In 2009 Google announced a new tag canonical that will help you to remove duplicate content on your site indicating the preferred version of the URL for the page.
If your website has identical or very similar content available on different URL, then the new format allows you to specify the URL, which should return to the search engine. Also, you can be sure that all characteristics such as the reference weight, etc. will be passed to the correct version of the address.
Now you can add this tag to specify your version of the address inside the tag <head> on pages with duplicate content:
<link rel=”canonical” href=”http://www.example.com/product.php?item=swedish-fish” />
In this way, Google will understand that all duplicates refer to the canonical address specified in the tag. Additional features of URL such as PageRank and related signals will transfer from the duplicate pages to the indicated page.
This tag will be useful mainly for the use of different CMS (phpBB, IPB, WordPress etc.)
For the popular WordPress there are already plugin canonical embed tag on the correct pages.
Answers to the frequently asked questions on the tag:
Is rel = “canonical” a hint or a directive?
This is a hint that we take into consideration and in conjunction with other signals when calculate the most relevant page to display in the search results.
Can I use a relative path to specify the canonical, like this: <link rel = “canonical” href = “product.php? Item = swedish-fish” />?
Yes, the relative paths are recognized as well as in the usual tag. Even if you enter a tag with a link to a document, then the relative path will be considered in accordance with the base URL.
It is okay if the canonical addresses contain not fully duplicate content?
Search engines allow small differences, such as the sort order in the table of products. We also understand that the canonical addresses are parsed by robot at different times, so everything is fine.
What if the rel=”canonical” shows 404 error?
Search engines (SEs) will continue to index your content and to use a heuristic approach to determine the canonical address, however, they recommend you to use the current URL as canonical.
What if the rel=”canonical” is not indexed yet?
SEs try to reach canonical URL quickly. As soon as they index it, they immediately consider newly the hint rel = “canonical”.
If the canonical URL can contain redirect?
Yes, you can specify a redirect, in this case the search engine will treat the process of redirect as usual and will try to index the new address.
What if I have a conflicting signals for the rel = “canonical”?
SEs algorithms are soft: SEs can follow canonical chains; however, we strongly recommend that you indicate on the pages a single canonical address in order to ensure optimal result of canonization.
Can this tag for links offer a canonical address on an absolutely different domain?
No. To migrate to a different domain, the 301 redirect is more convenient. Currently, Google supports the canonization only inside subdomains or within one domain. So site owners can indicate www.example.com instead of example.com or help.example.com, however, they can not indicate example.com instead of example-widgets.com.
It sounds interesting but can I see an example?
Yes, wikia.com helped Google as a trust tester. For example, you will notice that the source code at http://starwars.wikia.com/wiki/Nelvana_Limited contains rel=canonical http://starwars.wikia.com/wiki/Nelvana.
Two addresses are almost identical, except that Nelvana Limited, first URL, contains a short message at the heading. This is a good example of the use of the tag in the future. With tag rel=canonical characteristics of two addresses combine and search results display the correct version.