Quite often when we look at a websites prior to starting SEO on a website we find canonical problems with sites that have multiple domains – so what does a canonical problem mean when we use it in relationship to search engines and computers.
First the Science Bit
canonical is a term in mathematics ”The usual or standard state or manner of something”
Computing has borrowed this term and uses it in a similar way to describe problems where working out what is actually meant by something can cause problems.
If you have ever used a website and entered a date and had a problem because the website expected a us date format and you entered a UK format you will have experienced a canonical problem.
For example using the date format “month/day/year” instead of “day/month/year can cause some nasty problems.
01/05/2006 is the first of May in UK usage but the 5th of January in American this can cause problems so a lot of websites and experienced computer programmers will use a date format like 05-Jan-2006 which removes any uncertainty to what date is meant.
So how does this effect my ranking in search engines
With websites there can be problems with search engines when a website has multiple domains how to work out which is the definitive domain? Humans are smart and can normally work this out but computers are really dumb! and can get confused
For example a site we recently worked on has a lot of domains
they also have another 8 different domains
When Google or any search engine sees this as 2 or more different websites this can cause the different versions of the website to compete against each other this can cause the sites to perform poorly compared to just a single site.
Also each version of the website will have links to it some may link to the .co.uk some to the .com as the number of links to a site is a major part of how search engines decide to rank a site in their search results – its obvious that splitting your links across more than one domain is a bad idea.
Also website will often respond both to the version of the site with the www prefix and the one with out for example.
so in our example there are 20 Versions of the website.
To solve the problem you should configure the website to tell browsers and search engines what the definitive name is.
This is done by using a permanent redirect which in layman’s terms says
“hey Google this website is no longer here its over here now – please update your records”.
When you browse a website that is redirected you get sent to the correct (or canonical) version of the website. The same things happens when a search engine reads a site it gets told exactly what the definitive version of the site is.
It is vital to use the permanent redirect and not a temporary one.
Each websever will have a way of doing this for a system based on Unix/Linux this is normally done using mod rewrite and .htacess (this is normally already installed) for Microsoft web servers a separate third party piece of software (an ISAPI filter) has to be installed
For our experience after all the multiple domains have been redirected it take up to 2-3 months for the full effect of redirects to show up in the search engine results.
There is also the issue of which domain is appropriate for the target .co.uk or .com for example we blogged about geolocation issues a while back here