Is it true that Google really respects the canonical tag?

Why doesn't Google always follow the canonical tag?

The canonical tag acts as one signal among many within Google's algorithm. Unlike a strict technical directive, it does not impose anything on the engine.

Google makes its own decision by cross-referencing multiple sources: the tag itself, as well as inbound link architecture, content consistency across versions, URL structure, XML sitemaps, and even detected user preferences. If these signals contradict each other, Google decides — and not necessarily in your favor.

What signals could contradict your canonical tag?

The most common case: you point to version A, but the majority of external backlinks target version B. Google may then consider B to be the 'real' version in the eyes of the web and ignore your tag.

Another classic scenario: significant content variations between URLs. If the URL you declare as canonical differs significantly from the one Google crawls, the engine may decide they are not really equivalent contents — and choose its own canonical version.

Does this logic apply to all types of sites?

Yes, without exception. E-commerce sites with product variants, multilingual sites using hreflang, content platforms with parameterized URLs — all are affected.

The technical complexity of the site simply amplifies the risk of divergence between your intentions and Google's interpretation. The more pages, URL parameters, and mobile/desktop versions you have, the more contradictory signals multiply.

• The canonical tag is a signal among others, not a strict technical command.
• Google cross-checks this indication with backlinks, actual content, URL structure, and other factors.
• Complex sites (e-commerce, multilingual) are particularly exposed to interpretational divergences.
• A perfect implementation of the tag does not guarantee that Google will respect it.
• The engine can change its mind on the canonical URL over time based on the evolution of external signals.

Is this statement consistent with real-world observations?

Ableitly. We regularly see documented cases where Google indexes a different URL than the one declared as canonical. SEO forums are filled with examples — particularly on e-commerce sites where Google prefers a product variant over the generic listing.

What is less clear, however, is the relative weight of each signal. Google does not publish any weighting: is it 60% backlinks, 30% canonical, 10% structure? Impossible to know. [To be verified] — the official documentation remains vague about the precise tiebreaking criteria.

What nuances should be applied to this rule?

The first nuance: on well-structured sites with consistent signals, Google generally follows the canonical tag. The problem mainly arises when signals contradict each other.

The second point: Martin Splitt does not specify if certain types of signals have priority. For example, does a canonical in the HTML code carry more weight than a canonical HTTP header? What about the XML sitemap? Again, [To be verified] — tests show varying results depending on contexts.

In what scenarios does this logic pose the most problems?

Site migrations are the typical nightmare. You redirect the old domain to the new one, set clean canonicals, but Google continues to index the old URL for weeks — because the backlinks are still heavily pointing there.

Another critical case: sites with filters and facets. You canonicalize to the main category, but if a filtered URL receives direct links (forums, social media), Google might decide it deserves its own canonical status. Result: unintentional cannibalization between your own pages.

What concrete steps should you take to maximize compliance with the canonical tag?

Step 1: audit your backlinks. Use Ahrefs, Majestic, or Search Console to identify where incoming links actually point. If a non-canonical URL receives more links than your canonical version, that's a red flag.

Step 2: check the content consistency between variants. If you canonicalize A to B, but the content differs by more than 20-30%, Google may legitimately determine that they are not duplicates. Harmonize or create separate pages.

What mistakes should you absolutely avoid?

Never set a canonical to a page that returns a 404 or 301. It may seem obvious, but we still see implementations where the canonical points to a redirected URL. Google ignores this directive.

Avoid canonical chains: A canonicalized to B, which itself canonicalizes to C. Google could get lost — or simply ignore the chain and choose its own version. One step only, always.

How can you check that Google has understood your intention?

Search Console, Coverage tab, then inspect the URL. Google explicitly indicates which URL it considers user-defined canonical and which URL it has actually selected. If these two diverge, you have a problem.

Supplement with a Screaming Frog or Oncrawl crawl to detect orphan canonicals, loops, or indexed pages despite an outgoing canonical. These inconsistencies are all contradictory signals for Google.

• Regularly audit backlinks to detect non-canonical URLs that receive links.
• Check content consistency between variants before setting a canonical.
• Avoid canonical chains and canonicals pointing to redirected or erroneous URLs.
• Use Search Console to compare the declared canonical URL vs. the canonical URL selected by Google.
• Crawl your site to detect orphan canonicals and loops.
• If you are migrating a site, focus your efforts on redirecting backlinks to the new canonical URLs.