Matt Cutts: rel=’canonical’ed pages can pass PageRank

by Christopher Wendels on juli 13, 2011

UPDATE: Since writing this I have worked at Google. I decided when I joined Google to leave any previous posts untouched – to neither confirm or deny any information. I would however, like to clarify that the issue here is if links on canonical pages will influence the ranks of their target pages – or if those links are used solely for discovery (if at all).

Matt Cutts just threw this video out on the WebmasterHelp Channel on YouTube. Please watch it before you continue:

Embed video:

rel=’canoncial’ isn’t doing what we were told it did
Either that or Matt didn’t get it right. Matt started out by saying it was a confusing question – I did not find it confusing in the slightest, but that is just me. We have the following scenario:

  • Page A is the original page we want to rank in the SERPs
  • Page B is a very close replica of A and we are worried about duplicate content and we add the rel=’canonical’ tag. However links may be slightly different on this page – for purely “white-hat” reasons, of course!

The question posed is what happens to the links on page B. Up until today I assumed (and have not read anything, anywhere; nor have I seen any results in my own or others tests that point to the contrary) that a rel=’canonical’ed page would flow all PageRank from canonical pages to page A and consolidate there. In other words links pointing TO page B would pass PageRank to page A.
Follow so far, but what’s new?
Well, what Matt said in the video, and again I’m assuming that he knows what he’s talking about, is that links FROM page B will pass page B’s rank onward.
Wait, what?
This means, barring Matt Cutts not being correctly informed, at least one of the two following possibilities:

  • Page A receives page B PageRank for ranking (SERP) purposes, BUT is not able to pass that full value of PageRank to the pages linked to from there.

OR

  • Page A revives page B Page rank for all purposes AND page B is still able to pass the first amount of PageRank on through its links.

None of these scenarios seem reasonable to me, since both have awesome potential for abuse and link schemes. I would love for @MattCutts to clear this little mess up.