Forum Moderators: open
To give one example, my home page (which I think is also PR 8) mentions the film Sidewalks of New York. But on a search for "Sidewalks of New York" it ranks 530th. There may be several thousand links to my home page, but the chances are none of them contain the words "sidewalks, "new", or "york" - and 0 anchor weight x 1 link from PR8 + 0 anchor weight x 10 links from PR7 + 0 anchor weight x 105 links from PR6 + ... = 0. I suspect the only reason I rank even 530th on that search is that the word "new" accidentally occurs near some of the links!
There are two reasons why I think people get too carried away by PageRank. Firstly, it is the only documented part of Google's algorithm - the papers on PR were written before the company was founded and the algorithm became a big secret. Secondly, it's highly visible - through the toolbar - in a way no other element of the system is.
You run the other direction now in saying it is
not important for the place in the SERP. That is wrong in my opinion.
PR and keyword weight go together like "breathing in" and "out". We list high with low PR sites, but we have named links to these sites for our main keywords. We list low with high PR sites for other keywords, so far, so true. But a (quite) new site with lower pagerank and the (almost) same inbound links lists lower.
I just mean: PR without keywords and vice versa means nothing.
right?
Danny, do you think the "anchor weight" is a count of the number anchor texts that contain a keyword, or a ratio of the number of anchor texts that contain keywords divided by the total number of links, or some kind of keyword weight or prominence thing?
Is that what you mean Danny? If so then I agree 100%.
I haven't seen anything definitive regarding words inside links but I, like most on this forum, assume it is very important. But it's very difficult to control the language inside a link to you. So all you can do is try. You have a "link to us" page and make suggestions or you go out and request links in a particular form. But you can't absolutely control it.
Also, language in a link is fairly simple, even if you can't control it. You tell folks to link to you like "Find all the Widgets at Bob's Widgets" and that's the best you can do. Not very complex. PR on the other hand involves a lot of effort. How many inbound links, from whom, and how to interlink all your sites are very difficult questions. This is why I think we spend so much time on PR.
It's been said that there are so many elements to the serp formula that you can't really optimize for all of them. And that's part of the appeal of Google. And it's been said that the one element you can have some control over, after you've done all you other homework, is the inbound link thing. If you have great content and people respect your site, you can get lots of inbound links and that'll feed your PR which should push you up in the serps.
Your own site's pages should have good PageRank, if they link to other internal pages. Internal links within a site have some influence on SERP ranking, don't they? If so, and since this is the one area webmasters can control fully, then that probably explains why everyone cares about understanding PageRank and distributing it optimally within their sites.
Nick
you are right.
Hypothetically, if you would choose a search query, of which any of the words in the search query text, do not occur on any anchortext on the whole Google index, Pagerank would play a role.
The main value of Pagerank is the voting power it gives towards other webpages. If you do not motivate that power with an anchortext, you are just passing on Pagerank.
The other value of Pagerank is it produces the most threads in WebmasterWorld and it is the major ego valuator for webmasters. ;)
I am abolutly new in this: I do not know if all this already is something known and obvious.
...And I am no english speaker (only learning to read on the net).
Before knowing these forums I knew by my own one tells to the "google dance" and other things. Those coincidences made me "client" of the forum.
On the other hand, it is obvious that the PR is a relative question. I have pages in top 10 with PR 3 and 4 (on 200.000 and 2.300.000 total pages listed, respectively).
About general, I think that who made the search engines they tried to make them "humain", and for it they have been creating "outlandish" [extravagances] "algoritmes".
Although most "outlandish" he is not to have counted with linguistic experts, like me :-), at the time of making his "outlandish (żextravagants?) algoritmes".
The stupid results that give back all the search engines, in a part, are related to that deficiency of beginning.
Those that also studies the Search Engines on the other hand devise "extravagances" that sometimes agree, partly, with the extravagances of the creators of the finders.
A little is guessed right at random, without a method rigorously scientist.
The same it happens with "links."
There is something "with links, yes, but in 3 weeks of reading, I have not seen it here exposed the 100%.
I feel not to be able to express to me and with more precision in English better. I hope that the idea is understood.
Is everything a work craftsman, who never is going away to reveal in forums, certain?
In any case, thanks to all reason why contribute to these forums. Something is always learned really.
Sorry, "powered by Altavista traslation" .
[edited by: nobody at 4:35 pm (utc) on July 20, 2002]
Ranking = Pagerank (0-10) X Relevance* (0-10)
* (loosely definied as "everything and the kitchen sink")
If so, it is useless to argue which is more important.
I fully expect GoogleGuy to blow milk through his nose while reading WebmasterWorld over breakfast, shaking his head and feeling sorry for my little brain, but for a gross oversimplification this seems to explain a large percentage of search results. YMMV
my home page (which I think is also PR 8) mentions the film Sidewalks of New York. But on a search for "Sidewalks of New York" it ranks 530th.
PageRank is a mechanism used to determine the order in which equally relevant pages are returned. It is a sort function.
The degree to which PR plays a role in your own SERP's will always depend on the overall level of on-page optimization that exists within your space. There are many on-page factors that go into the mix. One of them is inbound anchor text.
By that I mean that Google considers inbound hyperlinks as an extension of your page.
If you think of it in terms of all the inbound link text that is out on the web pointing to your site, is actually on your page, I think it becomes a bit easier to understand. Going a step further, the weight that is given to these additional words added to your page will be dependant upon the PR score of the page containing the original link.
Danny's PR8 site doesn't show up for "Sidewalks of New York" because A) the page lacks the phrase in important areas of the page like the title, and B) there are no additional occurences of the phrase being added to the page through inbound link text.
The phrase being added to Danny's site via link text is "Book Reviews." The combination of his on-page optimization and his inbound links has placed his page #2 for the term "Book Reviews." (out of 3.5 million pages)
He is listed above the New York Times book review section (also a PR8) because the NYT wasn't smart enough to include the word "reviews" in their title.
The Boston Globe (#1) did use the whole phrase in the title, which enabled it to obtain the #1 spot for "book reviews" even though it is a PR7.
If you spent the time digging through all the back links, you would probably find that even though the Boston Globe's overall PR score is less than Danny's, the weight of the link text being added to the Boston Globe's page is probably greater.
All links contribute to increasing your PR score.
Links with relevant link text increase your PR score and and significant keywords to your page.
Danny's overall PR score is a significant factor that has provided him the opportunity to compete with the likes of the New York Times.
His on-page optimization combined with the anchor text keywords being applied to his page is what has enabled him to win.
If you take his home page and upload it to a different domain with an equal PR score, he would no longer show up at #2 for "book reviews."
Sean, your formula is simple, but it actually hits the spot. On the one hand, you multiply PR and relevance. The actual algo for the combination may be more complex, but both values are certainly not added to each other, and, so, it really does not make much sense to argue if PR or relevance is more important. On the other hand, the formula shows that relevance is limited. In "The Anatomy of a Large Scale..." Larry and Sergey say: "Count-weights increase linearly with counts at first but quickly taper off so that more than a certain count will not help." (Note: Count-weights also include the anchor text of inbound links.) So, the possibilities for on-site optimization are limited and other factors like PR become important.
PageRank certainly is a critical factor for the SERPs. But, of course, its importance shouldn't be overrated. Larry and Sergey say it this way: "We designed our ranking function so that no particular factor can have too much influence." I don't see any reason for them to drop this principle.
What I don't know is how the anchor text weighting works. I know that text not in anchors but adjacent to the links matters, but I don't know how far away. It would make sense if in HTML code like
<DT>widgets
<DD>good for sick cows
<DT>wodgets
<DD>good for sick horses
Danny.
