Indexing problem guide
Google selected a different canonical
A canonical link is a hint about which duplicate or near-duplicate URL should represent a page. Google can ignore that hint when other signals disagree: a redirect points elsewhere, the sitemap lists a different URL, internal links favor another variant, or the content is substantially duplicated. A “Google-selected canonical” result is therefore a consistency problem, not always a broken tag.
Start here
Run the Check canonical signals before changing templates. It checks what a public crawler can see; Google Search Console remains authoritative for Google’s own indexing decision.
Compare every canonical signal
Inspect the canonical in the raw HTML, not only after JavaScript runs. It should be an absolute, publicly fetchable URL on the preferred host and protocol. Then compare it with the final URL after redirects, the URL in the XML sitemap, Open Graph and social links, hreflang alternates, and the links your own navigation uses.
Common mismatches include a trailing-slash variant, HTTP versus HTTPS, www versus the apex host, tracking parameters, and a CMS template that emits the homepage as the canonical for every page in a collection.
When Google is right to ignore it
If two URLs have materially different intent, do not canonicalize one to the other just to consolidate signals. A product page and a category page, or a language variant and its English counterpart, may need separate indexable URLs with correct hreflang rather than a forced canonical.
Google may also select a different canonical when the declared target returns an error, is blocked, redirects, carries noindex, or has little relation to the source content. A self-canonical on a page with no internal links and thin copy is not a guarantee of inclusion.
Repair the cluster, not just one tag
- Choose one preferred URL format and make redirects, canonicals, sitemap entries and internal links use it consistently.
- Make the canonical target return 200, be crawlable and contain the content you want indexed. Remove accidental noindex or authentication on that target.
- Update templates so the rule is generated from the route’s real public URL, including locale, pagination and product identifiers.
- Inspect the URL in Search Console after Google recrawls it. External tools can spot contradictions; only Google can report its selected canonical.
Do not add dozens of cross-canonicals to “solve” duplicate content. Consolidate true duplicates, improve unique pages, and make the preferred version the one users and crawlers naturally reach.
Not sure this is your problem? Start from how Google indexing works and what to expect, which covers the whole process and links to every diagnosis from one place.