Hire a senior Perl / Python programmer today; download my up-to-date resume (PDF)
John Bokma MexIT
freelance Perl programmer

A PageRank myth busted

Thursday, January 4, 2007 | 0 comments

Yesterday in the newsgroup alt.internet.search-engines 'N-P-H' claimed that "the level 1 and 2 node web pages are more important than the level 3 and 4 nodes." I guessed that 'N-P-H' meant that the index page of a site is the most important one, and the pages linked from it are less important etc., and told 'N-P-H' that this is wrong. Today 'N-P-H' confirmed that I guessed right, but disagreed with my view because "since that is precisely how they (Google) compute page rank.". Well, it's not.

The "PageRank drops with every directory level" myth

A common myth that now and then shows its ugly face in alt.internet.search-engines, and probably many other SEO related forums is that Google calculates the PageRank of pages on a site by using the PageRank of the index page and subtracting one for each directory level encountered. For example, if your site (index page) has a PageRank of 7 the myth goes, the PageRank of /foo/bar/baz.html would be 4 because /foo/ will be 6, and /foo/bar/ will be 5 and since baz is not the index page of bar it's even a level deeper and so it will be 4.

A variation on this myth is that if the index (again say PR 7) links to a page, say foo.html, and foo.html links to bar.html, and bar.html links to baz.html, that the PageRank will be as follows:

Of course this is silly, because like the first method it doesn't take into account the fact that other sites can be linking to (voting for) baz.html. And that's what PageRank is all about.

From what I have read so far, this myth was born when Google used indeed the first method to assign temporary PageRank to pages. I am not sure if this method is still used for assigning a temporary PageRank, but it's not important at all because it's not how actual PageRank is calculated for pages belonging to a web site. Actual PageRank does take into account external links.

"PageRank drops with every directory level" myth busted

This myth can easily be busted. For example the blog entry Creating an XP Pro VM for the free VMware Player on this site should according to the "subtract one for each folder" have the following PageRank:

I checked the actual PageRank of the blog entry with a Live PageRank checker and it's 5.

As for the variation on the myth, the shortest path (as far as I know) to the aforementioned blog entry is as follows:

Close this time, but no cigar.

Another important issue worth mentioning is that PageRank isn't a major factor anymore (if it ever was) in deciding the order on a Google search engine result page (SERP). It might be very well the case that a page with a PageRank of 4 ranks well above a page with a PageRank of 8.


My conclusion is that it's silly to focus on the index page, then on the pages one level "below", then on the pages two levels "below" etcetera. Focus on content, because content is king. The page that at the time of writing gets the most traffic on this site via Google Search is a blog entry (Security webcam hunting with Google). It clearly shows that the "PageRank drops with every directory level" is nothing but a myth.

But even if this myth was a fact, it still doesn't matter much because PageRank is just one ingredient in the very complex SERP ranking algorithm Google uses. And according to some PageRank is no longer an active ingredient.

Again, what really counts is good content, because good content makes people link to it. And inbound links drive traffic to your site without relying on a search engine. It's always good to have a backup plan.

PageRank related

Also today

Please post a comment | read 0 comments | RSS feed