PageRank Sculpting and the Emperor’s New Clothes

Add Your Comments

In the history of SEO, few controversies have been more interesting than the debate over PageRank sculpting, and fewer have raised more troubling questions about the state of the SEO industry.

PageRank sculpting is an SEO technique where the flow of PageRank is blocked from less-important pages in a website in order to pass more PageRank to important pages. PageRank sculpting is hardly new. Stephan Spencer of Netconcepts, for example, has been preaching the merits of sculpting with the nofollow attribute since at least 2007. Other SEOs used JavaScript for sculpting long before.

Despite the potential benefits of controlling PageRank within a website, most SEOs (me included) stayed away from sculpting, out of concerns that Google may view it as a spam technique. All that changed, however, after Google’s Matt Cutts essentially endorsed sculpting in a series of interviews and conferences beginning in late 2007. In fact, PageRank sculpting became a mainstream optimization technique, with SEOs jumping onto the bandwagon en masse.

Fast-forwarding to the June 2009 SMX Advanced conference, Matt Cutts dropped a bombshell by stating that Google had changed its handling of the nofollow attribute in a way that negates PageRank sculpting — and that this change had been made more than a year earlier. This raised a number of disturbing questions, chief among them why Cutts discussed the nofollow attribute at industry conferences for at least a year after it was apparently rendered ineffective for the purposes under discussion.

Mind you, Google has no legal obligation to tell SEOs about any portion of its ranking algorithm. In this case, however, Google — and Matt Cutts specifically — had a responsibility to disclose the change because they specifically endorsed the technique. Not doing so puts Cutts (the leader of Google’s Webspam team) in the awkward position of encouraging SEOs to practice something that Google actually considered to be spam.

The second troubling question is why the buzz around PageRank sculpting had continued, even after its impact was negated by Google. To quote Matt Cutts, “At first, we figured that site owners or people running tests would notice, but they didn’t.” In the SEO world, we are supposed to test optimization strategies to learn what really influences Google’s rankings. How is it that a major optimization technique could be defused, yet no renowned SEO noticed or announced that fact? What about SEOs using PageRank sculpting on the websites they worked with? Did they notice it wasn’t having any effect?

That was precisely the experience at my company, BridgePose. As latecomers to PageRank sculpting (for reasons mentioned above), we first tested the technique on a handful of sites during 2008. It had a clear impact on rankings. Encouraged, we deployed sculpting on about a dozen other websites during the fall of 2008, but this time saw absolutely no impact on rankings. At the same time, rankings for our earlier sculpted websites began to go soft.

What was going on? Maybe the sites were too small to benefit from sculpting. Maybe other factors were driving down rankings. Maybe, maybe, maybe … we weren’t doing it right. The SEO community was buzzing about PageRank sculpting. We wondered if there was something we didn’t get about this?

Maybe that is the key. PageRank sculpting did indeed work at one time. Maybe if everyone else continued to say it was working, none of us wanted to publicly state that it wasn’t working for them. Perhaps the SEO community was guilty of groupthink, where conformity to the conventional wisdom suppresses independent thought. Personally, the lesson I have taken from this is to always trust my eyes, no matter what anyone else in the SEO community says. And that applies even to statements made by Google employees. Even if their employee badge reads “Matt Cutts.”

About the Author

Tom Dahm is the President of BridgePose Search Engine Marketing and past founder of NetMechanic. He has been optimizing websites since 1996.

Add Your Comments

  • (will not be published)