Guide library · Software Help HQ
Google Selected a Different Canonical: A Safe Troubleshooting Worksheet
Software Help HQ | Search indexing explanation
Short answer: A declared canonical URL is a preference signal, not an instruction Google must obey. If Google chooses another page as canonical, first decide whether its choice actually serves visitors better. Only then investigate competing signals, duplicated content, redirects and incorrect internal links.
A two-address example
Suppose your website serves the same guide at:
https://example.com/how-to-starthttps://example.com/how-to-start.html
You consider the first address the permanent one. Google may discover both and show only one. Before changing anything, record which URL visitors use, which appears in navigation, and whether either redirects. You cannot decide from the filenames alone which URL Google should display.
Capture four separate facts
| Observation | How to gather it | What it tells you |
|---|---|---|
| User-declared canonical | Inspect the page's source or CMS SEO settings | Your preferred version |
| Google-selected canonical | Search Console URL Inspection | Google's current selection for that inspected URL |
| Redirect destination | Open both URLs in a fresh browser or inspect response | Whether one route consolidates to another |
| Internal links and sitemap | Inspect navigation and sitemap entries | Whether your own site tells contradictory stories |
Write the observation date beside each. Google can change its choice after recrawling; an older report may no longer match the current site.
Troubleshoot in this order
1. Is Google's selection reasonable? If a .html address is the one all users and
existing links already reach, it may be better to accept that address rather than introduce risky redirects.
Canonical work is supposed to clarify the visitor's destination—not change its shape for cosmetic reasons.
2. Is the supposedly canonical page complete? Open it anonymously. Verify the full guide is there; confirm it isn't a draft, a placeholder, a noindex page, or a blank JavaScript frame. If it has less useful content than the alternative, improve the page before trying to override Google's choice.
3. Are there mixed signals? A sitemap entry for /how-to-start, canonical tag
pointing to .html, and navigation link pointing to a third redirect route create needless
ambiguity. Use one approved preferred address consistently in future internal links and metadata. Do not
silently rewrite indexed legacy URLs without assessing inbound links and current traffic.
4. Does the site need a redirect? Permanent redirects are a strong canonical signal, but only use them when the alternate route can safely be retired. In a visual builder, publishing or redirect configuration may affect more than one page. Save the current implementation, identify the precise intended route, and verify rollback before changing it.
5. Are pages merely related, not duplicates? Two software guides may discuss the same vendor but answer different user questions—pricing versus a setup tutorial, for example. They should not both point to a canonical that erases a legitimately distinct page. Strengthen their titles, introductions and internal links so readers and search engines can understand the difference.
When to leave a canonical alone
A Google-selected canonical different from the declared one isn't automatically a defect. If the selected page is the strongest accurate page, is accessible, and has appropriate links, an editorial note and future consistency may be better than a production change. The useful measurement is whether the right page is showing for the right query and serving readers, not whether two fields happen to match.
Example: preserving a fragile website
A site-builder project has a live /review.html page, while a redesigned draft is planned for
/review. The administrator is tempted to force a global rule mapping all .html paths
to extensionless paths. That could unintentionally redirect unrelated pages and damage existing search results.
A safer plan is to document the single review route, confirm the CMS's ability to apply one isolated redirect,
and test exactly that route after an approved release. Leave other URLs untouched.
Copyable resolution record
- Inspected URL: ___________________
- Intended primary URL: ___________________
- Google-selected primary URL and inspection date: ___________________
- Both routes return complete, correct content: yes / no / unknown
- Current sitemap URL: ___________________
- Current internal-link URL: ___________________
- Redirect observed: none / permanent / temporary / unknown
- Reader-facing issue: ___________________
- Proposed change and rollback evidence: ___________________
- Recheck date: ___________________
Avoid two common mistakes. Don't use robots.txt to force canonical selection, and
don't add noindex to a useful public duplicate simply because the canonical hints are inconsistent.
Google recommends proper canonical hints for that problem.
Related next step: If both pages are genuinely useful for different search intentions, write down each visitor's question and keep separate article-to-article links rather than collapsing them.
Sources and scope: Google's canonical selection guide; canonical troubleshooting. The two-address website example is hypothetical, not a reported Project HQ production incident.