Forum Moderators: open
Here is a small list I gathered (each name followed by a .com) -
Google
Adobe
Macromedia
Real (not sure - the page jumps)
AltaVista
HP
Apple
Suggest more?
seoRank
Now for simple link popularity, let's go with the numbers produced by www.alltheweb.com. They count all the *external links* to a domain, and not just to a page. They also don't deal with PR, so they mention all the pages that they have in their index.
The external links is an important distinction here. The vast majority of yahoo's links are from internal. In fact, yahoo has so many pages that they are not all in the google index.
So here are the numbers of external links.
Yahoo - 5,626,684
Google - 10,934,186
Looks to me like the Google domain (and that is just the .com) wins by almost a factor of 2.
Now when I check the number of pages in the google index for yahoo.com and for google.com
Yahoo - 15,000,000
Google - 612,000
So yahoo has to spread half as many incoming links over 25 times as many pages.
I'm not sure what version of alltheweb you're using but I'm counting 76 million external links not 5?
[alltheweb.com...]
To properly compare to google, you would have to check all the numbers on all their country sites.
And that still doesn't cover the fact that Google is linked from an awful lot of high PR pages that do not mention yahoo.
You also are talking about the traffic to the yahoo site. PR is page based. I go to hundreds of yahoo pages a day, but I don't think I have gone to their home page in 6 months. I go to the google home page dozens of times a day. I am almost positive that the google home page is significantly more popular than the yahoo home page.
PR has nothing to do with traffic, it has to do with the webgraph of the links. Do you understand that PR and link popularity are totally different creatures? If you have not read the original papers on PageRank, I highly recommend them.
What it comes down to is that all those PR9 and PR10 links that google has from w3c, who, macromedia, apple, etc, count a whole lot more than the link to my yahoo profile from my personal home page.
If Google did not give the highest PR site on the web an accurate PR, that would be dodgy. But you have not shown any reason that google should not have the highest PR other than a hunch.
Using a logarithmic factor of 6.22
It's wrong of course, because one minute we talk about a PR10 'site' but then we are actually referring to ranked 'pages'. Still it was mildly diverting. Does 199 PR9 sites stack up against gut feel?
PR..Number of Sites
10...............32
.9..............199
.8............1,238
.7............7,701
.6...........47,897
.5..........297,922
.4........1,853,072
.3.......11,526,108
.2.......71,692,390
.1......445,926,663
.0....2,773,663,843
Total 3,305,017,064
Steve
Your numbers of sites match pretty well with Alexa positions for site of corresponding pagerank (After 15 million Alex stops working).
That is quite interesting.
log: 6.2
PR........"points"
10........13537087
.9.........2183401
.8..........352161
.7...........56800
.6............9161
.5............1478
.4.............238
.3..............38
.2...............6
.1...............1
.0...............0
So, I'm assuming if a PR8 page links to 20 pages, that would be worth 17,600 points. That would propel a lowly PR3 page to PR6, guaranteed.
If this is correct, it has obvious implications for how we value and "buy" links from high PR pages- assuming of course that the PR will be transferred and will boost my position in SERPs. Of course, traffic is important too, but for now I'm mainly interested in figuring out the worth of PR.
Not that I'm considering "buying" pagerank. ;)
Of course nobody really knows for sure, but I for one would say a logarithmic scale matches what I have seen, and it's very hard to guess the base, since most of us are not in laboratory conditions and are just trying to help our sites :)
But I give those numbers the thumbs up. They at least "feel" good to me.
Does it really matter what base the log is in? To covert from on base of log to another you just simply multiply by a constant. For example log4 x is half the value of log2 x.
The simplest log to do on a computer is base 2 because numbers are stored in a base 2 format (binary) on the computer. To calculate log2 of a binary number you just count the position of the highest set bit. You can then multiply or divide that value by a constant to get a different base.
But my guess is that Google are using log but some form of linear approximation to convert the real page rank to the toolbar page rank.
The base would matter. If it was 2 a pagerank 6 would be worth double a pagerank 5.
If it was 10 it would be worth 10 times more.
No, Google has not acknowledged that it is a log scale. It is based on two things, first that it sure does look like some sort of log scale, and second that a similar graph shown in the early pagerank papers used a log scale.
And you are correct, the base of the toolbar PR does not matter, simply because toolbar PR doesn't matter. It also most likely changes from month to month.
Improving your real PR is a much more important goal than worrying about that little green bar. And the very best way to increase your real PR is to have a site that is worth linking to.
add phpBB.com
phpBB.com was PR 10 for a long time and then they become PR 9 and now back to PR 10 ...
Did anyone noticed those paid links in their homepage - they are not inheriting PR from phpBB :)
Great advice Dave, if you are providing information on a zillion products like you do. If you are a hard sell single product site (like mine), that is a different matter.
However just because I don't have thousands of product descriptions it doesn't make my site any less valuable than yours. My product is one lots of people need and which hundreds of other sites sell too. No lenghty description is necessary, or in fact possible. So how should I make my site worth linking to? that's a retorical question by the way
that's a retorical question by the way
Too bad, I'm going to answer you anyway ;)
Great advice Dave, if you are providing information on a zillion products like you do. If you are a hard sell single product site (like mine), that is a different matter.
No it is not. The better your site is, the more likely it will pass the screening of people that are likely to link to you.
If you get a link from the yahoo directory, or from DMOZ, your site needs to meet the requirements for them to link to you.
There are many different types of quality, you just need to provide the best type of quality that is appropriate to your site.
However just because I don't have thousands of product descriptions it doesn't make my site any less valuable than yours.
Certainly not to you. But it is obviously more important to the people that contribute to is and to those that link to it because they consider it to be quality content.
That is the whole point, you have to give value in some way to those that link to you. Having content worth linking to will even make buying links easier.
My product is one lots of people need and which hundreds of other sites sell too.
So, your site is one of hundreds, and it is as valuable as my one of a kind site?
Of course it is, because it is the one that puts food on your table. But it is all the more reason to be the best site on the subject.
Yeah, there are limits to what you can write about some products. But if your site gives a lousy shopping experience then when someone is deciding whether to link to you, they may just decide not to. Even things like how well your shopping cart works count towards the quality of your content.
