Google Choosing www as Canonical: Fix Non-www Indexing
A host-level diagnostic for fixing cases where Google keeps the www domain indexed after a site moves to the non-www version.
· 14 min read
A site owner moved a four-page website from www to non-www, set up 301 redirects, and still found the old www pages indexed while the new host was absent. When Google choosing www as canonical is the symptom, the practical payoff is a repeatable way to determine whether this is normal consolidation, a broken hostname migration, or an unrelated indexing block.
The key distinction is simple: www.example.com and example.com are separate hosts. A 301 redirect from one to the other is a strong canonical signal, but Google evaluates that signal alongside the rendered canonical tag, internal links, XML sitemap URLs, HTTPS handling, crawlability, and its own duplicate-content assessment.
Start by separating a normal canonical state from a migration failure
Google selects one representative URL from a duplicate cluster. Its documentation calls this canonicalization, and it can choose a different URL than the one a site declares. That alone is not necessarily an error: seeing an old www URL in search for a period after a hostname move does not prove that the non-www version cannot be indexed.
The original BigSEO case described a site with only four pages, a move from www to non-www, and 301 redirects already in place. For a site that small, a full host-level review should be quick. The important question is not “Has Google removed every old URL yet?” It is: does every old URL consistently lead Google to the intended new equivalent?
Treat this as a real migration failure when one or more of these conditions applies:
- The non-www URL is meant to be primary, but URL Inspection reports the
wwwURL as the Google-selected canonical long after all signals were changed. - Some
wwwpages return200 OKrather than redirecting, or redirect to a generic homepage instead of the matching non-www page. - The non-www URL is excluded for a reason such as
noindex, robots blocking, a server error, or “Crawled - currently not indexed.” - The site sends mixed signals: for example, a non-www page self-canonicalizes to
www, while the XML sitemap and navigation point elsewhere.
A normal duplicate consolidation may show “Alternate page with proper canonical tag” for the old host. That status commonly means Google recognizes the duplicate and indexes the selected canonical instead. It becomes a problem only if the selected canonical contradicts the intended hostname or the intended non-www URL is itself not indexable.
Test all four HTTP and hostname variants first
A www-to-non-www move has four starting points, not two. Each must be tested for the homepage and for representative internal URLs such as /services/ or /contact/.
| Starting URL | Expected result when non-www HTTPS is primary |
|---|---|
http://www.example.com/page/ | One permanent redirect to https://example.com/page/ |
https://www.example.com/page/ | One permanent redirect to https://example.com/page/ |
http://example.com/page/ | One permanent redirect to https://example.com/page/ |
https://example.com/page/ | 200 OK, indexable page, self-referencing canonical |
The ideal result is one hop. For example, http://www.example.com/about/ should not move through three locations such as HTTP www → HTTPS www → HTTP non-www → HTTPS non-www. Google can follow redirects, but extra hops make debugging harder and increase the chance that one route fails, loops, drops a path, or produces inconsistent headers.
Check the response, not just the browser address bar
A browser landing on the right page is not sufficient evidence. Test the HTTP status code and Location header with a crawler, command-line request, browser developer tools, or a local audit application. Confirm these details:
- The old host returns a permanent redirect, normally HTTP
301or308, rather than a JavaScript redirect or meta refresh. - Every redirect preserves the path and meaningful query parameters where appropriate.
- The final non-www page returns
200 OKand does not redirect back towww. - The canonical target is a live, indexable
200 OKURL. - The old hostname is not blocked in
robots.txtbefore Google can discover the redirect.
Google describes redirects as a strong canonicalization signal and specifically recommends choosing a preferred destination when a page is available through several URLs. A permanent redirect is therefore the primary mechanism for a hostname move; canonical tags are supporting evidence, not a substitute for redirecting a retired duplicate host.
Inspect the actual canonical delivered on the non-www page
On every intended primary page, the HTML should contain a self-referencing canonical, such as:
<link rel="canonical" href="https://example.com/services/">
This must be the canonical present in the rendered document Google can access, not merely a setting shown in a CMS dashboard. View page source, inspect the rendered DOM if JavaScript controls metadata, and crawl a sample of pages. A common migration mistake is carrying over an old base URL, so the live non-www page still outputs https://www.example.com/services/ in rel="canonical".
Google treats rel="canonical" as a strong signal, but not an absolute command. It may ignore a declaration if the target is unavailable, unsuitable as a canonical, contradictory to stronger signals, or materially different from the source page. A canonical is for duplicate or very similar content; it should not point unrelated pages to the homepage merely to force consolidation.
Avoid canonical chains and conflicting instructions
The canonical should point directly to the final preferred URL. Do not create a sequence such as:
https://www.example.com/page/canonicalizes tohttp://example.com/page/http://example.com/page/redirects tohttps://example.com/page/https://example.com/page/canonicalizes back towww
That is three different signals across the same content. Instead, redirect every non-preferred variant directly to the final HTTPS non-www page, then place the final URL in its own canonical tag.
Also inspect meta robots and HTTP headers. A valid non-www canonical cannot be indexed if the page has noindex, returns an X-Robots-Tag: noindex, is inaccessible to Googlebot, or produces a persistent 5xx server response. Canonicalization, crawling, and indexing are connected but distinct diagnoses.
Use Google Search Console to compare Google’s view with the site’s view
Google Search Console’s URL Inspection tool is the decisive page-level check. Inspect both forms of the same URL, beginning with the non-www URL intended to rank and then the corresponding www URL.
Under Page indexing, record these fields for each test:
| Search Console field | What it should show in a healthy non-www migration |
|---|---|
| URL is on Google | Usually yes for the final non-www URL after processing |
| User-declared canonical | The non-www HTTPS URL on the final page |
| Google-selected canonical | The same non-www HTTPS URL |
| Crawl allowed? | Yes for the final URL |
| Page fetch | Successful |
| Indexing allowed? | Yes; no noindex directive |
For the old www URL, Search Console may show it as a redirect, an alternate, or not indexed. That is expected if Google has processed the migration. The critical mismatch is when the non-www URL says “Duplicate, Google chose different canonical than user” and identifies www as Google-selected canonical.
The live test is useful for checking the page Google can fetch now, while the indexed result describes the version Google previously processed. They can differ after a deployment. If live testing finds a corrected redirect or canonical but the indexed report still shows the old state, the technical fix may be complete even though Google has not recrawled and reprocessed it yet.
For wider visibility, verify both hostname properties in Search Console. Google’s own crawling and indexing guidance explicitly advises site owners to have both www and non-www versions verified. A Domain property is also useful for consolidated reporting, but URL-prefix properties can make it easier to inspect exact protocol-and-host variants.
Reconcile every supporting signal with non-www HTTPS
Google does not select canonicals from one clue alone. It considers signals including redirects, canonical annotations, sitemap inclusion, and HTTPS preference. When the non-www host is intended to win, every controllable signal should agree.
Internal links and navigation
Update templates, menu links, footer links, breadcrumbs, pagination, in-content links, structured data URLs, Open Graph tags, and absolute image or script references where they influence page discovery. A navigation menu that still links to https://www.example.com/ repeatedly tells crawlers that the old host remains a central site address.
Google uses crawlable links both to discover pages and as a relevance signal. A site does not need every link to be absolute, but no internal link should deliberately send users and crawlers to a hostname that is supposed to be retired. For a systematic review, a local crawl can list every internal URL and reveal whether www links still appear in templates or isolated legacy content.
XML sitemap, hreflang, and feeds
Submit an XML sitemap containing only canonical non-www HTTPS URLs. Remove old-host URLs rather than submitting both versions “for coverage.” Sitemap inclusion is a weaker canonical hint than redirects and rel="canonical", but contradictory sitemap entries create avoidable ambiguity.
If the site uses hreflang, each language or regional URL must reference the equivalent non-www HTTPS URL, and reciprocal annotations must match. A migration can otherwise look correct in English while an old www URL remains in a French, Canadian, or Spanish alternate set.
This same rule applies to RSS feeds, product feeds, schema markup, canonical URL fields in CMS plugins, and cached environment variables. One stale setting does not always make Google choose www, but several stale signals can outweigh a partially implemented migration.
Diagnose the symptom before applying a fix
The following table prevents the common mistake of treating every indexing report as a canonical issue.
| Symptom | Likely cause | Correct next action |
|---|---|---|
www is indexed and non-www redirects correctly | Google has not fully processed the host migration yet | Confirm all signals; monitor both hosts rather than undoing the redirect |
| Non-www reports “Alternate page with proper canonical tag” | Its canonical points elsewhere, or Google sees it as duplicate | Inspect user-declared and Google-selected canonical; correct the non-www self-canonical |
Non-www reports noindex | CMS, plugin, staging setting, or HTTP header blocks indexing | Remove noindex, retest live URL, then request indexing |
| Non-www is blocked by robots.txt | Crawl restriction prevents normal processing | Allow crawling of the final public URLs; do not block the old redirecting host prematurely |
A www URL returns 200 OK | Redirect coverage is incomplete | Configure server-level redirect rules for all paths and protocols |
Google chooses a different page, not merely www | Pages may be too similar or canonical points to an unsuitable URL | Make pages meaningfully distinct or fix the canonical target |
| Non-www has server errors | Hosting, TLS, CDN, origin, or application failure | Resolve the 5xx, SSL, or timeout issue before focusing on canonical tags |
This is why the advice to “just submit the new URL” is incomplete. A request for indexing may prompt recrawling, but it does not repair a redirect loop, stale canonical, blocked page, or inconsistent sitemap.
Apply the fix in a controlled order
Hostname migrations are easier to validate when changes are staged logically. Google’s site-move documentation recommends preparing URL mappings, updating URL details, configuring redirects, updating internal links, and monitoring the move. For a small www-to-non-www move, the practical sequence is:
- Choose the sole preferred version: normally
https://example.com/. - Create a one-to-one map from every old
wwwURL to its non-www equivalent. - Configure server or CDN redirects across HTTP and HTTPS, preserving paths.
- Confirm the final non-www URLs return
200 OKand have self-referencing canonicals. - Replace old-host internal links, sitemap entries, hreflang annotations, feeds, and structured-data URLs.
- Check robots directives,
noindex, server errors, and TLS coverage for the final host. - Inspect representative URLs in Search Console and submit the updated sitemap.
- Request indexing only for a few priority final URLs after the technical evidence is clean.
For agencies managing more than a handful of pages, a local-first audit is useful before requesting any recrawl. Audra can crawl the accessible site from a desktop environment to identify redirect chains, canonical mismatches, non-200 responses, internal links to the old host, and technical issues that might complicate the migration. That makes the Search Console review a verification exercise rather than a hunt through isolated URLs.
Teams that need a wider post-migration checklist may also find the guidance in this domain migration indexing triage guide useful, particularly when a hostname change coincides with a redesign or platform move.
Verify after recrawling without relying on a fixed timeline
Google does not publish a guaranteed number of days for a canonical or hostname change to appear in its index. Processing time varies with crawl frequency, site size, server reliability, link discovery, and the scale of concurrent changes. A four-page site may appear straightforward, but even a small site can be delayed when Google encounters inconsistent technical signals.
Verification should focus on evidence, not a calendar promise:
- Re-run the four-variant redirect test after each deployment.
- Inspect the homepage and several important internal URLs in Search Console.
- Compare user-declared and Google-selected canonical fields.
- Review the Page indexing report for
noindex, crawl errors, server errors, and duplicate statuses. - Confirm the XML sitemap contains only final non-www URLs.
- Monitor impressions and clicks for the final hostname once the URLs are indexed.
Do not use the site: operator as the only acceptance test. It can provide a rough spot check, but URL Inspection and server-level tests offer much stronger diagnostic evidence. If a page is indexed but does not receive impressions afterward, the issue has moved beyond basic canonical selection; page indexed but no impressions requires a different review of query relevance, competition, content, and search appearance.
Keep canonical work distinct from SEO and AI visibility work
Canonical consolidation determines which duplicate URL Google is most likely to index. It does not guarantee rankings, traffic, inclusion in an AI-generated response, or strong performance on technical quality metrics. Once the correct hostname is indexable, the next work includes content quality, crawlable architecture, page experience, accessibility, structured data where appropriate, and link equity.
That separation matters for reporting. An agency can accurately state that a non-www migration is technically sound when all four variants resolve correctly and Search Console confirms the final canonical, while still identifying separate issues such as slow Largest Contentful Paint, weak internal linking, or poor answer-engine visibility.
For a broader explanation of what remains after canonical repair, see AEO vs SEO: a practical guide to AI and Google visibility. The migration is infrastructure work; earning visibility is the next phase.
FAQ
Why is Google choosing the www version as canonical instead of the non-www URL?
Google may see stronger or more consistent signals for www: a live 200 OK response, self-canonicals pointing there, internal links, sitemap URLs, hreflang entries, or historical links. It can also be a temporary processing state after a move. Test all HTTP/HTTPS variants and compare the user-declared and Google-selected canonical fields in URL Inspection before changing anything.
How can a site owner tell whether Google is ignoring a canonical tag or selecting a different canonical?
Inspect the final non-www URL in Google Search Console. Under Page indexing, compare User-declared canonical with Google-selected canonical. If they match, Google accepted the preference. If they differ, inspect the selected URL, the redirect path, rendered rel="canonical", sitemap inclusion, internal links, and whether the declared target is a valid indexable 200 OK page.
Should a site redirect www to non-www, or use canonical tags on both versions?
For a deliberate www-to-non-www hostname move, use permanent one-to-one redirects from www to non-www. The final non-www page should use a self-referencing canonical. Canonical tags support consolidation, but they should not replace redirects when both hostnames are publicly reachable versions of the same site. Avoid leaving the www host available as a normal 200 OK duplicate.
How long does Google take to index the new non-www version after a canonical or redirect change?
There is no reliable fixed timeline. Google states that crawling and indexing take time and does not guarantee when a URL will be crawled or indexed. The productive approach is to verify the redirect, canonical, crawlability, XML sitemap, and internal links first, then use URL Inspection to see when Google’s indexed information updates.
Can inconsistent sitemaps, internal links, redirects, or HTTPS versions cause Google to choose www as canonical?
Yes. Google evaluates several canonicalization signals, including redirects, canonical annotations, sitemap inclusion, and HTTPS. If the site redirects to non-www but its navigation, sitemap, hreflang tags, or canonical elements still point to www, it sends conflicting evidence. Aligning every signal on the final HTTPS non-www URL gives Google the clearest migration path.
Sources
- https://www.reddit.com/r/bigseo/comments/1ws33mn/google_choosing_www_version_as_canonical_and_isnt/
- https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls
- https://developers.google.com/search/docs/crawling-indexing/canonicalization
- https://developers.google.com/search/docs/crawling-indexing/301-redirects
- https://support.google.com/webmasters/answer/12482179?hl=en
- https://developers.google.com/search/docs/crawling-indexing/site-move-with-url-changes
- https://developers.google.com/search/help/crawling-index-faq