'No follow' Links are PageRank “Black Holes” – Maybe
Update: Subsequent to this post being written Matt Cutts has confirmed that the change to the handling of nofollow described below is real, and that in fact it has been treated this way for over a year.
This week is awash with bloggable news, but within SEO circles the most "buzzy" has been Google's apparent shift in its treatment of nofollow, a story originally broken on Search Engine Land. In this post I want to demystify what nofollow actually is and what the implications of the change really are. I'll caveat that by saying that I'll deal strictly with the theoretical assigning of PageRank or, more generically, "link equity", to pages within a site based on linking relationships between those pages.
"Nofollow", for the uninitiated, is a value of the HTML rel attribute that can be applied to a link preventing search engines from crawling the target of that link.
Before nofollow existed, pages simply passed a diluted version of their own PageRank (donatable PageRank) equally to each page they linked to. Page A passes a third of its donatable PageRank to pages X, Y and Z;
With the advent of widespread search engine support for nofollow ( first announced by Google in January 2005) , websites could use the tag to theoretically block the flow of PageRank to pages they deemed less important, thus increasing the relative value of other links that were followable. A bit like damming up parts of a river to force more flow down other channels. This practice became known as "PageRank sculpting".
The link to page X has been nofollowed (let's assume it's a terms and conditions page, a typical target for PageRank sculpting), so pages Y and Z receive more PageRank from page A;
After the recent announcement, links that are nofollowed neither pass PageRank or improve the relative value of other links on the page. Instead, they "absorb" the PageRank that would have been passed through them. It's this behaviour that makes them akin to black holes. The link to page X is nofollowed, but pages Y and Z still receive the same amount of PageRank as they would if the link wasn't nofollowed;
therefore, in a heads up competition between using nofollow and not
using it, you're better off not using it at
all. That way at least a small part of the PageRank
passed through the link you would have nofollowed will
find its way back to important pages, rather than being lost
entirely to the nofollow black hole. The nofollow tag has
been removed from the link to page X. Some PageRank is spent
on the page but at least it can subsequently pass a portion of this
to page Y;
Remember that if the prospect of page X being accessed by search engines is particularly terrible, you can use the robots directive noindex, follow in the robots meta tag to let PageRank flow through the page without the page itself being indexed.
As with any sudden changes and announcements made by the search engines though, sound advice for now is to sit back and see what effect this actually has. For all we know, these effects may already have manifested, or they may be too miniscule to detect or be concerned about anyway.
Will this move kill the notion of PageRank sculpting? I doubt it. The nofollow tag was never the only way to go about optimizing your PageRank distribution, it was merely a tool introduced by Google to nullify link spam that didn't really have the desired effect and just played into the hands of some SEO's who tried to use it to their advantage in funneling PageRank around their sites.
What this does do though is take away an easy and inconsequential to use technique, whereby you could avoid answering difficult questions about site structure and fixing the underlying structural problems a site has, and happily plaster nofollow all over your site and call it a day. This move by Google may just force a slight return to more "Thoughtful" SEO, where knowing how to design an information architecture and internal linking schema actually matters.