What each SEO category checks
What each SEO category checks: a plain-language guide to every SEO category we check, from metadata and semantic HTML to performance, plus what it means when performance comes back unmeasured.
The categories, in plain language
The SEO score is made up of several categories. Here's what each one looks at:
- Crawlability: can search engines reach and read your pages (robots rules, sitemap, no blocking errors)?
- Indexability: are your pages eligible to appear in results (canonical tags, no accidental "noindex")?
- Metadata: do your pages have clear, useful titles and descriptions?
- Heading structure: is each page organized with a sensible heading hierarchy?
- Content quality: is there enough clear, relevant content on the page?
- Content relevance: do your pages actually target the topics and keywords you track, and are they easy to read? See content recommendations.
- Images: do images have alt text and load efficiently?
- Internal linking: do your pages link to each other in a helpful way?
- Structured data: do you use markup that helps machines understand your content?
- Semantic HTML: can a machine tell your content apart from your menus and footer? See below.
- OpenGraph: do pages have the tags that make shared links look right on social?
- Performance: do pages load quickly, measured on a phone-like connection? See below.
- Mobile: does the site work well on phones?
- Links: are outbound links healthy, without broken destinations?
A closer look at Semantic HTML
Anything that reads your page automatically (a search engine, a browser's reader mode, an AI assistant answering a question) has to work out which part of the page is the actual content and which part is the wrapper around it. Your menu, cookie notice and footer repeat on every page. Your content does not.
The usual way to mark that boundary is with a handful of standard HTML tags: <main> for the content of the page, and <header>, <nav> and <footer> for the shell around it. When a page uses them, a reader can go straight to what matters. When every part of the page is a generic <div>, there is nothing to aim at, so the reader takes the whole page and your navigation gets treated as if it were content. Pages that are mostly menu then look very similar to each other, and what makes each one distinct gets diluted.
This category checks that each page marks its content with <main>, that the surrounding shell is marked too, and that the content actually sits inside <main> rather than beside it. Pages with very little text (a login screen, for example) are skipped, since there is no content to judge.
It is usually a quick fix: these tags live in your site's layout template, so changing them once normally covers every page at the same time. If swapping the tag is risky, adding role="main", role="banner", role="navigation" or role="contentinfo" to the wrappers you already have works just as well.
Worth being clear about one thing: Google has said that these tags are not a ranking factor on their own. We check them because they change how reliably a machine can pull your content out of the page, which matters more as answer engines quote sites directly.
When Performance shows "not measured"
Performance is the one category that depends on loading your site from the outside, so it is the one that can come back without a score. When that happens the card tells you which of two very different situations you are in.
We could not measure it. The message says performance could not be measured this time and that there is nothing to do on your side. That is exactly what it means: the measurement did not complete, and your score is not affected, because a category we could not measure is left out of the average rather than counted as a zero. Re-running the scan later is enough.
Your site turned the measurement away. This one is worth acting on quickly, and the card says what happened: a firewall or bot protection refused the request, DNS did not resolve, the security certificate was rejected, or your server returned an error. These are not really page-speed problems. Whatever refuses a speed test usually refuses search engine and AI crawlers too, which means the affected pages may be missing from results and from AI answers altogether. That is why it shows up as a high-priority item rather than a quiet note, and why it appears in your quick wins and action plan.
If the cause is a firewall or CDN blocking automated visitors, the fix is the same one described in why AI search bots get blocked. After changing anything, re-run the scan to confirm the page can be read again.
How to use it
Each category contributes to the overall health score. Tackle the weakest categories first, and lean on the quick wins for the fastest gains.