Or using rel=canonical for fun and profit.
I’ve been toying with this post for about two months now, because this technique is so good that I didn’t really want to share it! Quite honestly I’ve not had enough of a chance to implement this across the network of sites I work on, but I have tested it and it does work.
A picture of a cute kitteh because I'm always being told I need images in posts
Rel=canonical background
When Google, Microsoft and Yahoo! announced support for the rel=canonical tag, many webmasters rejoiced – in fact I was rather excited and posted about duplicate content and rel=canonical here. At the same time I was looking at some of the anchor text used in the internal links off a number of the large sites I work on, and getting frustrated by navigation links being given generic names like ‘news’, ‘reviews’ and so on while appearing at the top of the page when you look at the raw HTML.
So I did some scouting about and came across an old post by Rand on SEOmoz and Branko on SEO Scientist that appear to confirm my fears. Only the first link on a page counts, and only the anchor text for that link counts. So if you’re linking to your awesome technology news page in the navigation of your site just using the word ‘news’ as anchor text, because let’s face it technology is quite a long word when you’re trying to fit a bunch of stuff onto a top navigation, all those other links to the same page with great anchor text further down the page count for jack.
In my day job we have issues with people stealing our RSS feeds and reusing them. As I have mentioned previously in my post on using blog content theft to your advantage I don’t generally see this as much of an issue. We’re covered by having sensible solutions in place which allow us to effectively garner at least five links per article with anchor text we chose. What irks me somewhat is that because we track all the visits and pageviews per visit to our sites by adding campaign tracking variables to our RSS we don’t see the full benefit of these links. Or we didn’t until we implemented rel=canonical.
Joining the dots
Pulling all these strands together, I realised that it could be possible to use rel=canonical to our advantage. Let’s track the links from site sections – navigation, footer, sidebars, wherever using tracking strings. Let’s make sure that the search engines know what we’re doing and what the canonical page is and let’s get a load of links from one page with different anchor text which all points towards the same page.
So I tested it. And what I found was that by using tracking strings of the type ?attr=stuff&src=somethingelse on the end of a URL – so http://www.example.com?attr=stuff&src=somethingelse you’re telling search engines that is a different URL to just http://www.example.com, and the anchor text you use for both counts and ranks.
Getting more links from high ranking pages
So how does this help you get more links from high ranking pages? In general the page on your site with the highest pagerank will be your homepage. If you’ve got a navigation with sub-optimal anchor text, don’t use nofollow, use a ‘tracking string’ on it. When you find a high ranking page that will link to you more than once, and you want to rank for more than one keyword term for the same page, use a ‘tracking string’. While this method is nowhere near as good as having a sensible navigation and rel=canonical isn’t a foolproof method, it will help you in your efforts to rank for more than one keyword term using both internal and external links.
I will caveat this post by reiterating that I’ve not run large scale tests on this method and I’m unsure as to the ‘hat’ that this may be construed as. I’m thinking it’s light grey at best. Do though let me know how you get on with using it, and if you do see improvements in traffic from natural search as a consequence.
Kitten image via Timylee on Fickr
{ 8 comments }
