info [ AT ] redcardinal [ DOT ] ie
Interesting Case Study: rel=canonical Google Fail
Michael Wall has written about a short test he ran on a small site using the rel=canonical tag. Very interesting, and certainly one test does not a trend make, but I’d have expected more of Google…
This is why I’ve tried to tell people that rel=”canonical” is simply a bandaid, and the best solution is prevention not cure – publish content on one, and only one, URL.
I would of thought it would behave the other way i.e. doesn’t really work all that well in keeping pages out of the index. I have used for duplicate content issues but I don’t see it anywhere near as effective as 301 redirect.
Comment by Kieran — October 19, 2009 @ 8:30 pm
ha, i get confused by my real name and my online name
hmm that is a bad sign !!!
Comment by searchbrat — October 19, 2009 @ 8:31 pm
Hey Kieran
I’m really not that fond of this element. I think in Michael’s case Google failed to see that the inner pages were unique. The desired behaviour is that the pages will only be replaced with the canonical if they are identical/near-identical. It’s always better to prevent problems rather than cure them.
Rgds
Richard
Comment by Richard Hearne — October 21, 2009 @ 9:25 am
Hi Richard,
thanks for the reference. You’re right, Google didn’t spot that the 30 or so pages were totally different, different pages title, meta description, content, images, folder and file paths etc and that the canonical url element should have only applied to the homepage at most.
There were no other factors robots file, noindex etc to signal to Google that the pages were not to be indexed.
Months passed and still it didn’t index the pages until I removed the canonical url element. Bing and Yahoo had no problem in indexing all the pages.
Comment by Michael — October 22, 2009 @ 3:16 pm