A page names a different page as the main one
In the reportCanonical mismatch
canonical_mismatch · Review item
This appears when a page names another page on your site as its canonical, and that page loads normally but does not name itself as canonical. It can be intended, so it is kept for review. Either let the page name itself, or confirm the other page as the main one by giving it a canonical to itself.
Severity: Low. Review it in your next round of site changes. It is Medium on your homepage and on pages that other pages of your site link to.
Why it matters
A canonical is a strong hint that the named address should be shown in search instead of this one. Google recommends that the canonical page names itself, and that your own links point to the canonical address rather than a duplicate.
What to do
- Look at the finding: it shows the canonical the page declares and the links that point to the page.
- If this page should appear in search on its own, change its canonical to its own address.
- If the other page really is the main one, add a canonical to itself on that page, and point your links to it where practical.
When to leave it as it is: if the page is meant to hand its place to the other page, for example a print or filtered version, and the other page should be the one in search, leave the canonical and let the other page name itself.
What we read
We read the canonical of each crawled page, from the <link rel="canonical"> in its head and its Link header, without running JavaScript. A page is reported when it names exactly one other address on your site, with or without www., and that page answered 200 with HTML directly, without a redirect, carries no noindex, is not blocked by robots.txt, and declares no canonical or more than one.
Other cases are left to their own checks: a named page that redirects, is noindex, blocked or fails; one that names a different page again; a page listed in your sitemap; an https page that names http; and an address that differs only by tracking parameters such as utm_source or gclid. The named page is fetched separately when the crawl did not reach it, up to 20 such pages per audit by default.
Example & evidence
Illustrative example — one page and the page it names, not a live audit.
https://bakery.example/cakes/lemon-drizzle/ is not in the sitemap. The homepage and the cakes page link to it, and it declares:
<link rel="canonical" href="https://bakery.example/cakes/">https://bakery.example/cakes/ answers 200 directly, carries no noindex, is allowed by robots.txt and has no canonical of its own. The lemon cake page is reported, and because it is not in the sitemap and other pages of your site link to it, the finding is Medium.
A possible correction: let the lemon cake page name itself, <link rel="canonical" href="https://bakery.example/cakes/lemon-drizzle/">. If folding it into the cakes page is intended, give /cakes/ a canonical to itself instead.
Confirm & check again
Check for yourself: view the page source and read its rel="canonical", then open the named page and check whether it names itself.
Check again: Full audit — run a new audit of the site. The finding closes when the page is crawled again as a complete HTML page and no longer meets the conditions, for example because it names itself, or when it answers directly with 200 and a file that is not HTML, or starts with a permanent redirect within your site. If the page still falls under this check and names another page on your site, the new audit must hold a current record of that page. A robots.txt block settles the comparison once that record exists, even when the page was never requested because robots.txt forbade it; otherwise the named page needs usable HTML that is not only a frame for scripts. A named page with no record cannot confirm the change. If the page leaves this check, for example because it is now listed in your sitemap, the named page need not be read.
Limits & sources
The check does not compare the two pages' content or know your intent, and it does not show which address Google chose. Canonicals added by JavaScript are not seen, and named pages beyond the 20 fetched separately are not checked.
Confidence: Medium because a canonical that points elsewhere is a sign of a mistake, not proof; the finding records the signals it relied on.
Sources
- Google, How to specify a canonical URL with rel="canonical" and other methods — calls rel="canonical" a strong signal, recommends a canonical on the canonical page itself, and advises linking to the canonical URL. Checked 25 September 2026.
Editorial review: 29 September 2026