted sum of the category scores. Checks that cannot run (for example when a site blocks our fetcher) are excluded from the denominator rather than counted as failures.
Score bands
| Score | Label | Meaning |
|---|---|---|
| 90–100 | Healthy | Nothing urgent. Keep an eye on performance and freshness. |
| 70–89 | Minor aches | A handful of fixes, usually on-page or structured data. |
| 50–69 | Sore | Real problems that limit crawling, speed or AI eligibility. |
| 0–49 | Ouch | Fundamental issues. Start with the errors in the fix list. |
What the AI visibility category does and does not measure
The category measures whether a page is eligible to be read, indexed and quoted by AI search systems: crawler access rules for search and user-triggered bots, whether the page can be fetched with those user agents, snippet eligibility, extractable content, authorship and entity markup, and citability signals such as sourced statistics. It does not measure whether a brand is actually cited; that depends on off-site mentions and is covered by the separate AI visibility checker.
Google says no special files or markup are required for AI features, and that llms.txt is not used by Google Search. We check llms.txt and similar files for completeness and score them as informational only.
Data sources
- Our own fetch of the page, robots.txt, sitemap and related files.
- Google PageSpeed Insights and the Chrome UX Report for performance data.
- A commercial backlink index for authority and backlink data in the full report.
Scores are automated and informational. They are not endorsed by Google or any AI provider.