Simple But Catastrophic Errors With Canonical Links
The canonical tag, introduced by Google early in 2009 was supposed to resolve many duplicate content worries. It seems to simple to be true though and even though Google claim that the canonical tag is a “hint” rather than a directive, many people have had catastrophic results just because they didn’t apply it right.
In theory, pages with duplicate content (but different URLs) are the ideal ones for canonical tags to be in place. Even though many people think that they are equivalent to 301 permanent redirects this is definitely not the case. However, they will be treated similarly to redirects and you have to be very careful because unlike redirects, it won’t be instantly obvious that something has gone wrong.
There are cases where 301s are the best solution, whereas in other occasions canonical instructions would be ideal. Because they are easy to implement that doesn’t mean they cannot be disastrous.
Common Mistakes
- Canonical link that points to a broken link (non-existent URL). This is definitely something you wouldn’t like to do but unfortunately it does happen a lot by careless web developers. I’ve seen well-ranking pages disappearing over night and not coming back unless the canonical instruction has been corrected. All incorrect canonical links will be treated as broken links and Webmaster Tools will report them as 404 errors.
- Canonical links that point to the homepage. Again, another catastrophic situation with quick and disastrous effects usually caused by a coding error.
- Canonical link that points to the same page (itself). This is not necessarily catastrophic according to Google and some web masters but there is absolutely no reason to do something like that as it may cause problems in the future.
- Canonical link of a product page, pointing to a category (parent) page. There isn’t any benefit doing something like that as the two pages should not have duplicate content anyway.
- Canonical link of a category page pointing to a sub-category page. This is definitely not a way to pass link power down to deeper pages and should be avoided.
- Rely on plugins that automatically generate canonical tags. WordPress, Joomla and Magento all offer canonical tags plugins but some users have experienced problems by not being able to configure them correctly.
- Missing forward slash in URL. That will result in a 404 page not found with all its disastrous consequences.
Hot to Resolve Such Problems
- Always copy the canonical link and paste it in a browser to see if it resolves to the desired page. Looking right doesn’t mean that it’s right!
- If you are 100% confident that the canonical link has been done incorrectly then just fix it and wait for up to 4 weeks. If nothing happens just revert back, get rid of all the canonical instructions and work out another solution.
- Always update the XML sitemap after you’ve taken any action.
Modi is an SEO consultant working for a company that organise Cunard Cruises.
Latest posts by DST Contributor (see all)
- Survival Guide to Multilingual SEO - May 14, 2013
- Five Killer Link Bait Tips That Can Provide You With ‘Passive Marketing’ - May 11, 2013
- 5 US SEO Events to Visit to Spend a One Month Vacation with Purpose - May 9, 2013
11 Responses to “Simple But Catastrophic Errors With Canonical Links”
Newsletter
Join the Contest!
Tags
Our Projects
I was advised to use the following point to avoid google ranking mydomain.com and www.mydomain.com as two separate websites:
•Canonical link that points to the same page (itself). This is not necessarily catastrophic according to Google and some web masters but there is absolutely no reason to do something like that as it may cause problems in the future.
Was I incorrectly advised?
.
@munday21
No, you were not advised incorrectly….
Diff people (designers, users, backlinkers, bookmarkers etc) use diff URLs (with WWW, without WWW, with HTTP, without HTTP) and these diff URLs of the same page get backlinks (internal and external) over a period of time. Now when Google will find backlinks to all these versions, it is bound to index all of them in its DB (If a URL is getting backlinks, Google will index it in spite of the URL being blocked via robots.txt). By canonical tag, we are not only telling Google the preferred version of the URL but also asking to index only that URL and give the entire juice to this URL. So to avoid such a situation and be on the safer site you have been advised to add the canonical of the same page itself.
@munday21
Yes, you were correctly advised but I’d suggest doing that in a web server level rather than adding canonical tags. Depending on using Apache or IIS there are different way to configure the web server so any request to a non www URL will redirect to its WWW equivalent. The reason I’m saying that is because with the canonical you won’t be able to tell sraight away whether it’s done correctly or not whereas in the other way you will be able to check it instantly.
Good stuff, but a couple of points where I take issue.
“Canonical link that points to the same page (itself)….” Not only is this “not catastrophic” but I believe it is beneficial. It is also virtually impossible to encode via a CMS if you’re only trying to do this for non-canonical versions of the page. Rand did an entire post addressing this:
http://www.seomoz.org/blog/dispelling-a-persistent-rel-canonical-myth
Just what sort of future problems may it cause?
“Missing forward slash in URL….” In what part of the URL. If in referenced canonical link (aside from the home page) then this is dependent on your server set up. E.g., if you’re set up so that:
example.com/folder/index.html
and
example.com/folder/
301 to
example/folder
Then the correct canonical URL for that page is
example/folder
… *without* the trailing slash. Or maybe I’m misunderstanding your point here?
Article content is very interesting! Original ideas! I like your article! Thank you to share such a wonderful content!
Hi, i found your blog through search engine.you have the great knowledge it is such a great information. you have really a great a great blog about the Catastrophic Errors. Thanks for sharing the nice information.
There isn’t any need for a canonical to point to itself as it is absolutely rendundand. But what is even more important is that if the canonical URL is incorrect (e.g. 404, broken etc) then the error wouldn’t affect that page as there you wouldn’t instruct the bots to prefer the broken or not found page to the current one. Being cautious and not taking unnecessary risk should always be a priority.
@Aaron
I was talking about missing the forward slash after the domain name resulting in 404 not found pages(for example http://www.site.comcategory/product.htm instead of http://www.site.com/category/product.htm
@macmodi
I dont think it is totally redundant. Here is a point that I would like to make:
Please see the URL patterns as an example:
http://www.example.com/
http://www.example.com/?utm_source=citysearch&utm_medium=banner&utm_campaign=christmas OR
http://www.example.com/?utm_source=feedburner&utm_medium=email&utm_campaign=newsletter
Here, the first URL is the preferred version and the other 2 are tagged URLs (Google’s URL Tagging Tool). Now suppose these 2 tagged URLs are routed through various campaigns and people might hyperlink to these URLs, bookmark them, share etc. If any of these 2 URLs gets a few backlinks, Google will index the tagged version too thereby creating 2 or 3 URLs for the same page/content, which is exactly the problem that canonical is trying to sort out i.e. removing duplicate pages/urls from the system. Here, since the tagged versions are not separate file/s on your server, where would you put the canonical tag?
Therefore putting the canonical tag on the page itself in such instances is correct and not redundant.
To add further, I would like to share something from my recent experience. I was doing a campaign for a client who used tagged URLs for various campaigns. Some of these tagged URLs were shown in the “links to your site” section of Google webmaster tool i.e. these URLs were getting backlinks. After implementing the canonical tag, Google webmaster tool stopped showing backlinks to these tagged URLs (after a few weeks).
Also, when I used the site: command for the site, some tagged URLs were displayed and after implementing canonical tag, they went away (again after a couple of weeks)
Regards,
Sujay
Simple But Catastrophic Errors With Canonical Links http://tinyurl.com/23rw3ov #SEO // ojo es importante
Would love to hear your opinion on canonical urls suddenly missing and Google only listing the main domain? Is that Google or is that a plugin, in this case All in one SEO pack – I heard some rumors about bad attributes and such…