IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
> PageRank: Where it helps, where it doesn't help, and other facts
Michael Martinez
post Jul 11 2006, 11:14 AM
Post #1


SEO Theorist
*******

Group: Super Mod
Posts: 2183
Joined: 25-January 04
From: Seattle, WA
Member No.: 2



I wanted to write a concise list of PageRank facts this morning. Of course, as life would have it, my phone rang off the hook, so my thoughts fled away while in their most productive state. Nonetheless, here is a start. This is specific to Google, of course. There are indications that MSN, Yahoo!, and Ask all use something like PageRank to one degree or another in their own algorithms.

PageRank: Where It HelpsPageRank: Where It Doesn't HelpPageRank: Other FactsAbsolute Crap About PageRank Believed And Shared By Many SEOs In Forums, FAQs, and Tutorials
  • Linking out to other people's sites will "bleed" your PageRank away. (See Comment 3 below)
  • A link from a (Toolbar) PR 7 page will boost your (Toolbar) PR to 6.
  • The quality of a link is determined by its (Toolbar) PR.
  • Having a 0 (Toolbar) PR value means your page is deemed unworthy by Google. (See Comment 4 below)
  • You can increase your PageRank by getting more links. (See Comment 3 and Comment 5 below)
Comments To The Above Lists
Comment 1:
PageRank is, so far as we know, used to enhance the relevance scores calcuated on a per-query basis. Hence, Google says: "By combining overall importance and query-specific relevance, Google is able to put the most relevant and reliable results first."

Comment 2:
SE Roundtable used the headline "GoogleGuy Claims PageRank Statement to be False". However, the citation they provided only reads:

QUOTE
I'd strongly disagree with the statement that the toolbar PageRank is for "entertainment purposes only"--millions of toolbar users use the PageRank display to judge the quality of pages....


That's not quite the same thing as saying the claim is false. He disagreed with it. In fact, how does Google know whether "millions of toolbar users use the PageRank display to judge the quality of pages". I sure don't. I don't know anyone else who does, except people who are trying to get links. That's an unverifiable claim from GoogleGuy (and Google if they actually say that in any authorized literature or presentations).

Comment 3:
You cannot "bleed" PageRank. PageRank is an approximation of how likely it is that someone will click on a link leading to your page. Whether you put 1 or 1,000 outbound links on your page doesn't change the number of links coming in to your page. Increasing the number of outbound links on your page decreases the likelihood (probability, chance, whatever you want to call it) that anyone clicking on a link from your page will go to any other particular page. You dilute or diminish the amount of PageRank-per-link that your page confers by increasing the number of outbound links. See Comment 5 for why this doesn't matter in the least from anyone's perspective.

Comment 4:
Google's Toolbar PageRank (or, more accurately, "external PageRanks") is only updated "every 3-4 months". What does Matt Cutts mean by "external PageRanks"? He doesn't explain. It is generally assumed in the SEO community that he is referring to the Toolbar PageRank values. For all we know, we could all be full of crap in making that assumption. Maybe he means something else. Whatever.

Comment 5:
New content is continuously being added to the World Wide Web. According to the original The Anatomy of a Large-Scale Hypertextual Web Search Engine (link goes to archive.org) page, "...the PageRanks form a probability distribution over web pages, so the sum of all web pages' PageRanks will be one." If today there are 20,000,000,000 pages in Google's index, the sum of all their PageRanks is 1. If tomorrow there are 21,000,000,000 pages in the index, the sum of all their PageRanks will still be 1.

All other things being equal, every time Google recalculates PageRank, the base starting value for all pages in the index will be lower than it was the previous time. A calculation of PageRank requires a number of iterations (passes through the database in which all the inbound links are evaluated for every page). Estimates vary on how many iterations are made for each calculation, but it is generally assumed to be more than one and many sources indicate more than 10 iterations are required.

If you want to truly increase your PageRank for any 1 given page, you have to accumulate links and/or PageRank through your links at a faster rate than that at which the Google index grows. If the only links you get are links from low PageRank pages, you cannot possibly grow your PageRank. If you only get links from pages with high actual PageRank, maybe you can increase your PageRank.

So why is it that pages appear to increase their Toolbar PageRank value all the time? What many people don't understand is that the Toolbar PageRank is a proxy value. It may or may not be determined on the basis of a logarithmic scale (link opens a .PDF file). Regardless of how the Toolbar PageRank value is determined, it is not a probability (which can only be expressed as a value between 0 and 1 or more colloquially as a percentage). Whatever the Toolbar proxy value is, it is not the actual PageRank calculated for and conferred upon and conferred by individual Web pages in the Google index. This proxy value may go up or down. It doesn't represent the likelihood or probability that anyone will click through to any particular page.
Go to the top of the page
 
+Quote Post
Michael Martinez
post Jul 12 2006, 09:21 PM
Post #2


SEO Theorist
*******

Group: Super Mod
Posts: 2183
Joined: 25-January 04
From: Seattle, WA
Member No.: 2



Matt Cutts has confirmed that he was indeed referring to Toolbar PageRank (with respect to Comment 4). Would be nice if someone could fix the wording for me....


(IMG:http://forums.spider-food.net/style_emoticons/default/whistling.gif)
Go to the top of the page
 
+Quote Post
Michael Martinez
post Dec 29 2006, 03:52 PM
Post #3


SEO Theorist
*******

Group: Super Mod
Posts: 2183
Joined: 25-January 04
From: Seattle, WA
Member No.: 2



J.K.

Please consider this post a request that this thread be pinned. I keep referring to it but I have to sift through old threads to find it.

Tanks.
Go to the top of the page
 
+Quote Post
mandar-seo
post Apr 23 2007, 12:50 AM
Post #4


Member
**

Group: Members
Posts: 10
Joined: 22-June 06
Member No.: 839



This post was really worth $1000. It seems that you have taken lots of pain for collecting the data. I appreciate your efforts.

With regards,
Mandar Thosar
Go to the top of the page
 
+Quote Post
Kevin Peterson
post May 17 2007, 07:11 AM
Post #5


Frood
*****

Group: Members
Posts: 175
Joined: 26-January 04
From: OZ
Member No.: 10



Glad to read this, Michael. Thanks for the work.

I've been working haphazardly on improving my online backup page, while doing the right things. It's amazing how much misinformation is out there. This reaffirms some of my own opinions regarding PR..
Go to the top of the page
 
+Quote Post
Mike Bradbury
post May 18 2007, 11:59 AM
Post #6


Member
*

Group: Members
Posts: 3
Joined: 15-May 07
Member No.: 1910



Danny sullivan made a similar post recently at search engine land. Here is the link:
http://searchengineland.com/070426-011828.php
Go to the top of the page
 
+Quote Post

Reply to this topicStart new topic

 



RSS Lo-Fi Version Time is now: 2nd January 2008 - 12:19 PM