Should you really use rel=canonical across different websites?

Is cross-domain canonical a directive or a suggestion?

Contrary to what many believe, the rel=canonical tag is not an imperative directive but an indicative signal. Google consults it but may decide to ignore it if its algorithms detect inconsistencies or deem that another version deserves more visibility.

In the case of a canonical pointing to another domain, this recommendation becomes even more fragile. Google systematically checks the consistency of the signal: does the target site confirm this relationship? Are the contents truly identical? Is there a legitimate editorial relationship between the two sites?

Why does Google prioritize this approach for syndicated content?

Syndication of content presents a massive duplication issue that Google must resolve to avoid polluting its index. When an article is republished on 15 partner sites, the algorithm must choose which version to display in the SERPs.

Cross-domain canonical theoretically allows the secondary publisher to signal: "This content exists elsewhere in its original version, please prioritize this source." But this logic mostly works when the canonical site has more authority than the republishing site. If the reverse is true, Google may ignore the signal.

What are the alternatives when canonical is not enough?

In the face of duplicate content between independent domains, several options coexist: the canonical tag, the meta robots noindex on the secondary version, or the addition of a unique paragraph to differentiate the versions.

The choice depends on the business intent. If site B absolutely wants to rank with this syndicated content, placing a canonical to site A is counterproductive. If site B just seeks to enrich its catalog without cannibalizing the source, the canonical becomes relevant but remains a gamble.

• Cross-domain canonical is a weak signal that Google may ignore if the sites do not have a clear editorial relationship.
• Google naturally favors sources with more authority, even without an explicit canonical.
• The tag must be reciprocal in its logic: if site A canonicalizes to B, but B no longer exists or redirects elsewhere, Google rejects the signal.
• A poorly configured canonical can render a page invisible by transferring all its ranking potential to a URL that should not be prioritized.
• Modern CMSs sometimes generate automatic canonicals that create loops or conflicts between affiliated domains.

Is this recommendation consistent with field observations?

On paper, yes. In reality, cross-domain canonical mainly works between sites owned by the same entity or in clear editorial relationships (news agency, official syndication). Between independent sites, the signal is often ignored.

I have seen cases where site B placed a canonical to a more authoritative site A, yet Google continued to rank the page from site B. Why? Because site B had a better backlink profile for that specific query, or because the search intent matched better with the context of site B. Google does not follow blindly.

What nuances should be added to this official directive?

Google says, "use canonical to indicate the preferred version," but does not clarify what happens when both sites want to rank. This is the typical case with shared content platforms (Medium, LinkedIn Articles, etc.): the author publishes on their blog AND on Medium. Who should canonicalize to whom?

The real unspoken rule: cross-domain canonical primarily works in a downward direction (from the weaker site to the stronger site). If you are a small blog and your article is picked up by a major media outlet, canonicalizing to the media will not change anything: Google would rank it anyway. But if the media places a canonical to you, that could help.

In what cases does this rule not apply?

When the content is not strictly duplicated. If site B republishes an article from site A but adds an intro, editorial comments, or local context, Google may legitimately consider that these are two distinct pieces of content. The canonical then becomes counterproductive.

Another problematic case: e-commerce sites sharing product descriptions through feeds. A retailer that canonicalizes to the manufacturer could shoot themselves in the foot by giving up the chance to rank for queries where they could differentiate themselves (fast delivery, local customer service, etc.). [To verify] if Google really takes the canonical into account in these multi-brand configurations.

What should you concretely do when syndicating content?

If you are the source publisher (the original creator of the content), ensure that the sites republishing your content place a canonical to your URL. It is rare that they do so spontaneously. Include this clause in your syndication agreement.

If you are the site republishing, place a canonical only if you have no intention of ranking with this content. Otherwise, differentiate your version: add local intro, expert opinions, or a specific angle. Google rewards added value, not slavish copying.

What mistakes should be absolutely avoided with rel=canonical?

Classic mistake: placing a cross-domain canonical without verifying that the target URL is indeed indexable and accessible. If your canonical points to a page in noindex, 404, or behind a paywall, Google ignores the signal and you lose all control.

Another trap: canonicalizing all pages in a category to a "template" on another site. Google detects these automated patterns and massively ignores them. The canonical should remain an editorial decision page by page, not a global CMS setting.

How to check that the canonical is being acknowledged correctly?

Use Google Search Console to see which URL Google considers as canonical. If you have placed a cross-domain canonical but GSC shows your local URL as canonical, it means Google has ignored your signal.

Also monitor the positions in the SERPs. If your competitor ranks with content that you canonicalized to them, that’s normal. But if you are ranking despite the canonical, Google has not followed your recommendation. You then need to understand why: insufficient authority of the target, signal inconsistency, or added value detected in your version.

• Verify that the target canonical URL is indeed indexable (no noindex, robots.txt, or 404)
• Ensure that both contents are truly identical, otherwise Google may reject the canonical
• Check in Google Search Console which URL is retained as canonical by Google
• Avoid cascading canonicals (A to B to C): Google only follows the first level
• Never canonicalize a high-performing page to a less authoritative page without a clear strategic reason
• Test on a few pages before massively deploying a cross-domain canonical