How does PageRank work, really?

I’ve read a lot of comments over the last few weeks with wild theories about how PageRank works.  Some things I’ve heard:

• appropriate content
• good design
• don’t use flash
• get incoming links
• good meta tags

Four of those are completely wrong.  One is right — incoming links.  When calculating PageRank, all that matters are the links coming into your site.

Understand, of course, that hundreds of other things can affect your ranking in Google, including those items listed above.  However, for PageRank it’s just simple math.

Here is the formula Google used when PageRank first came out.  The currently formula is possibly a bit different, but certainly quite similar:

PR(A) = (1-d) + d(PR(t1)/C(t1) + … + PR(tn)/C(tn))

Let’s break it down:

PR(A) — The PageRank of your site

(1-d) — d is the “dampening factor”, typically 0.85

+ d(PR(t1)/C(t1) — PR(t1) is the PR of a site linking to you.  That PageRank is divided by the number of outgoing links on that page so that each link from that page gets an equal slice of the PageRank.

PR(tn)/C(tn) — Same idea as the last one, just showing us that it keeps adding the totals from every page that is linking to you, the multiplies that total by d, the dampening factor.

The dampening factor serves two purposes.  First, it makes sure that every page has some PageRank, even if there are no links coming in to it.  In that case, the PageRank is simply (1-d), or 0.15.

The second reason is so you couldn’t get a page up to a certain level and then run single links off of it to other pages and have them all instantly at the same high level.

It’s also important to note that the number calculated here is not the same as the 0-10 number that you see in the Google toolbar.  That number is based on this one.

The 0-10 number in the toolbar is set on a logarithmic scale.  Google won’t say what the scale is, but most people figure it’s around 5.  This means you need five times as  much PageRank to move up to each new level.  This also means that there are 1/5 as many sites at each higher level.

Here’s the bottom line.  PageRank is based on three things and three things only:

1. The number of links coming into your site
2. The PageRank of the links coming into your site
3. The number of links on the pages of the links coming into your site.

That’s it!  Nothing else goes into it.  Again, though, hundreds of things go into determining your exact ranking position, but nothing else goes into your PageRank calculation.