Does Google really follow your canonical tags or does it decide on its own?

Why doesn't Google always respect my canonical tags?

Google collects a variety of signals to choose the representative URL for a group of duplicate pages. The canonical tag is just one hint among others: 301 redirects, mentions in the sitemap, internal link structure, and URL consistency.

Specifically, if your canonical tag points to URL A, but all your internal links direct to URL B, Google will often favor B. The engine looks for the overall consistency of the site, not blind obedience to an isolated directive.

What constitutes a 'clear preference' for Google?

A clear preference means that all your signals converge on the same URL. The canonical tag, redirects, the XML sitemap, internal links, and URLs shared on social media must unanimously point to the same version.

When your signals contradict each other (canonical pointing to /page-a/, sitemap contains /page-b/, internal links point to /page-c/), Google has to decide for itself. It will choose the URL it considers to be the most representative and accessible, regardless of your supposed preferences.

Does this decision impact my ranking in the results?

No. This is probably the most reassuring aspect of this statement. Google's choice of the canonical URL affects which URL appears in the SERPs, but not the ranking power of the consolidated content.

Google attributes ranking signals (backlinks, authority, engagement) to the canonical version it selects. If your preferred URL differs from Google's, you may simply see an unexpected URL appear in the results, without losing positions.

• Signal consistency: align canonical, redirects, sitemap, and internal linking towards the same URL
• No absolute guarantee: Google reserves the right to choose a URL different from your preference
• No impact on ranking: only the display of the URL in the SERPs changes, not your positioning
• Mandatory verification: use the URL Inspection tool in the Search Console to see the canonical URL retained by Google
• Common divergence cases: HTTP/HTTPS variations, www/non-www, trailing slash, superfluous URL parameters

Is this statement consistent with field observations?

Yes, largely. SEO audits regularly reveal cases where Google ignores canonical tags, especially when the internal linking contradicts the directive. A site that self-canonizes its product pages to /product/ but keeps 80% of its internal links pointing to /product.html will create this confusion.

The fact that this does not impact ranking is also confirmed by experience: sites with partially ignored canonicals maintain their positions, simply with unexpected URLs in SERP. It's more an issue of appearance than performance.

What nuances does Google deliberately omit in this statement?

Google remains deliberately vague about the exact weighting of signals. We do not know if a canonical tag accounts for 30% or 70% in the final decision. [To be verified]: What is the exact priority between HTML canonical and HTTP header canonical?

Another absent point: the consideration delays. When you correct contradictory signals, how long before Google reevaluates its choice? Silence. Observations show that this can take several weeks, if not months, on little crawled sites.

In what cases does this rule become problematic?

E-commerce sites with product variations suffer particularly. A product page available in 12 colors generates 12 almost identical URLs. If canonicalization is not perfect, Google may choose a red variant URL as canonical even though you wanted the blue version in SERP.

Multilingual sites with partially translated content also encounter this issue. Google may prefer an incomplete EN version over a complete FR version if signals (external backlinks, misconfigured hreflang) lead it towards EN.

What should you prioritize auditing on your site?

Start by extracting from the Search Console all URLs for which the canonical URL retained by Google differs from the one you declared. The URL Inspection tool reveals these discrepancies page by page, but for a comprehensive diagnosis, export your index coverage data.

Next, analyze your internal linking: do the links predominantly point to non-canonical URLs? A crawler (Screaming Frog, Oncrawl, Botify) will quickly show you if your internal links sabotage your canonical directives.

How can you strengthen the consistency of your canonical signals?

Align all vectors: HTML canonical tag, HTTP header if applicable, XML sitemap containing only canonical URLs, 301 redirects for outdated variants, and internal links exclusively pointing to the canonical versions.

Check your socially shared URLs and in marketing campaigns as well. If you are generating traffic to non-canonical URLs through Facebook Ads or newsletters, you are sending contradictory signals to Google. Standardize your destination URLs.

What critical errors should you absolutely avoid?

Never create canonical chains: page A canonical to B, which in turn canonical to C. Google often stops at the first jump and ignores the rest. Each duplicate page should point directly to the final version.

Avoid circular canonicals (A to B, B to A) or misconfigured self-referential ones. And most importantly, do not canonical to a URL that is 404 or 301: Google will simply ignore the directive and choose arbitrarily.

• Export canonical discrepancies from the Search Console (Coverage tab)
• Crawl your site to map the internal linking structure and spot links to non-canonical URLs
• Ensure your XML sitemap contains only canonical URLs (no variants)
• Test your canonical tags on a sample of pages using the URL Inspection tool
• Audit your 301/302 redirects to make sure they point to the same URLs as your canonicals
• Clean up superfluous URL parameters (utm_, session_id) via Google Search Console (URL Parameters)