Not by design, but effectively yes in practice. Google’s documentation confirms cross-domain canonicals are supported and can be fully honored, so there’s no rule that discounts them mechanically. What actually happens is that Google’s threshold for trusting a cross-domain canonical tends to be higher than for a same-domain one, because getting a cross-domain consolidation wrong has larger consequences (transferring signals between two unrelated entities) than getting a same-domain one wrong (which stays contained within a single site you control). The honest answer sits between “flatly weaker” and “identical treatment”: the mechanism is the same, but the practical bar for Google acting on it is not.
Why cross-domain and same-domain canonicals aren’t evaluated identically in practice
Google’s Search Central documentation on cross-domain canonicalization states plainly that a canonical tag pointing to a different domain is a supported pattern, used for cases like syndicated content, migrated domains that haven’t fully redirected, or multiple domains serving near-identical catalogs. The documentation does not say cross-domain canonicals carry a reduced weight relative to same-domain ones as a matter of algorithmic policy.
What it does say, and what practitioners consistently observe, is that Google evaluates whether the cross-domain relationship “makes sense” given other available evidence. That’s the operative distinction. A same-domain canonical is inherently lower-risk to honor: both URLs belong to the same site, controlled by the same owner, so consolidating signals from one to the other doesn’t create any cross-entity ambiguity about who should receive credit. A cross-domain canonical asks Google to move ranking signals from Site A to Site B, two different entities that may have different owners, different trust histories, and different reasons someone might want to exploit consolidation (scraped content pointing a canonical at the original to disguise duplication is the abuse case Google has to guard against).
Because of that asymmetry in downside risk, Google’s systems appear to require corroborating evidence before fully honoring a cross-domain canonical: matching or near-identical content between the two URLs, consistent internal and external linking patterns that support the claimed relationship, and no contradictory signals suggesting the two domains are actually competing rather than cooperating. When that corroboration is present (a true syndication partnership, a domain migration in progress, a multi-regional mirror setup), Google has shown it will honor cross-domain canonicals fully, consolidating indexing and ranking signals as directed. When the corroboration is thin or contradictory, the same tag is more likely to be treated as a hint that gets overridden, exactly the way a same-domain canonical would be overridden if backlinks or content similarity disagreed with it, just with a lower bar for Google to decide “this doesn’t add up.”
What this means for implementation
Don’t treat a cross-domain canonical as inherently underpowered and compensate by stacking redundant signals defensively; that’s the wrong mental model and can create its own confusion. Instead, treat it as a signal that needs the same supporting evidence a skeptical reader would want: make sure the content on both URLs is genuinely equivalent (not just similar), make sure any 301 redirects or hreflang annotations you’re also running are consistent with the canonical claim rather than contradicting it, and make sure there’s a plausible, discoverable reason for the relationship (a stated syndication agreement, matching business registration/branding across domains, or clear historical continuity if it’s a migration).
If you’re monitoring whether a cross-domain canonical is being honored, the same URL Inspection tool in Search Console that surfaces same-domain canonical mismatches works here too: check the “Google-selected canonical” field on the non-preferred domain’s URL. If Google is still indexing that URL independently rather than consolidating it to your declared canonical, that’s your signal the corroborating evidence isn’t strong enough yet, not that cross-domain canonicals don’t work.
There is no published figure from Google quantifying a reduced weight for cross-domain signals, and claiming one exists would be fabricating precision Google has never provided. What’s verifiable is the documented behavior: the mechanism is identical, the supported-use-cases language is unambiguous, and the practical difference comes from Google requiring more corroborating context before it will act on the cross-domain version of the same hint. That’s a real distinction, but it’s a matter of evidentiary threshold, not signal strength.
Hypothetically, imagine a publisher, “Example News,” syndicates an article to a partner site, “Example Aggregator,” and sets a cross-domain canonical pointing back to the original. Let’s say the content matches closely, the syndication relationship is referenced on both sites, and internal linking patterns are consistent with a genuine partnership. In this hypothetical, Google would have the corroborating evidence needed to honor the canonical and consolidate signals back to Example News. Now imagine a second hypothetical: a scraper site sets a canonical pointing at Example News’s article without any real relationship, matching content but no linking pattern, no disclosed partnership, and contradictory backlink signals. In that case, the same canonical tag would be far more likely to get overridden, not because cross-domain canonicals are inherently weak, but because the corroborating evidence needed to trust it simply isn’t there.
