I know that aggressive ad positioning, especially when your site is new, is a flag for spam with Google’s algorithms. I created a blog on fishing once that had ads in the top left and in much of the sidebar, which led to me having to make it through a Captcha for each post. What’s interesting to me now is the question, does aggressive ad publishing decrease PageRank?
The answer is, or should be, that aggressive ad positioning (mainly those PPC and aim to send visitors clicking through to an advertiser’s site) decreases the PageRank that a page passes to the other pages it links to. PageRank, if you’ll recall, is the probability that a random surfer will visit a given page. The more probably a visit, the greater that page’s PageRank (PageRank’s name actually comes from Google co-founder Larry Page, though (Page’s Rank[ing system]); it doesn’t refer to ranking web pages).
When an ad is displayed in such a way as to encourage visitors to click it rather than the other links on a site, it makes it less likely that a visitor will click the other links. Therefore, they’re less likely to visit those pages. So the PageRank that the page with the ads on it can pass to other pages it links to is thus decreased.
This application of PageRank theory isn’t just good for external links though. If you’re trying to get someone clicking your ads rather than your navigation, for example, they’re also less likely to visit the pages your navigation links too. So you should be passing less PageRank to your own internal pages as well.
The flip side is that this means that if you advertise using AdWords, you should be increasing your PageRank. People are more likely to visit you because you’re advertising. And the greater the CTR on your ads, the greater your PageRank ought to be.
All of which brings me to nofollow and paid links. If a link is sending traffic, it should count towards a site’s PageRank. Saying that you can stop a link increasing the likelihood of someone visiting the page by putting a nofollow attribute in the link tag is nonsense. It’s not like you’re telling visitors not to click through and visit the page you’re linking to (except maybe for SEOs, who, like me, use Aaron Wall’s SEO for Firefox to highlight nofollow-ed links). It’s like I posted at Youmoz; Nofollow and sculpting PageRank don’t make sense.
Bottom line: Brin and Page’s discovery of probability of visiting as a model for ranking either needs to be trashed entirely – wealthy advertisers could just dominate smaller competitors – or else
a) the proxies for measuring probability need rethinking, or
b) probability needs to be defined more narrowly. For example, it could be redefined as “the probability of visiting because of non-commercial influences, commercial influences being here either direct (AdWords, paid text links) or indirect (branding).”
Liked this piece of search theory? Get my RSS feed.