These definitions are the ones the product itself uses. The same text appears in the dashboard, in the client report and in the API, so what a stranger reads here is what a client reads in the document you send them.
The list is deliberately not long. A rule earns a place by being detectable without guessing, explainable to somebody with no SEO vocabulary, and worth a developer's afternoon. Rules that stop paying off on a site are retired for that site by the ledger itself.
Critical — something is unreachable or unreadable
We could not read this page
When a site answers an auditor with a bot challenge, nobody checked the page and search engines may meet the same wall. How to let a named crawler through.
Page returns a server error
A page returning a server error teaches search engines to crawl less often and eventually to drop it. How to tell a real outage from a bot-only failure.
Missing page title
With no title tag, Google writes its own headline for your result, usually from whatever text is nearest the top. How to find the template that dropped it.
Redirect loops back on itself
URLs redirecting in a circle never load for anyone: not for visitors, not for search engines. Usually two rules rewriting each other's output.
Link to a page that does not work
A link to a page on your own site that returns an error wastes a visit, wastes crawl budget and passes nothing on. Link to the new URL, not to a redirect.
Warning — costing traffic or clarity
Page not found
A missing page costs you something only while it is still linked or still in the sitemap. When to restore it, when to redirect it, and when to leave it.
Missing meta description
Without a description, Google picks an arbitrary sentence from the page as your snippet. Rarely the one you would have chosen, and rarely the same twice.
No main heading
The main heading tells readers and search engines what a page is about, separately from the title in search results. Missing H1s are almost always a template.
Page is hidden from search results
A noindex instruction tells search engines to show a page to nobody. Often deliberate, and often a staging setting that shipped and went unnoticed for weeks.
Conflicting canonical tags
When a page names one canonical URL in its HTML and another in its headers, Google resolves the disagreement for you, and may not pick the page you wanted.
Very little content
Thin pages rarely rank, and a lot of them can affect how search engines judge the site as a whole. When to expand a page, and when to merge and redirect it.
Images without alt text
Alt text describes an image to screen readers and to search engines, which cannot see pictures. An accessibility failure before it is lost image traffic.
Slow to respond
Over about 1.5 seconds to first byte, search engines crawl less of the site and visitors leave. Response time is not the same problem as page speed scores.
Secure page loading insecure files
A secure page pulling images, scripts or styles over plain HTTP gets them blocked or warned about by browsers, which can visibly break the page for everyone.
Same title on several pages
When pages share a title, search engines cannot tell them apart and nor can the person choosing a result. Usually one template with an unfilled variable.
Same description on several pages
Pages sharing one description produce search results that all read the same, so the snippet stops helping anybody decide which of your pages to open.
Same content on several pages
When several URLs carry effectively the same text, search engines pick one and ignore the rest, and their choice is often not the page you wanted to rank.
Redirect passes through several stops
A URL that redirects to a URL that redirects again is slower for visitors and costs an extra fetch for every crawler. Point the first hop at the destination.
Page nothing links to
A page in the sitemap that nothing links to reads as unimportant to search engines, and cannot be found by a visitor who is browsing rather than searching.
Informational — often correct as it stands
Title is long enough to be cut off
Past roughly 60 characters a title is truncated in search results, so the end is never read. Which part to keep, and why the count is a guide not a rule.
Title is very short
A title under ten characters usually means the page is not describing itself, which wastes the single strongest signal you have about what it is for.
Meta description will be cut off
Descriptions past about 160 characters are cut off in search results, so any reason to click that you put at the end is not read by anybody.
Meta description is very short
A very short description leaves most of the available room in the search result unused, which is free space you already own and are not using.
More than one main heading
Several H1s are not penalised, but they usually mean the page is unclear about its own main subject, which shows up in what search engines choose to show.
Heading levels skip a step
Jumping from H2 to H4 makes a page harder to follow with a screen reader and slightly muddies the structure search engines read out of the document.
Page points at a different URL as the original
A canonical tag pointing elsewhere removes this page from search results by design. Correct for duplicates and print views, expensive when it was unintended.
Page links nowhere
Search engines follow links to discover and weigh pages. A page that links nowhere passes nothing on, and readers who arrive there have nowhere to go next.
No structured data
Structured data labels what a page contains so search engines do not have to guess. Only the types that still produce rich results are worth the implementation.
Buried deep in the site
Pages that far down get crawled less often and are treated as less important. Depth is a signal about how much the site itself seems to value the page.
Try it on your own site, free
No card, and the free tier does not expire — the first verdict takes about six weeks to arrive.