Can google stop link buying?

[cuzco]

We know google does not like webmasters buying links/PR, so what do you think they are doing to stop this?

How would they automate the detection process? What would they look for to distinguish link buying from natural linking?

Some have said the google “sandbox” effect is to stop link buying, what evidence is there for this? could the sandbox effect be to stop something else, like 100’s of doorway pages or sites that used to rank well for few months then die etc?

Thanks

[seobook]

We know google does not like webmasters buying links/PR, so what do you think they are doing to stop this?
try to plant seeds of doubt to make buying or selling pagerank less appealing until they can eventually come out with an algorithm that focuses on local hubs and less on overall web link popularity.

How would they automate the detection process? What would they look for to distinguish link buying from natural linking?

I do not work at a search engine, but they can look for patterns. Sites with many sitewide links. Sites with many links in a row...stuff like that.

Some have said the google “sandbox” effect is to stop link buying, what evidence is there for this?
some sites that should have ranked well based on extreme amounts of optimized inbound text link link popularity were not / are not

could the sandbox effect be to stop something else, like 100’s of doorway pages or sites that used to rank well for few months then die etc?

100's of doorway pages cost a decent amount of money to set up. Google can and most likely significantly does devalue links from the same IP C block. setting up hundreds of different hosting accounts is hard to manage. also all those different sites will not have any link popularity to parse out unless the webmaster can get other sites to link into it.

[cuzco]

some sites that should have ranked well based on extreme amounts of optimized inbound text link link popularity were not / are not

100's of doorway pages cost a decent amount of money to set up. Google can and most likely significantly does devalue links from the same IP C block. setting up hundreds of different hosting accounts is hard to manage. also all those different sites will not have any link popularity to parse out unless the webmaster can get other sites to link into it.

Thanks


Maybe i should of left the last paragraph as “could the sandbox effect be to stop something else”

As far as i was aware the sandbox effect stopped new sites ranking for a few months no matter where the backlinks came from. That would only effect link buyers for a few months when most link buying is an ongoing year round strategy for some SEO's

Doorway pages was just meant as an example of something else that sandbox effect could be going after.

I also know of at least one company that specialise in creating 100’s of doorway pages, it might be expensive but they aren’t cheap, as soon as one page gets hit that make another – they’re subject of many ethical forums.

[Irony]

Who would link to an obvious doorway?

I would not. Even for money.

And if a site at a doorway domain is good enough to link to it - I think it is not a doorway anymore... but just another site? Or am I wrong?

[cuzco]

Forget about doorway pages for now.

Back to link buying.

I think we can all manually spot most sites selling links.

I am looking for peoples thoughts on what factors google will look for to automate the process – all speculation but some might be spotting anti-link buying already.

Curious what strategy’s google will implement to combat link buying without effecting natural linking.

One possibilities of the top of my head.
Only count x amount or a % of the links from the same domain/IP to site A. –What legitimate sites would this hit?
Other ideas?

Thanks

[Irony]

Suppose a webmaster looks for relevant sites (say, 100 of them) with good PR... or good positions in SERPs... or whatever. Then buys exactly one link from each.

It's a hard job, of course, but not harder than any other link building activity, be it a reciprocal linking campaign or anything. But the resulting pattern will be totally undetectable.

There is nothing Google can do to stop it IMHO

[seobook]

There is nothing Google can do to stop it IMHO

they can delay pagerank update
do the short term sandbox thingie
block pagerank from being parsed on some sites
pr 0 some sites for blatently selling tons of links

the biggest thing they can to is plant seeds of doubt to lower demand and raise costs to make ROI less predictable. Of course as they lower demand supply will exceed demand and lower price to a new equilibrium.

they will eventually move away from pagerank some and more toward their hilltop and localrank stuff. in the meantime their main defense is raising cost and extending out the time required to create a bogusly high ranking site (thus further raising costs).

[rustybrick]

They might eventually move away from PageRank but it won't stop the selling of links for keyword rich anchor text.

What might slow the business of selling such links are the other points you mentioned. But I know that link building companies have already made adjustments to their strategies to think more in the future.

[Anthony Parsons]

Ok then! How about this theory. I don't think it could be possibly for Google to track the selling of links automatically. Impossible feat even by super computer.

You could analyse website links when the bot runs this time, then next time and detect increases, but really, how in gods green earth could you tell which sites are selling them. The site that adds one link or 100 links between crawls? That is a natural part of every site. Links go in and out, people change their sites daily and lots more.

IMO, the hubs would be where to start and the places to monitor. Such hubs as Linkadage, etc who are clearly selling PR value and not advertising would be the spots to monitor. That is the only logical solution I could think off at present, but certainly not automatically. The SE's could set up something similar to spam report, but then like that, nothing would likely happen.

Then with everything, there are simple ways to get around it. If you where selling PR, something like, $100 per month link on this PR8 page you could change to many possibilities:

# $100 per month for link listing upon this page
# This page receives "x" visits daily. $100 per month 3 word phrase listing. Download Google toolbar for further market analysis. (maybe a link to a page explaining what PageRank means on a marketing scale)

This could be a real tough one to monitor. Lets face it, everyone does it if they sell advertising because you always consider the PageRank within the costing. Pagerank + page views = $$$.

[rustybrick]

Yup, the nature development of the Web is by links. Viral marketing is all about Web site plugs. You can not simply look at percentage increases or decreases in link count.

[cuzco]

OK, so google can not stop all link buying, and buying single links on high PR sites will go unnoticed unless someone opens their big mouth :D

On these auction/link buying sites you can see people offering 5k-10k PR5/PR6 links – basically every page of their site.

Wouldn’t 5k links from one site look a bit fishy?
What if the buyer had the majority of their links from this one site?
Surely this is much easier to spot and do something about?

Doubt they would ban a site if this was auto-detected but maybe they would just stop the seller from passing on PR.

Or maybe they would only count a few of the 1000’s of links in the ranking if its seen they all come from the same site. Just a thought!

[Anthony Parsons]

Links all through the pages are irrelevant really as they all come from the same domain. Links on each page are for advertisement purposes, but the domain will only technically pass the one qualified link.

[David Wallace]

I don't know if this would stop it or not but it would definitely hinder link buying:

Get rid of the PR indicator and backlinks checker. If someone cannot see their PR and if they cannot check backlinks at Google, how will they sell links based on PR6, PR7, etc., etc.?

[Brad]

Google will not stop link buying. In fact my prediction is that it will only increase. Too many people know about it and there is too much money to be made buying it and selling it. All you have to do is never mention PR and just call it advertising or sponsorship or whatever.

I agree that they could slow things down by getting rid of the PR bar on the toolbar and directory but they seem reluctant toget rid of the one marketing gimmick that keeps their toolbar on every SEO's computer. ;)

Besides, Google is selling Adwords to people that are selling (or brokering PR) so they help fuel the obsesion with PR and get paid for it. :eek:

[Webmaster T]

We know google does not like webmasters buying links/PR, so what do you think they are doing to stop this?
Where does it say Google doesn't like this? It's not mentioned in the guidelines or anywhere else I could see. What they don't like is what SearchKing did which was to sell links based on PR. Buying links is an ad spend if Google tried to control that they'd have a very bad affect on content sites and niche directories that depend on ads for their livlihood. The sites that are setting up as link brokers and "PR Clubs" likely have some bad news coming but IMO, if you stay away from these buying ads/links from relevant or targeted sources is a good strategy that is being rewarded.

[bhartzer]

The only way they'll stop PR buying is through a manual review. Once G. goes public, they'll suddenly have a lot more money to hire more people to do manual reviews. And since spammers are really G's biggest threat, I'm sure we'll see a lot of changes after the IPO.

[Jeff Martin]

Let me offer this view:

Whats wrong with buying links? Any time you buy an advertisment, a link back to your site is a given. Why would you place an ad where no one could reach your site?

If I have a PR 8 site and I offer ads for run of site and sectional like every other web site on Earth why would Google need to do anything about it? What "law" have I broken? None. To say its wrong would be like penalzing SEW because of all of the SEO advertisers Danny has.

Now if I publicly market and base my prices of my inventry based on PR, then I am using a Google service to make money from, which Google has publicly said is a no-no (Search King anyone? Love ya Bob!).

Thoughts?

[cuzco]

I presumed everyone would know I meant buying links for PR -sorry. There are hoards of sites out there selling unrelated advertising based on PR.

So just to clear things up, this is the kind of text links I am talking about. Links that are bought just for improving ranking.

Have seen a few networks get hit for selling links, but as far as i am aware these were hit after the owners went around blatantly adverting PR for sale.

Wanted to see if people have seen any signs of google going after these sites other than manually, some have suggested “sandboxing” is an attempt at this. Also what precautions people are taking when buying links(if any), and what they think google will do next – all speculative but something i find interesting.

Didn’t really want to debate if it was right or wrong for google to go after these sites, i buy link *advertising* myself - in the end its up to google if they like it or not.

[Phoenix]

Google will not stop link buying.

But they could! Turn off the the green bar. Yet at least they have tried, in the form of manual penalty or automatic in the algo, its depends. As bhartzer pointed out an IPO might give it the fuel to develop solutions to combat link buying. Yet, why the spammer are the parasite on the back of the camel, its doesn't mean its Google's biggest problem. I am sure they have bigger fish to fry and I don't mean professional spammers. They will and always will do it one way until it gets stopped. However, I think the "sandbox" effect doesn't necessarily have all that much to do with link buying only. There is a lot of things it does to fix certain problems with PR. In my mind I can see it as benefical to some degree.

I also read in another forum, that Google might have possibly prevented PR8-PR10 and up links from passing on pagerank. (searchking, phpbb.com, etc..) I know I have several PR8 links that don't show up in the backlink checks. Yet, some do. I ones that show up are natural links. So maybe its a slow progression of stopping the link broker or monger from corrupting what they developed as a tool for a specific use other than it used for now (partly).

:)

[Mel]

I doubt that webmasters who are making large (dollar volume) buys of links are doing it for the PR as much as for the anchor text links. Still the defacto standard for evaluating the price of a link is the PR.

I would suspect that Google could (and probably would) step on sites that blatently sell links on the basis of PR, but as far as putting the link merchants out of business, I would suggest that simply implementing a version of LocalRank (probably in addition to PR, not in place of) that also removes the link from the relevancy equation would create a situation where a large number of links from any one site would be removed from the equation and thus they could only sell one effective link per customer and keyword per site.

When (if?) the buying public ever get wise to the fact that a text link from PR2 site is just as good as a text link from a PR8 site, and that the ranking factor of PR is not all that great, this will largely disappear in the form we see it now. Could be that an education campaign would also do the trick.

[bhartzer]

The spammers are one of Google's biggest threats to their business model. They say that in their SEC/IPO document. I believe that once they IPO they'll hire a buch more people to go after the spammers.

[Mel]

Putting the thread title and your post together I assume you are equating the buying of links with spam?

If so then I could not agree that all link purchases are spam, as then you would have to say that every paid advertisement on the web is spam.

But to be the devils advocate for a moment if you make the assumption that paid for links are spam, then what of reciprocal links where you do not pay in cash, but barter for the links? What about links from the Yahoo directory (or from any of the paid inclusion directories for that matter) are they spam too? Are we going to start saying that links from pages that have paid for inclusion in search engines are spam too?

Certainly there is a point where easily obtained links (like those that you just buy) should be devalued if the search engines are to maintain their relevancy, but where are we going to draw the line? I suspect that we may be seeing a devalation of forum signature links in the recent Google links update, and that is probably a good start.

IMO the means to drastically reduce the number of paid for links is available to google via their LocalRank patent but since that is not happening is it a question of when or if they are going to do it?

[Mel]

Links all through the pages are irrelevant really as they all come from the same domain. Links on each page are for advertisement purposes, but the domain will only technically pass the one qualified link.
Tony could you explain what you mean here? Certainly IMO this is not true for PageRank which only considers page, not domains, in its formula, but are you thinking that text links are handled in this way, and if so how do you know?

[seobook]

Putting the thread title and your post together I assume you are equating the buying of links with spam?

If so then I could not agree that all link purchases are spam, as then you would have to say that every paid advertisement on the web is spam.

Totally agree there Mel. There is no black and white really when it comes to this issue. Anchor text is extremely powerful and until their algorithm can learn to devalue random unrelated links link selling will remain a big business.

I don't think they need more capital to fight spam. I think more likely that they may not want to make extremely huge changes to their algorithm right before an IPO. I think they might do some more aggressive stuff after the IPO, but I do not think the need for more money or better engineers is the problem. They probably do not want to change their core product that much right now (as they do not want to lower their relevancy or get more bad press about how their algorithm changes have put small businesses out of business).

[robwatts]

Agree with most of the sentiments.

Just wanted to toss in a point about so called 'natural' and unnatural links.

Many of the strategies to which are being referred, are often seen in grouped links at the bottom of pages or in some column to the left or right of a table.

Just a thought here, but if google (http://www.google.com) wanted to identify more natural linking patterns it might help to consider that many people who link to sites in this way often link in a way that embeds the link itself within a paragraph or group of other non linked words. The 'vote' for the other site is often contained within the context of the discussion.

Links found in this way are far more likely to be genuine votes for other sites in that they havent been created for the purposes of ibl creation or pr inflation. They exist because the site owner or authour took the view that they added genuine value to the point that they were making and chose to link to it to illustrate or add value to his point. Whereas in many many cases footer links to external sites are often simply building pr or attacking the ibl aspect of the se algo.

Im not saying that they should completely devalue footer type links, or ratchet up the importance of plain text surrounded links, but it would be interesting to see the impact, thats for sure. :)

[seobook]

totally agree there. if they did a deep discounting of links outside of content then they would curb a bunch of the link selling.

many people are willing to give up a banner or two worth of space to make thousands per month, but a much smaller segment of the web population is willing to give an extra screen worth of content to sell a bunch of links. plus adding a whole bunch of random giberish content no only causes people to lose trust in you but it also reweights the contents and meaning of the page significantly in the eyes of search engines.

[Anthony Parsons]

Tony could you explain what you mean here? Certainly IMO this is not true for PageRank which only considers page, not domains, in its formula, but are you thinking that text links are handled in this way, and if so how do you know?

What I'm saying Mel is, I don't think Google is dumb enough that if you have one site with 1000 links throughout that site all pointing to the same place, that it is going to count those 1000 links, instead mark it as one link.

Now if you placed a unique link on each page pointing to a different page (and each page being like content) then you would see some benefit from it. What many people seem to forget is that it isn't just about PageRank. Google clearly stipulate that it goes further than just assigning PageRank, and that each page must be like / relevant content to give any benefit what so ever.

1000 links through a site all pointing to the same location won't do squat. Google is a lot smarter than that.

My directory for example is sitting on a PR5. Big Deal, who cares. It has links from directories and other like sites pointing at it, thus giving some importance within those links. I have a PR8 homepage link from a web design site pointing at my homepage and it hasn't done anything significant to the PR value of that page or site as a whole for quality purposes. A web design site pointing at an Internet Directory....relevance is nothing. Now if I changed the content of my homepage to reflect more in the SEO aspect instead of having my SEO info on a sub-domain, then that directly benefits the page and would probably see an increase to PR7 on the next update.

Relevancy is the key, not quantity of links stuffed into pages pointing at a site.

[seobook]

Relevancy is the key, not quantity of links stuffed into pages pointing at a site.

PageRank is fairly independant of relevancy.

[Mel]

What I'm saying Mel is, I don't think Google is dumb enough that if you have one site with 1000 links throughout that site all pointing to the same place, that it is going to count those 1000 links, instead mark it as one link.

Now if you placed a unique link on each page pointing to a different page (and each page being like content) then you would see some benefit from it. What many people seem to forget is that it isn't just about PageRank. Google clearly stipulate that it goes further than just assigning PageRank, and that each page must be like / relevant content to give any benefit what so ever.

1000 links through a site all pointing to the same location won't do squat. Google is a lot smarter than that....

Relevancy is the key, not quantity of links stuffed into pages pointing at a site.

I agree that is a nice idea but I do not see that in what Google actually does today. I have seen 5000 links pointing to one page with the same keywordtext shoot that page to the #1 ranking for a very competitive term literally overnight and I have seen rings of interlinked sites take the prize positions as a result of site wide anchor text links.

But if you can show us where Google is doing something that takes relevancy and the number of links from a site into consideration, I'm always willing to learn.

[Anthony Parsons]

Just go and have a close look at some of the most popular highly competitive terms. You will see for an example:

#1 has 1,000,000 links
#2 has 5,000 links
#3 has 1,500,00 links

In such a competitive arena for those type words, link quantity has nothing to do with much in Google. You assess some of those sites that register hundreds of backlinks from the same site. It registers, but they all point to the same page. You can do this and rank a page overnight, then they dissappear again and that site owner becomes another whinger when they dissapear. Why? Because Google reads what you did and takes it away automatically. Revancy is the key in links. If the pages your linking between are not relevant, you lose just about everything and end up another site who tried 30 times harder than a smart operator to achieve the same ranking. Lets face it, when you step into the competitive arena, links are everything. You can do it the smart way or the hard way.

Google even state PageRank works on relevancy, so you cannot deny that fact. The other major engines state they want relevant content between links. It is only peoples assumptions that change this. You can go and put your link on a PR9 Cooking sites page, point it towards your hardware page and your going to see just about nothing in value from that link because they are not relevant and the link should not exist for PR purposes. You might get lots of traffic, but not Pagerank. Put the same link pointing towards your restaurant site and well, you could get a PR8 instantly and blitz the rankings for relevancy because you just obtained such importance from a complementing site. Its all relative.

[AussieWebmaster]

Okay if a site has a thousand links to one page...
Well if it has say 500 pages with two links to the same thing on each page it would seem feasible... someone creates a template with a link on the left and a link on the right to the same place... 500 content pages later...
I would say an authority site is making a huge vote for a single place... how does Amazon, Google, etc get PR10s...

[AussieWebmaster]

Just go and have a close look at some of the most popular highly competitive terms. You will see for an example:

#1 has 1,000,000 links
#2 has 5,000 links
#3 has 1,500,00 links



My guess would be #2 is either new site and popping in for it's 15 secs of fame... or is content rich with the keyword... or has 2,500 links from each of #1 and #3 pages

[Anthony Parsons]

My guess would be #2 is either new site and popping in for it's 15 secs of fame... or is content rich with the keyword... or has 2,500 links from each of #1 and #3 pages

Now come on.....lets think about this!

#1 - New sites don't normally pop into the SERP's with 5000 respondant links
#2 - Content becomes a small factor within the highly competitive arena
#3 - 2500 links from #1 & #3? A real possibility. Though would you do it with a direct competitor? I think not.

I think it is interesting when you actually go and analyse some top highly competitive terms. Pull apart a few of the top ranking sites and see what makes them top ranking. You see it more and more now, especially with Googles backlink tweaking, link relevancy is what its all about. PR is nothing.

PR10 sites get their PR10 from a combination of quantity & quality. Buts lets face realism shall we. PR10 sites are only a few and they are the hub of the net, such as Google.com, w3c, adobe, etc. Look at the theory behind those sites. You link to google because its google, you link to w3c because they are the web compliancy standards, you link to adobe because they hold the worlds leading software and pdf reader, etc.

A PR10 site does not need a PR10 sites to link to it. How did a PR10 site ever eventuate to begin with? They eventuate because they have relevant sites linking to them. Many people seem to think you need a higher PR site to get a higher PR yourself? Couldn't be further from the truth. If you're a webdesign site with no PR value, and you get 20 websites in the SEO and graphics industry to give that webdesign site a PR5 page link each, you will find that the webdesign site will most likely get a PR6. Why? Because it only has 20 sites linking too it, but all are off relevant content, not higher PR value. That is how a PR10 sites gets to PR10.

We can't confuse PR with rankings though. Yes, I don't disagree that 5000 links from one site may give you a ranking boost, though if the linker was linking each of those 5000 page links, to 5000 of your pages, each page link containing like or relevant content, the difference that would make would be enormous. Time consuming, but well worth it.

Back to the original point though, most of those mass link sites are only going to give you mostly rubbish links. Is it worth the payment? I think not. I would rather pay the same amount of money for the one good relevant link that will do more good overall than 5000 rubbish mass site page links. Yuk!

[Mel]

...
Google even state PageRank works on relevancy, so you cannot deny that fact. ...


Can you point me to where Google says this? And are you suggesting that the proprietary Stanford software called PageRank (which is only licensed to Google) is now being replaced by something else and still called PageRank?

I totally agree that PR is a very small consideration in google rankings and it has been that way for some time now, but I still see anchor text links working very well for rankings, and I don't see that the relevancy of the links seems to be a factor in PR at all.

[Anthony Parsons]

Absolutely agree MEL,

I still see anchor text links working very well for rankings, and I don't see that the relevancy of the links seems to be a factor in PR at all.

Anchor text is one aspect, but the relevancy of that page the link is anchored upon is the key ingredient to PageRank.

Source (http://www.google.com/technology/)

PageRank relies on the uniquely democratic nature of the web by using its vast link structure as an indicator of an individual page's value. In essence, Google interprets a link from page A to page B as a vote, by page A, for page B. But, Google looks at more than the sheer volume of votes, or links a page receives; it also analyzes the page that casts the vote. Votes cast by pages that are themselves "important" weigh more heavily and help to make other pages "important."

Important, high-quality sites receive a higher PageRank, which Google remembers each time it conducts a search. Of course, important pages mean nothing to you if they don't match your query. So, Google combines PageRank with sophisticated text-matching techniques to find pages that are both important and relevant to your search. Google goes far beyond the number of times a term appears on a page and examines all aspects of the page's content (and the content of the pages linking to it) to determine if it's a good match for your query.

That's the problem Mel. The full and complete story is a different thing. Note what google says about important pages, then their definition of important pages and how it analyses them and the links upon them to match for relevance towards a site and query. Anchor text is one aspect, but useless if the page it is listed upon does not relate to the page it points towards. Hence, 1000's of links across a site is useless if the pages are not relevant to each page they point towards.

[qwerty]

Google combines PageRank with sophisticated text-matching techniques to find pages that are both important and relevant to your search.
I've always taken this to mean that Google looks for relevance based on the text on a page, but not that it's part of the PageRank calculation. PageRank is based on the number of links pointing to a page, the PR of the pages on which those links are located, and the number of pages to which those pages are linking. However, relevance doesn't enter into that particular calculation.

What I think G's saying in that quote is that PageRank is a factor, and relevance is a factor, but not that relevance is a factor in PageRank.

[Anthony Parsons]

Mel pretty much well answered the question by relating to PR10 sites. The first PR10 site did not come about linking to higher PR sites now then, did it?

Page relevancy that the link is listed upon is a must consideration when you simply apply commonsense IMO. If not, then PR would not exist because where would it start?

As I said earlier! Example: SEO sites, all PR8 homepage, 20 off, link to a web design page. Possiblities? Web design page takes up a PR9 because so many relevant content sites / pages, descriptive anchor text included, link to that page, hence how a higher PR develops to begin with. I have never seen PageRank as a pyramid structure.

[Mel]

I really don't think anything is put on Googles webmaster pages before it has had a full treatment by the resident public relations spindoctors, and therefor do not look for them to reveal anything about thier algos through that process, though they often offer tantalizing hints which can be taken in one of several ways, any of which may not be correct.

Here is how I would personally interpret that quote:

PageRank relies on the uniquely democratic nature of the web by using its vast link structure as an indicator of an individual page's value. In essence, Google interprets a link from page A to page B as a vote, by page A, for page B. But, Google looks at more than the sheer volume of votes, or links a page receives; it also analyzes the page that casts the vote. Votes cast by pages that are themselves "important" weigh more heavily and help to make other pages "important."

This says that google looks at the PR of the page that casts the vote, and that is fully in accordance with the Published PageRank equation...

Important, high-quality sites receive a higher PageRank, which Google remembers each time it conducts a search. Of course, important pages mean nothing to you if they don't match your query. So, Google combines PageRank with sophisticated text-matching techniques to find pages that are both important and relevant to your search. Google goes far beyond the number of times a term appears on a page and examines all aspects of the page's content (and the content of the pages linking to it) to determine if it's a good match for your query.

Here is the first hint that this might not be a show and tell of the Google algo, since I pretty well think everyone agrees that sites are not the recepients of PageRank Pages though pages are.

IMO this is simply saying that Google uses more than PageRank to judge the page, but it does not say that they are using relevancy terms in PageRank, if fact it says that Google uses more than PageRank, it also uses sophisticated text matching techniques, a process which is described clearly in the Stanford paper Anatomy of a Large Scale Hypertext Search Engine, which describes briefly how this process works, including the fact that the text matching and PageRank terms are seperate and added together after the relevancy matching is done. In that paper they describe how they look at the text surrounding the searchterms for more information and the fact that anchor links are also used.

All in all, I see this as nothing that cannot be found in or Page and Brin. Google have carefully woven two topics into what to the casual observer might appear to be one, but in fact they are not revealing anything that you cannot find in the Standford papers published around the year 2000.

It is IMO well to remember that PageRank (while often considered a proprietary portion of the Google ranking algo) is in fact a process owned by Stanford University and licensed to Google. Also that the exclusivity of that license is due to expire in a few years. This lack of ownership of the process coupled with the fact that it will soon be available to all is another reason I doubt it has changed since inception or that there will be changes to PageRank per se. Thier recent patents also address the issue of replacing PageRank as both LocalRank and Topic Specific Page Rank point to.

In fact, I believe that the several patents which Google have received in the past year or so may be a better indicator of Googles current thinking and recent events seem to indicate that they may be putting the duplicate content patent into some use a month or so ago.

It is IMO also prudent to remember that things like the duplicate content patent do not have to be incorporated into the general ranking algo, as it can be run as a seperate project outside the ranking algo and the duplicate content removed or penalized without disturbing the operation of the ranking algo, and indeed there is precedent for doing things that way if you remember the sporadic hidden text blitzes which occured a couple of years ago.

As you may have guessed, I am quite cautious about accepting Google public statements without a large grain of salt and a careful analysis of what it may or may not mean. There is no reason for Google to reveal any of their ranking algo to the public, and every reason not to disclose the secrets of its greatest asset.

[Mel]

Mel pretty much well answered the question by relating to PR10 sites. The first PR10 site did not come about linking to higher PR sites now then, did it?

Page relevancy that the link is listed upon is a must consideration when you simply apply commonsense IMO. If not, then PR would not exist because where would it start?

As I said earlier! Example: SEO sites, all PR8 homepage, 20 off, link to a web design page. Possiblities? Web design page takes up a PR9 because so many relevant content sites / pages, descriptive anchor text included, link to that page, hence how a higher PR develops to begin with. I have never seen PageRank as a pyramid structure.

Well you have me confused there Anthony, as there is no need to get links from high PR pages to get a high PR. You can get the same PR7 effect by linking to a PR 4 page from a PR8 page with no other links, or by linking to the PR4 page from several thousand PR4 pages. Thats the nature of the beast. Consider that fact that every sites PR is recalculated every month or so with no reference whatever to its previous PR.

Asking where PR comes from is answered by looking at the published PR equation, and there are no relvancy terms in there.

[AussieWebmaster]

I agree my comment was in jest...
Though the chicken and the egg approach does not quite get it either...
Content... pages and pages of relevant content and inbound links are a great start...
A site that has numerous areas covered and large number of pages with the respect of the various areas other top sites are a great start...

[Chris_D]

Brad wrote:
Besides, Google is selling Adwords to people that are selling (or brokering PR) so they help fuel the obsesion with PR and get paid for it.

So the first way that Google could make it harder to sell text links would be to stop selling the triggerword 'pagerank' in Adwords.

Pretty easy step.


Webmaster T wrote:
Where does it say Google doesn't like this? It's not mentioned in the guidelines or anywhere else I could see.

GG recently clarified that the guideline statement (below) pretty much applied to buying / selling links for their PR value.

"Don't participate in link schemes designed to increase your site's ranking or PageRank..... " http://www.google.com/webmasters/guidelines.html#quality

Notice it covers both increasing ranking AND increasing Pagerank?

Don't confuse PageRank with 'themed anchor text links' - they are 2 different things.

:)

[Mel]

So the first way that Google could make it harder to sell text links would be to stop selling the triggerword 'pagerank' in Adwords.

Pretty easy step.




GG recently clarified that the guideline statement (below) pretty much applied to buying / selling links for their PR value.

"Don't participate in link schemes designed to increase your site's ranking or PageRank..... " http://www.google.com/webmasters/guidelines.html#quality


Notice it covers both increasing ranking AND increasing Pagerank?

Don't confuse PageRank with 'themed anchor text links' - they are 2 different things.

:)

But like so many GG "clarifications" it still leaves a lot open to question:

For instance if I don't buy links to increase my PR, then its OK?
I can just buy them to increase my rankings and traffic, since I don't care about PR. :D

The problem with such a clarification is that it is based on intent, and there is no way to judge intent.

[Irony]

Today I saw an Adsense offering PR7, PR8, PR9 links for sale. Google accepts it. ;)

I wonder why.

[Mel]

Yep, there's the irony, Irony. Well Google has always said that there is no connection between the adsense results and the search results. :D

[Anthony Parsons]

I really don't think anything is put on Googles webmaster pages before it has had a full treatment by the resident public relations spindoctors, and therefor do not look for them to reveal anything about thier algos through that process, though they often offer tantalizing hints which can be taken in one of several ways, any of which may not be correct.

To that end Mel, I also don't think anything is put on stanfords pages that is going to give you the complete answer either. As for when you said:

This lack of ownership of the process coupled with the fact that it will soon be available to all is another reason I doubt it has changed since inception or that there will be changes to PageRank per se.

You can't honestly be serious Mel! If that algorithm hasn't changed, then how the hell did Google change how backlinks are counted, displayed, etc etc etc in the last couple of update. Please, get serious. If the PageRank algo is changed, then so is how Google interprets and counts backlinks. Gee....last time I looked that is exactly what happened last year, hence the reason so much importance is now placed upon backlinks to your site. Stanford papers from 4 years ago......technology can't remain the same for 4 months let alone 4 years mate. You could not honestly think the PageRank algorithm has not changed since its conception?

[seobook]

You can't honestly be serious Mel! If that algorithm hasn't changed, then how the hell did Google change how backlinks are counted, displayed, etc etc etc in the last couple of update. Please, get serious. If the PageRank algo is changed, then so is how Google interprets and counts backlinks. Gee....last time I looked that is exactly what happened last year, hence the reason so much importance is now placed upon backlinks to your site. Stanford papers from 4 years ago......technology can't remain the same for 4 months let alone 4 years mate. You could not honestly think the PageRank algorithm has not changed since its conception?

The PageRank algoright and Google's filtering mechanisms are two entirely different things there Anthony. The PageRank algorithm can surely shift around as they change the dampenening factor (originally d ~0.85) but the algorithm is just one part of what Google is using. There are other add ons which protect the usefulness of PageRank and filter out garbage links, but those are not part of PageRank.

PageRank is just this algorithm. Nothing more and nothing less...

PR(A) = (1-d) + d (PR(T1)/C(T1) + ... + PR(Tn)/C(Tn))

[Anthony Parsons]

Excellent point Aaron.....thank you. I think it is exactly as you state, the changes around and beyond our scope to what Google does, is a major impact on the way linking strategies have changed in recent months. If what Mel stated earlier about the facts in which Google is shifting, PageRank could be well and truly out the door and exceed its usefulness in coming years with regional search focus. That would be nice, as I personally can't stand PageRank myself. From reading its conception to everything else I have studied on it, it sucks in no uncertain terms.

It creates more trouble that what it is worth.

[Mel]

For even more interesting reading look at the details of the Localrank patent, which has the potential to remove all but one link from any Class C IP address to a page, and then to combine it with what they call Oldscore (which I interpret as the present ranking scores) in such a way that either the OldScore or the LocalRank score can have the greatest impact on the final rankings.

When I think about the potential implications to buying links, using internal anchor text links, forum links, ranking rings, etc it is clear to me that PageRank as originally conceived is horse and buggy technology by comparasion.

[Anthony Parsons]

Here's some answers for most of this direct from the London SES Conference and Mr Matt Cutts Mouth.

"thematic incoming links from authority sites carry more weight than on-page optimization."

Mr. Cutts emphasized that the quality of links is more important than quantity.

If you're not alternating your link text, especially if you're buying links, you may be harming your rankings. Identical link text, "looks like (and 9/10 times is) a sign of manipulation of its PageRank algorithm." Natural links, such as those appearing here in this article, use varying link text. Be sure to mix things up if you're linking for the express purpose of raising PageRank.

It was also suggested - by Google I assume - that you only buy links for the traffic they may provide. It's apparently easy for Google to tell when sites allow links for pay because the links go off to such unrelated sites. Not that this changes the fact that links to your site give your site better ranking. I guess if Google can tell they may work something into their algorithm, but link vendors will just spread out their links more, or create themed sites based around their clients.

Source WPN (http://www.webpronews.com/insiderreports/searchinsider/wpn-49-20040608SESLondonLinkBuildingAndDomainNameIssues.h tml)

Well, I guess google can track it.

[seobook]

Well, I guess google can track it.

or that is what they want people to expressly believe. their technology will rarely be as good as they want people to believe it is. their database is extremely huge and it is probably somewhat hard to make hard set rules that work when applied across the entire database.

some "spam" still works because there is legitimate reasons that use similar techniques

I still mix up link text though because it just makes sense

[Brad]

Of course it is also in Google's interest for Matt Cutt's to spread a little disinformation. ;) Google has one of the best spin departments in the business so I would take their ability to detect smartly done link placement with a grain of salt.

The flip side is when they are wrong, they distort their own results and shoot themselves in the foot, re.: Florida.

[seobook]

Blue smoke and mirrors
yup. nothing more, nothing less :)

[Mel]

I am amazed by the nuggets of wisdom that are dropped by Google reps in public:

"thematic incoming links from authority sites carry more weight than on-page optimization."

Now at first glance that appears tantalizing, but suppose instead he had said this:

incoming anchor text links from any old site carry more weight than on-page optimization

Wouldn't that also be true? Or are they one and the same thing?

[Webmaster T]

Agreed Mel, that seems to be the problem with that statement. How does Google determine "authority"? There seems to be lots of proof that a large quantity of themed text links can make a site an "authority" seemingly ignoring the relevance of the linking site to the linked site. IMO, that is what makes that statement confusing and why the truth in Matt Cutts statement is likely somewhere between "smoke and mirrors" and fact.

IMO, Matt's statement reflects their preference/goal the problem is they haven't been able to do it without having "authority" sites coming up for terms which they aren't "topical authorities". If Google determined "topical authority" would paid links from irrelevant "authorities" be worth buying? I doubt it that would make it a question of whether it is a "targeted audience" rather than a PR/ranking decision.

[cuzco]

Google employees do seem to talk in ideals rather than facts or are deliberately vague – who can blame them with people like us around ;)

There maybe some truth in what Matt Cutts said - that they can spot some of the worse offenders, but it doest look like they are doing anything about it yet.

IMO they wont be able to effectively filter the majority of artificially linked sites from natural sites. I think they may start doing something with the more extreme cases, such as sites that have 100’s of links from the same site and nowhere else. If they were only to lower the benefit of these links, instead of a complete block, it would make it tricky for us to spot.

[michaeljl]

As long as people can buy and sell links that transfer PR, then relevant returns for search engines that rely on link popularity will continue to get worse.

Case in point. When I did a Google search for the word "directory", among the top ten returns for this highly competitive word was a website for a Real Estate Company in Michigan.

What in the world does this Real Estate Company have to do with directories? Absolutely nothing! Not one bit of relevance at all.

How did they get there? Well it appears that this company has somehow acquired Magellan's old domain (http://www.mckinley.com) and all the links that used to be pointing to a search engine/directory are now pointing to this Real Estate Company.

Link Popularity (and Google) says that this website is a relavant important search engine/directory. (PR9) Nothing could be further from the truth.

[seobook]

As long as people can buy and sell links that transfer PR, then relevant returns for search engines that rely on link popularity will continue to get worse.

The web will continue to grow and competitive market forces will make link buying less profitable...as more people jump into the market the link prices will go up AND they will be less effective.

Those competitive market forces will dictate that the sites buying links will eventually need to become more relevant (similar to how Google AdWords is self regulating).

The one really large negative effect of link buying is that it makes the web increasingly commercial but not necissarily of degrading quality.

[nuclei]

Just go and have a close look at some of the most popular highly competitive terms. You will see for an example:

#1 has 1,000,000 links
#2 has 5,000 links
#3 has 1,500,00 links


I think you are very very confused by something here. It is not the links themselves that hold the power. It is the anchor text.

if site #3 has 1.5 mil links with its url in them and site #2 has 5,000 links with its keyword in them, of course #2 is going to beat #3.

BUT, Google does in fact see 1000 links from one site pointed to your url and does indeed count every single one of them. The site Mel is talking about was a site put in #1 for a competitive seo term. And it was done with links from mainly one site.

[nuclei]

Anchor text is one aspect, but the relevancy of that page the link is anchored upon is the key ingredient to PageRank.


Anthony, I highly suggest that you go read this in depth so that you can understand how PageRank works mate.

http://webworkshop.net/pagerank.html

It does not give a damn about relevancy. Nor does googles algo regarding where links come from and go to. In fact most "authority" sites have most of their links coming from highly non-relevant pages due to someone who owns a site liking their page and linking them in a blog, forum, or something similar.

Matt is a google employee, it is part of his job to tell us what google wants us to think. Always remember that.

All evidence tells us that anchor text links from any old sites that are pr4+ can help boost you in the serps. Keeping those links relevant does not help or hurt you algorithmically at all.

Down the road, it may, which is why most of us have gone toward only relevant links. But at present, there is nothing which suggests anything of the sort, other than google spin doctors.

[nuclei]

It is IMO well to remember that PageRank (while often considered a proprietary portion of the Google ranking algo) is in fact a process owned by Stanford University and licensed to Google. Also that the exclusivity of that license is due to expire in a few years.

In fact, I believe that the several patents which Google have received in the past year or so may be a better indicator of Googles current thinking and recent events seem to indicate that they may be putting the duplicate content patent into some use a month or so ago.


Thanks Mel, You have a much better reading way of putting what I was thinking than i would have done. I think a lot of people are ignorant of the fact that PageRank is not really G's, and that they think it is being manipulated daily by google to include some semblence of relevancy.

[Mel]

Yes there are many misconceptions regarding PageRank including the common ones like it is the most important part of Googles ranking alog, or that it somehow includes relevancy terms or that it counts some links but not others.....

If you review the Google section of almost any SEO forum, a great many of the threads are concerned with PageRank, when it is the simplest (and perhaps the least important) of Googles algos.

[steve sardell]

Yes there are many misconceptions regarding PageRank

Mel, as you know already, I absolutely agree with your statements of PR. The question becomes: what has bread all these misconceptions? Is it because it is a measurable factor? Is it because it is Larry's namesake? Or is it because of the confusion between page rank and PR. Might it be because of that green bar so prominately displayed on the tool bar? Or is it something altogether different or a combination? Never have I ever seen so much confusion and over complication on something so simple. That's my two cents worth.

[Mel]

IMO we have the Google Toolbar to thank for all the interest and resulting confusion over PageRank, plus a contribution from the name itself, which is easily confused by newcomers as somehow having to do with how pages rank.

The Google toolbar contributes because it provided the first (and at the time the only) public measurement of a pages worth (?), and thus became a source for bragging rights, even though it really only provided a relative indication of PageRank, and not of how well your pages would rank or even how important they were relative to other pages.

It is surprising to me that you will often see messages posted on SEO fora, together with congragulatory replies, from supposedly knowledgable individuals stating that this month they finally got their PR7, as though it were a medal for valor.

But IMO the toolbar serves a very useful purpose for Google in that it diverts so much time and effort away from research and optimization. Google fuel this fire with statements like

The heart of our software is PageRank™, a system for ranking web pages developed by our founders Larry Page and Sergey Brin at Stanford University. And while we have dozens of engineers working to improve every aspect of Google on a daily basis, PageRank continues to provide the basis for all of our web search tools.

But those who have to get results for clients soon learn that they should concentrate on other aspects of Google if they are to rank well.

I used to think that the first thing an aspiring SEO had to learn was that keywords were the basis of traffic, BUT I am gradually coming around to the opinion that the first lesson shoud be:

Take all public pronouncements by search engines with a large dose of salt, especially Google.

[nuclei]

I am gradually coming around to the opinion that the first lesson shoud be:

Take all public pronouncements by search engines with a large dose of salt, especially Google.

This is how it has always been. The search engine staffers are experts at telling you only what they want you to think. They have to really. Every large company does.

Misinformation is common.

[AussieWebmaster]

I know that is the most amusing part of it the Page in PR is actually named after Larry Page!!!

[massa]

It's not illegal. It can be profitable and if your site has value, you deserve to be compensated if a third party wants the benefit of that value. My argument has always been that it matters not WHY or HOW your site attained value. What matters is that it has it and if it has it, it belongs to the person who owns the website. If that value or some part of it, was attained by some kind of endorsement or award, it is still yours regardless of who gave it to you. If the awarder did not want you to benefit from it, they should not have awarded it.

My intent in posting is not to create more controversy or to change anyone's mind. It's been long enough now that I know there are some who think one way and some who think another and little I'm going to do or say is going to change that. I will say that there is a misconception that Searchking sold Page Rank when in fact, from the VERY begininng, it was made clear that we did not control PR. We did price our text ads based on the PR of that site, but that is not the same thing as selling page rank. The fact that it increased the page rank of the page we linked to had NOTHING to do with me. That was Goolge's doing and I clearly gave them the credit for that.

I believe the misconception is mostly perpetuated by people who are trading in links yet feel the need to find some distinction that what they are doing is good because it is different somehow from what I was doing. Well, whether you trade for links, spam guestbooks, sell them or buy them, there is value being traded and that is not so different.

Look folks, there is really no need for this debate. It's just business. If something has value and there is no law against trading in that market, it is going to be traded in. Period. That is the essence of free enterprise. Doing it and trying to find some way to convince yourself you are NOT doing it, only deludes you. Like it or not, you're doing it. And you have no choice but to do it. That is the world we live in and let us not forget it is a world Google itself, in no small part, created.

If you want a link, (and I believe we can all agree that you do), you can tell yourself and others all day long that you only want one that sends you direct traffic and that you don't care if it helps your site climb higher in the search engines results. If that were true, then you have no right to consider yourself an internet marketer. Especially not an SEO. So, OK, maybe we might secretely admit that part of why we want a link is because of the PR and the perception that it may help increase our own. Then THAT becomes at least part of the value and you should be prepared to pay a fair price for that value.

Also, let me remind you that Google is not the only reason you should want links. If google did remove the pr part of the toolbar, there would still be a market for text based links. Just ask Inktomi.

My point in making this post is to try to alter one little bit of perception. That saying it is one thing when it is in fact another is not a good thing. I believe that is a mistake far too many of us are making. Saying it is bought solely for the advertising value, yet you would be willing to pay more for advertising value on a site with a PR 9 than a site with a PR 4, is hypocritical at best.

I've been accused of being stupid because I came out and said it was about the PR. Maybe it was stupid and maybe it wasn't but what it was was honest. I based my prices in part on the value associated with the toolbar. In my opinion, the whole world is better off being stupid and honest as opposed to trying to find ways of doing one thing while using semantics to try and convince ourselves that is not what we are doing.

If someone wants to buy it, we should be able to sell it if it is not illegal. We should have no reason to try to find things to call it or ways to describe it to make it "look" like something else. If we have to try to lie about or try to disguise what we are doing because of a fear of someone, then there is something wrong.

I broker text links as a service to my clients. I take their orders and I try to find what they asked for. I don't try to make them call it something else. I don't try to convince them that what we buy and what we say we buy are different. I provide a simple, honest service to my clients and I'm proud of the service I provide. I believe acquiring links is nothing more than off-site optimization and I also believe that Google may come or go but off-site optimization IS the SEO of the foreseeable future.

I know there are those who say that is trying to make my business by using Google's. To me, this is a ridiculous argument. Does the Fiest phone book trade off of South Western Bell? Where would Goodyear tires be without Henry Ford. Where would SEO's be without search engines? For that matter, where would Google be without webmasters? Business is not about trading off of someone else's idea. That is what trademarks, patents and copyrights are for. Business is about identifying markets and offering products and services to satisfy demand at a profit.

I've broken one of my little rules not to get involved in these types of discussions with this post. Please excuse me. It's just that I have a great deal of respect for many of you here. I've done business with some, co-moderated with others and read posts and followed careers of just about everyone I've seen post here over the years. I felt that I may have a little different perspective to offer. I hope you all take my comments in the spirit in which they were intended.

[seobook]

Good post Bob :)

[nuclei]

And quite true in several aspects.

[orion]

Welcome Bob to the forum.

For some reason, Google is tolerating the practice of selling text links in connection with PageRank. The message is that

1. Bob and SearchKing were right in their perception; that connecting link offers with link demands is a valid service. Paid-results essentially is a variant of this. Using metrisized links in the equation is just a sound marketing strategy.
2. Many now view the whole issue as an offer-demand thing. The key question appears to be more of a power issue than anything else; ie., who should or will control revenues and cash flow? Everything else is a futile exercise for others to display their semantic abilities.

Now that marketers and others are showing their true colors, it is fair to ask: what SEOs/SEMs that were so critical in certain "seo experts" forums will say about the selling of services in connection with PageRank? Sure that some of them may come with carefully crafted words and pr lines to justify their change of heart.

I always had serious reserves about PageRank, not because I'm in favor or against Google's business model; their revenue model makes sense, their PageRank metric doesn't. This is not a contradictory position and does not invalidates my perception that PageRank is the biggest blunder that came from Stanford db, at least when applied to the commercial Web. My reserves are more for theoretical IR reasons than anything else. Under detailed inspection, one can see many obvious fallacies planted in the metric right from the get-go (98).

Glad to see that others can realize the obvious.

Orion

[nuclei]

welcome to the thread orion :)

[seobook]

sure pagerank has some problems, but it also gives webmasters a tie in with the google system and gives seos a semi tangible gauge to show they are doing something.

the fact that google gave a bunch of marketers who like to talk about search engines a tie in with their algorithm surely brings them more manipulation, but also brings them a bunch of buzz.

[AussieWebmaster]

There are two issues in this thread that I see and a couple of conversations have clouded the separation.
Inbound links using keywords helps improve a page for that keyword. The recent competition at SearchGuild well proved the point.
PR seems to help position in the SERPs but how exactly is still to be comprehensively defined.
The fact that sites with low PR can still capture a number 1 for keywords, seems to raise the question of unknown variables.
Google reflects its measurement in PR for certain elements... site authority, site popularity and a few other elements not as readily defined.
Given recip links are being down played it would seem, the argument of value of links is a very valid one. Google, IMO, should stop worrying about people getting value from their work (even if they are the measurer of the value)... link exchanges may be perceived as Google spam, and they are working to eliminate them, but buying a vote of popularity is done all the time offline, so online should not be seen so suspiciously.
The further Google goes to eradicate others making money from the system they set up will only take away from their own development.
The books that sell tips and tutorials for using AdWords seem okay. They even have a hard time cleaning the people who advertise 'black hat' methods from organic and paid results...
I think as the company goes public the demands to monetize more may make this thread moot.