Why is Google confused by your conflicting redirects and canonicals?

What do we mean by 'conflicting signals' in this context?

When you set up a 301 or 302 redirect from page A to page B, you are explicitly telling Google: 'This page A no longer exists, I want you to consider page B instead.' This is a strong, technical signal that is enforced upon crawlers.

At the same time, if you leave a canonical tag on page A pointing to page C (which is different from B), you create a schizophrenic situation. The redirect says 'Go to B,' while the canonical says 'The preferred version is C.' Google is faced with two incompatible instructions from the same site.

The engine then has to arbitrate between two signals of the same hierarchical level, which is not accounted for in its normal operation. Canonicalization algorithms seek consistency: when it is lacking, they switch to 'I do what I can' mode, with results that fall beyond your control.

When do we typically observe this conflict in practice?

The typical scenario: a poorly prepared migration or redesign. You redirect your old URLs to the new ones but forget to remove or modify existing canonicals that pointed elsewhere (to HTTPS versions, www, or other variations).

Another common case: pagination or filtering management systems. You redirect empty or outdated pages to page 1, but the CMS continues to automatically inject a canonical to a URL with parameters. The result is an invisible conflict that undermines your consolidation strategy.

Hybrid configurations between CDN and origin server can also generate this problem. The CDN imposes a redirect for performance reasons, while the origin server still returns a different canonical in the source HTML. Google sees both but does not know which to prioritize.

Why does Google talk about 'unpredictable behaviors' instead of giving a clear rule?

Because Google cannot guarantee which signal will prevail. Depending on the context (history of the URLs, respective authority of target pages, consistency of other signals), the algorithms may decide differently. It’s not simply 'the redirect always wins' or 'the canonical prevails.'

In practice, we observe three possible outcomes: Google follows the redirect and ignores the canonical, Google follows the canonical and ignores the redirect (rare but documented), or Google chooses a third URL it considers the best canonical version based on other signals (internal links, sitemaps, etc.). This last option is the most frustrating for SEOs.

• Absolutely avoid combining redirects and canonicals pointing to different destinations
• Audit technical configurations after each migration or infrastructure change
• Check in Search Console which URL Google actually selected as canonical
• Clean up obsolete canonicals before implementing massive redirects
• Test HTTP headers and the source HTML to detect contradictions before Google discovers them

Is this statement consistent with what we observe on the ground?

Yes, and it’s actually one of the rare instances where Google's official position matches the observations. I have documented dozens of cases where sites lost positions on strategic pages simply because a redirect and a canonical contradicted each other. The detection delay can be long (several weeks of crawling), masking the cause.

The interesting point: Google does not say 'we always follow the redirect' as many SEOs believe. It explicitly admits that the outcome is unpredictable. It’s a polite way of saying: 'Our algorithms are not designed to handle this anomaly, so they are improvising.'

What nuances should we consider regarding this general rule?

First special case: temporary 302 redirects combined with a canonical. Technically, a 302 says 'this page exists elsewhere temporarily,' so a canonical to the original URL might seem consistent. However, Google often ignores this distinction and treats 302s like 301s after a few weeks. The conflict remains unresolved.

Second nuance: cross-domain canonicals. If you redirect domain-A.com to domain-B.com with a 301, but domain-B.com contains a canonical pointing to domain-C.com, Google may interpret this as a legitimate chain of canonicalization. [To be verified] I don’t have solid data showing how well Google handles this case, but the logic suggests it may follow the chain if the domains have a clear relationship (same owner, same theme).

Third point: language variants with hreflang. A geolocated redirect (e.g., .com to .fr for French visitors) combined with a canonical to .com might seem contradictory, but it’s actually a valid configuration if it is accompanied by proper hreflang tags. Google understands that the redirect serves UX, not canonicalization. However, be careful: if your hreflang is misconfigured, the conflict becomes real again.

When does this rule not apply?

Case one: redirecting to the same URL as the canonical. If page-A.html redirects to page-B.html AND contains a canonical pointing to page-B.html, there is no conflict. It’s even redundant but harmless. Google follows the redirect, sees the canonical confirms it, and everything runs smoothly.

Case two: self-referencing canonicals after redirect. If you redirect page-A to page-B, and page-B contains a self-referencing canonical pointing to itself, that’s the normal configuration. The canonical of the source page (A) becomes obsolete since nobody can reach it anymore.

What should be prioritized in an audit of an existing site?

First action: extract all redirected pages that still contain HTML. Use a crawler like Screaming Frog in 'follow redirects' mode to identify source pages that, before redirecting, return a 200 code with a canonical. These ghost pages are your ticking time bombs.

Second check: compare redirect destinations with canonical URLs declared in Search Console. If Google has selected a different URL as canonical than the one you are redirecting to, you probably have a conflict. The 'URL Inspection' tool shows you which URL Google considers canonical for each inspected page.

Third check: verify the HTTP headers Link rel=canonical. Some servers or CDNs add a canonical in HTTP headers in addition to the HTML canonical. If the two do not point to the same URL, or if a redirect points elsewhere, you create a triple contradiction that even Google cannot resolve properly.

How can we prevent these conflicts during a migration?

The golden rule: remove or update all canonicals before implementing redirects. If you are migrating from old-site.com to new-site.com, ensure that all pages from old-site.com either have a canonical pointing to their equivalent on new-site.com or have no canonical at all before launching the redirects.

In practice, this means: carefully map each old URL to its new destination, then automatically generate consistent canonicals in an intermediate phase. Allow Google to crawl this configuration for 1-2 weeks before activating the redirects. This allows the engine to pre-understand the target structure.

For large sites, use progressive section-by-section redirects rather than a big bang. First migrate one category, check in Search Console that Google is correctly following the signals, then continue. If a conflict arises, you detect it on 500 pages instead of 50,000.

What tools can be used to automatically detect these contradictions?

SEO crawlers like Screaming Frog, Sitebulb, or OnCrawl can all identify redirected pages that contain canonicals. Configure them to follow redirects and extract canonicals: any page with 'Status Code 301/302' AND 'Canonical Tag present' deserves investigation.

The Google Search Console remains the ultimate referee. The 'Coverage' report shows you the 'Redirected' pages that are still indexed (a sign of a conflict), and the 'URL Inspection' tool explicitly tells you which URL Google has chosen as canonical and why. If the reason given is not 'Redirect' but 'Canonical tag', you have your diagnosis.

For recurring technical audits, set up monitoring scripts that regularly compare redirect mappings with canonicals extracted from HTML. A simple diff between a redirect file and a complete crawl can reveal discrepancies introduced by CMS updates or interventions from non-SEO teams.

• Crawl the site in 'follow redirects' mode to identify source pages still containing canonicals
• Check in Search Console that the canonical URLs selected by Google match the redirect destinations
• Compare HTTP headers and HTML canonicals to detect contradictions at the server level
• Clean all obsolete canonicals before deploying a redirection plan
• Test the redirects in pre-production with tools like curl or browser extensions displaying headers
• Document the redirect/canonical mapping precisely in a migration tracking table