Subscribe in a reader
HOME
Do blogroll links bleed pagerank?
I believe that blogrolls do not bleed pagerank, this is why I am not afraid to offer sitewide links to those who comment on my blog if they follow a few basic rules. Remember how back in the day before blogs search engine optimizers would fill their footers with text links to cross promote all their websites? This unfair practice was very similar to circular linking in that if you had many sites you could own keyword real estate simply by placing text links in the footer. If you noticed over time this technique lost it’s value as Google improved it’s algorithm.
Today by default we link to SEOBOOK.com and other places without much thought from our blogrolls. What about the rest of us unknowns? Would it be fair to allow the massive power of the blogroll to go unchecked as it promotes only a few popular bloggers? The answer is no and it comes down to a pretty basic equation (and I have no idea what the math is).
Links that exist in a blogroll do not change so they do not bleed pagerank. Linking that does not change is also easy to detect and flag as blogroll links just like links in the footer. In fact I would go as far as to say that all external links that go offsite in footers, headers, sidebars AND blogrolls do not carry as much weight as in content linkage accompanied by surrounding text. So do you agree that blogrolls do not bleed pagerank? I have proof BTW but do not want to keep linking the poor folks who work @ Google. ;)
Similar Post:
February 17th, 2007 at 2:59 pm
My research shows that external links are just fine, with a couple of caveats.
- they must be ontopic (so hope you don’t get a top-commenter with a blog on, umm, fairf floss tasting in Newark. That could do weird things to Google’s perceptions of your linkworth)
- they must ‘fit’. Footer links in a small font can also do weird things to Google’s opinion of you (not that a blog roll normally appears in the footer, or in small font.) You could probably also get around this, at least for this week, by being up to date and using external css rather than font elements.
- probably a coupel of others that don’t occur to me right now…
February 17th, 2007 at 6:58 pm
Lea - Yes indeed and your blog is kind of related, you have been added to my blogroll. :)
February 18th, 2007 at 12:27 am
From what I have seen they leak, and Google discount the value afterwards and don’t give it back to the other links on the page.
February 18th, 2007 at 5:34 pm
I’m with Andy on this one. I think blogroll links (along with footer and header links) bleed PR. I like to see your proof even if its more anecdotal rather than actual proof.
I do agree with you though that they’re not as effective to the destination site as is a link that is surrounded by text.
February 19th, 2007 at 3:22 am
Aaron, from my experiments this past year new sites seem to bleed a lot more PR than old sites with the same PR. The sites I have tested have also had about the same links inbound, due to very normal linking.
So for example these old sites have been around for years, but have few links from low PR sites.
I did notice that sites that have been around for less than 6 months without a strong linking strategy seem to leak PR faster than most. Popular blog sites, even out of the gate seem to avoid this. Due to their amazing ability to drive content based links, rather than navigational links.
My theory is that Google uses an algorithm that has a lot to do with the amount of links that appear that are relevant within the body of the document, rather than links within the navigation of the site.
This especially applies within the case of sites exchanging links with one another, for example this year I exchanged links with the main site of my primary company and it seemed to have little to no effect. Even though the main site has a PR of 9 and millions of useful links.
I imagine that Google can detect at this point common parts of a site, for example the navigation. So if all the pages look the same, except the main content than nothing is more important than the text that appears within the corpus of the document. Including links.
Just my random thoughts
February 19th, 2007 at 7:26 am
Aaron Shear wrote:
That’s it, you win the prize, you are correct sir! ;)
December 19th, 2007 at 2:21 am
I imagine there are also other things that they take into account for similar pages:
- On the same IP range
- All pages of one site linking to other site (e.g. cross-site navigation) as opposed to the link being in one high-page ranking text page.
I use three domains for my three sister sites. Rather than existing for page rank, each site exists on a slightly different topic. I was having difficulty getting links into sub-pages - people will more generally link to the top domain page, unless linking to a specific content article.