Is PageRank still influencing the selection of canonical URLs?

Why is Google still using PageRank for canonicalization?

PageRank has never truly disappeared — it has merely transformed. Google stopped publicly displaying scores in 2016, but the underlying algorithm continues to run behind the scenes. When Gary Illyes mentions its use for canonicalization, he reveals a little-documented aspect: faced with several versions of the same page, Google has to make a decision.

And this choice isn't based solely on the canonical tag or the content. The authority signal conveyed by links — the very essence of PageRank — guides the decision. A URL with a better PageRank score is more likely to be designated as the reference version, the one that will appear in SERPs and accumulate ranking signals.

In what concrete scenarios does this mechanism apply?

Let's take a common case: you have three versions of a product page. One on the main domain, one in a subsection with parameter filters, and a third generated by a mobile variation. No canonical tag is explicitly defined — a classic mistake. Google detects the duplicate content and must designate a canonical.

This is where PageRank comes into play. If your main page receives the majority of external backlinks and has strong internal linking, it will likely be chosen. But if, due to poor architecture, a parameterized URL captures all internal navigation links, it could become the indexed version — even if that’s not your intention.

Does this mean that canonical tags are useless?

Absolutely not. PageRank is one signal among others. Google combines the canonical directive, the structure of XML sitemaps, 301 redirects, page content, and yes, PageRank. When these signals converge, the decision is clear. When they contradict each other, Google weighs them.

A well-placed canonical remains your best control tool. But if your internal link architecture systematically sabotages your directives (by sending more juice to the wrong URL), Google might choose to disregard you. This is why consistency between technique and linking strategy is so important.

• PageRank remains an indicator of internal and external authority for Google, even if it’s no longer visible publicly
• Canonicalization relies on multi-signal arbitration: tags, content, redirects, and PageRank
• Poor link architecture can reverse your canonicalization intentions, even with correctly placed tags
• The weight of external backlinks influences which URL becomes the reference in the case of unresolved duplication
• Optimizing internal linking is not just about on-page SEO, it's also about controlling indexing

Is this statement consistent with field observations?

Absolutely. For years, SEO practitioners have noted that Google sometimes ignores explicit canonical tags. The most frustrating cases? E-commerce sites where the version with UTM parameters becomes the indexed page or blogs where the AMP URL supersedes the desktop version despite a contrary directive.

Gary Illyes's response provides a missing piece of the puzzle. If these "undesirable" URLs receive proportionately more authority signals — through unrefined social shares, misconfigured backlinks, or poor internal linking — PageRank favors them. Google follows the strongest signal, not necessarily the one you’ve declared.

What nuances should be added to this claim?

To say that PageRank "is used" doesn’t mean it is dominant. The exact weighting remains opaque. [To verify]: Google has never communicated a precise ratio between the different canonicalization signals. We know that the canonical tag carries significant weight when it’s coherent, but what margin of error does Google tolerate before favoring PageRank?

Another blind spot: Gary refers to "duplicate pages," but what is Google precisely referring to as duplication? Strictly identical content? Variations at 80%? Paginated pages? The granularity of detection inevitably influences the frequency of activation of this PageRank mechanism. The more duplications Google detects, the more this signal comes into play.

In what situations does this logic not apply or become problematic?

If you have a massive technical duplication — say 10,000 URL variants for the same product via dynamic filters — PageRank cannot arbitrate everything. Google may simply choose randomly or demote everything out of caution. Relying on PageRank as an arbiter then becomes a risky gamble.

Second edge case: multilingual or multi-regional sites. When you have cross hreflang + canonical tags, regional PageRank can favor a geographically irrelevant version if it dominates in backlinks. A .com with strong authority could overshadow a .fr that is better targeted but less "powerful." Here, the PageRank signal contradicts user intent — and Google must choose between relevance and authority.

What concrete steps should be taken to leverage this information?

First, audit your internal link architecture. Identify the pages you want indexed as canonical, then check that they are indeed receiving the maximum amount of quality internal links. Too many sites send 80% of their juice to navigation pages, filters, or technical URLs with no SEO value.

Next, map out your problematic external backlinks. If third-party sites consistently link to duplicate versions (tracking parameters, sessions, etc.), either 301 redirect them or contact webmasters to correct. Each backlink to a bad URL reinforces its PageRank and complicates Google’s canonicalization task.

What mistakes should absolutely be avoided in this context?

Do not create signal conflicts. If your canonical tag points to URL-A but your XML sitemap submits URL-B, and your internal links favor URL-C, Google is faced with an impossible choice. It will default to PageRank — and it may not be the version you want.

Another trap: ignoring chained redirects. If your preferred URL is at the end of a series of 301 redirects, its PageRank erodes with each hop. Google might decide that an intermediate version of the chain, better linked, deserves canonical status. Simplify your redirects to avoid this kind of drift.

How to check if my site is correctly optimized in this regard?

Use Search Console to compare the URLs indexed by Google with those you've declared as canonical. The "Coverage" tab and the "URL Inspection" report reveal cases where Google has chosen a different version than the one you specified. If the discrepancies are numerous, that’s a warning sign.

Next, analyze the internal PageRank distribution using an SEO crawler (Screaming Frog, OnCrawl, Sitebulb). Identify pages that accumulate juice abnormally without being strategic. Redirect the linking towards your target pages. For detected duplications, add clear canonicals and check that they are respected in the following weeks.

• Audit internal link architecture and prioritize strategic pages in linking
• Clean up backlinks pointing to duplicate or non-canonical URLs
• Check consistency between canonical tags, XML sitemap, redirects, and link structure
• Eliminate chained redirects that dilute the PageRank of preferred URLs
• Use Search Console to detect discrepancies between declared and indexed canonicals
• Analyze the internal PageRank distribution with a crawler to identify leaks