Forum Moderators: Robert Charlton & goodroi
I would like to know how a page containing 100+ links is processed by Google. Does it read the first 100 links and ignore the rest?
I would also like to know how the PR is distributed when a page has many links. We all know that generally all the links of a webpage will inherit PR-1 (PR minus 1). But what happens when it's a farm link? Will the X first links inherit PR-1 and the others nothing? Will inheritance PR value decrease following the number of links and be the same for all of the links?
We all know that generally all the links of a webpage will inherit PR-1 (PR minus 1).
Huh? Can you elaborate on that please? Is this external links going out to other sites, or internal sites links, or both? If this is external, then every site that links to me is giving me a MINUS 1 PR?
Thanks.
I couldn't find anything definitive on your question and it's been asked over and over again. I think it's a good idea to keep the links to less than 50 on a page if it's "link exchange" type page simply due to the potential confusion of so many links. If says "100", then it's a good idea to assume they won't see anymore links than that, or may not FOLLOW them, and I've also heard the G Bot stops after 100k of the page's size. But, that's not true, I just tested that. I have a couple of pages at my site that are a "specifications database" of some kind and the pages are about 135k in size. The pages are 1st on the 1st page at G for their [widget widget database] search, and G shows them as 136k in size. I went down to the extreme bottom of the page and got some of its text and I put that in G IN QUOTES to see if it was indexed. Strangely, there was only ONE RESULT at G, and it was the almost exact same 135k page but on another site, and it was in raw text format! But when I clicked the "In order to show you the most relevant results, we have omitted some entries very similar to the 1 already displayed. If you like, you can repeat the search with the omitted results included, that page of mine WAS included on the 1st page. Why it didn't show originally I don't know, but the fact is G indeed WILL index pages of at least 136k in size.
So that will at least tell you something about page size, and that everything on it is at least indexed.
I did some more looking into this and from what I see, it's a myth. Here's one quote: "I just went to the guidelines page at Google and saw that this recommendation is for ranking well". From what I've been able to ascertain, you don't get penalized for more than 100 links, G does indeed see more than the first 100 links, but you'll get a higher PR if fewer than 100 links. I SM'd you a search result URL.
This is not always the case anymore, in the past it was almost guaranteed a site would gain PR-1 from a link, but not so now, other factors come into play now.
It's also commonly known that sites that have many links do not pass the full strength of PR to all links. PR dilution comes in when to many outbound links are found. I would say this can kick in around the 50 mark... but again other factors can effect this. Nothing is set in stone, there are far to many factors in play, one of the important ones being sites history.
The PR-1 theory is not one I agree with, although in certain circumstances it may be true (usually when a page has 10 links or less). There is a problem with even that simple scenario though:
Once you get more than 1 link, how do you estimate the incremental benefit of a given link? Pasted answer below.
A link from a PR4 page could give you a PR3 if it's your only inbound and there are not too many links on the page. That's no good as a measure when you are trying to find out how much benefit it will give you on top of your existing PR.
When trying to explain PR to anyone that cares I have always used a 'PR points' analogy as such (which is simplified and ingnores dampening etc):
1. You need (your guess here) 10 times more points to get to the next level of PR than the previous one so PR1 is 10 points, PR2 is 100, PR3 is 1000 etc
2. You don't know if a page is a low or high PR? so there could be up to a ten fold difference in the benefit you get from a link from any page depending on that, hence assume a mid value for calculations, e.g. a mid PR2 has 500 points to 'give out'
3. Also take your score as a mid value unless you have a better idea of where you stand.
4. Calculate the benefit of a link from a given page and by dividing the points that page has to 'give away' by the number of links that will be on the page with your link included. eg PR2 page with 4 current links= 500/(4+1)=100 points
5. Add this to your point score, do this with all links gained and you get a rough guide as to your upcoming PR.
E.G. a site has no inbounds but gains these:
PR4 with 4+1 links = 10000/5 = 2000
PR6 with 49+1 links = 1000000/50 = 20000
PR5 with 9+1 links = 100000/10 = 10000
PR3 with 4+1 links = 1000/5 = 200
Total = 32200 points = Low/Medium PR4
The very important part of this is that to get from PR5 to 6 you will need loads of low PR links, a few high PR links may do it.
> I've read on Google's site that it is strongly recommanded to limit the number of links to 100 maximum per page.
I was surprised when I saw Google add that to their site. The >100 links on a page thing was already part of their rankings but they don't make a habbit of providing useful SEO tips. ;-)
> Does it read the first 100 links and ignore the rest?
At various times, Google have given considerably less PageRank than the normal calculations for pages with more than 100 links. At other times, the links have been divided normally. While in effect, this has applied to all links on the page. I don't know the position now; I haven't checked for months.
From a practical point of view, I would recommend having less than 100 links per page.
> I would also like to know how the PR is distributed when a page has many links. We all know that generally all the links of a webpage will inherit PR-1 (PR minus 1).
The PR-1 rule was never the case; Google divide the _raw_ PageRank (minus a small proportion) by the number of links on the page. Two or more links to the same URL count as one link. Links to dead (404) URLs have sometimes counted, and sometimes not.
When the raw PageRank donated from the incoming links has been added, the Toolbar and Google Directory PR graph values are calculated by taking a log of the raw value. _If_ you have a roughly normal number of links on the page, then the PR given is roughly one less than the PR of the donating page (hence PR-1).
It should be noted that 5+5 = 10, but ln(e^5+e^5) <> 10
There was some conjecture that Google picked the log base of the Toolbar (or the coefficient if log base 10 or e), to match a PR-1 drop from PageA to PageB, where PageA has roughly the average number of links per page. I don't know if that was true.
> But what happens when it's a farm link? Will the X first links inherit PR-1 and the others nothing? Will inheritance PR value decrease following the number of links and be the same for all of the links?
If the URL has been identified as being part of a link farm (or some other types of pages that Google don't want to count), then the page itself might be assigned zero PageRank, or it could be flagged with "can have PageRank but not pass it on" stutus. The latter has been the case for an increasing number of PageRank link sellers since Januarry 2002 and, at various times pages with guestbook-like URLs.
What with Sitemaps being used its not even a spiderring issue anymore.
Thanks for correcting my thought that "pages usually inherit the PR of the inbound link minus 1" ...
Huh? Can you elaborate on that please? Is this external links going out to other sites, or internal sites links, or both? If this is external, then every site that links to me is giving me a MINUS 1 PR? Thanks.
hmmm it's just something that I've noticed (seems to be false, obviously).
The PR showed in the Google toolbar is not up-to-date : it's the PR of the last "Google Dance". In fact the PR value changes in real-time (Googleguy said it on the forum).
I've found a tool on the web which allows to see the realtime PR value of nearly any site and here is the result :
My home page has a realtime PR5 with a fair number of links on it (less than 50)
=> all the pages linked from my homepage have PR4 (ie: PR5 minus 1)
=> Same thing for the pages linked from the links of my homepage : they have PR3 (ie: PR5 minus 1 minus 1)
=> More generally I have noticed that the deeper you go, the more PR you "lose" (-1 per level)
This brought me to think that a page "generally inherits PR-1 from the highest ranked backlink" (conclusion made by myself).
As this "rule" was true for internal links (at least those of my site), so I tried to see if it was also true for external links. So I went to the DMOZ which has very high-PR categories and tried to analyze how the sites of this webdirectory profited from DMOZ' high PR. I found out that when DMOZ was the highest PR backlink of a site, this site inherited PR-1 (to verify the other backlinks of one site, I just used the "link:" command on Yahoo or MSN, which are more up-to-date than Google's "link:" command, then checked the realtime PR of each backlink).
That's why, now, I always try to put in the footer of all the pages of my site a link to the most important pages, so that when a page is linked from a high-PR site, it will directly benefit to the pages of my footer.
Well that's all.
If anyone wants more details about :
- The URL of the tool which allows to see realtime PR
- or the URL of my site (that I used in the example described above)
- or the URLs of the DMOZ sites that I've tested
Don't hesitate to ask ;) (I've not mentionned any url to prevent spam/ads)
And sorry for the inaccurate english (I'm french ;) )
Quite funny ... Google recommends not to put more than 100 links in one page, but they do it :D (122 links)
Well anyways it proves that when a page has more than 100 links, those links are ALL crawled by Googlebot (there must be a high limit though). However, like "inbound" and "cml" said, PR transmission is more than surely diluted in that case.
No. That notation is "maths speak" for "your page will have a PR value one less than the page that links to you". It doesn't mean "a PR of minus 1". But in reality it rarely works like that at all.
1. You need (your guess here) 10 times more points to get to the next level of PR than the previous one so PR1 is 10 points, PR2 is 100, PR3 is 1000 etc
Actually, FYI the breakdown is:
Toolbar PageRank.....Real PageRank
0..................0 - 10
1..................100 - 1,000
2..................1,000 - 10,000
3..................10,000 - 100,000
and so on.
Otherwise all sites of the web would end up with PR0 :D
This is not always the case anymore, in the past it was almost guaranteed a site would gain PR-1 from a link, but not so now, other factors come into play now.
How do you gain, a minus PR? I know you're saying that is no longer the case (I guess?), but what other factors now come into play?
Also when you say "It's also commonly known that sites that have many links do not pass the full strength of PR to all links", do you mean entire websites that have "many links", or webPAGES that have "many links", and you talking about outgoing links to OTHER sites, or links to other pages at YOUR site?
I'm really getting confused about this. I have the same PR on my homepage, as I do many of my other pages. One of those other pages is one of my link exchange pages (page 1) but page 2 has a PR 2 pts. lower! To my knowledge, NO ONE is linking to my link exchange pages! So, I don't understand this....nor how can so many of my pages have the same PR has my homepage.
Thanks.
There were examples (Chris_R 2002) to show that the PR probably can go above 10 on the Toolbar.
Due to the relationship between the values wer're talking about, log base and coefficient, we can't infer raw PR numbers in a direct way. Fortunately, we don't need to - that's partly why logs are useful for this type of scale.
The factor from ToolbarPR(n) to ToolbarPR(n+1) is not 10.
I believe that the intention of G's recommendation about less than 100 links per page is to enhance its crawling efficiency. Less number of links on the page helps its bots to crawl faster and better; it does not stop crawling after 100 links. It's mere recommendation, but people take it as the Ten Commandments. From there, people began to break down their long link page down to smaller ones with directory-style pretension.
I believe that the intention of G's recommendation about less than 100 links per page is to enhance its crawling efficiency.
Not really. It is more about encouraging usability.
A well designed page can carry a few hundred links without seeming excessive, but there aren't many well designed pages out there with lots of links.
Google used to ding you in PR passed-on for having somewhere in the neighborhood of 200 links. Now they seem to just ding the page with excessive links in the rankings themselves.
Google will crawl and pass PR to every link on your page. They always have.
