The ability to pass pagerank. Can it be lost?

[kgun]

Listen to the WPW video

PubCon Las Vegas 2007: Matt Cutts of Google and Vanessa Fox » WebProNews Videos

At the start of that video Matt Cutts says in clear terms:

"Sites selling links can loose the ability pass pagerank".

Can is as a relative concept.

Natural questions:

1. Does it apply to the whole site?
2. Where is the limit for the whole site?
3. Where is the limit for each page?

Related video: PubCon Las Vegas 2007: Rand Fishkin of SEOmoz » WebProNews Videos

[mjtaylor]

Well, what's new? Google did cut the PR of quite a few major link selling sites this year - and those sites can no longer pass it ... because they don't have it.

The question, does it apply to the whole site? can be answered by looking at whether any of the pages of a given site have PR. If they do they can pass it ...

Limits? I think if Google saw one link sold it would remove the PR of the page where it's sold ... if the link selling is site wide ... then bye bye PR period ...

They *can* if they are *caught* ... or if they fit the pattern that has been worked into the algorithm to address the issue.

[wige]

My understanding of this method (based on comments from Google that seem to be intentionally non-specific) is as follows:

Google knows and acknowledges that although the SEO community stays on top of every tidbit of information coming from the Googleplex, the average webmaster/small business person is relatively oblivious to the hows and whys of rankings. As a result, most systems err on the side of caution - punish a site, not a server; punish a page, not a site; punish a link, not a page.

As the Googlebot reads a page, it breaks the content down, trying to figure out what the different parts of the page are. For instance, the bot will look for a footer so that can be ignored from the keyword calculations. Most sites have visual separations between different types of content - think of your own site, a visitor can immediately spot "this is navigation, this is the logo, these are ads, this is the footer". These distinctions are carried over into the code. Googlebot tries to find the same elements to determine how to rank different things. Wording that is found in the header, footer and navigation are discounted relative to the value of the content. In the same way, links that are separated from the content are discounted (or completely ignored) when the value of outbound links is calculated. In most cases, the bot will "ignore" sections of the page if it detects possible paid links. The idea that links within content are worth more because the surrounding text imparts more relevance has already been discussed, this idea simply expands on that concept.

Nofollow enters the equation because if the spider finds these links greatly outnumbering the editorial links, or in some other way these paid links trigger a red flag for the spider, it may decide that the purpose of the page is to spam the index. In that case the page could be penalized by having all pagerank removed, or even be dropped from the index. If this happens to enough pages on the site, then the site would risk a penalty. Finally, if enough sites on the server get penalized, the server itself may be penalized. Nofollow simply gives the webmaster a way to avoid the risk of too many false positives.

Bear in mind, paid direct links were created specifically for the purpose of gaming search engines. If you bought a link on a web site for any reason other than to increase your IBLs, they would not be direct links. If you sold links for any purpose other than to pass on page rank, you would not use direct links.

[kgun]

Originally Posted by mjtaylor

Well, what's new? Google did cut the PR of quite a few major link selling sites this year - and those sites can no longer pass it ... because they don't have it.

Of course I read and participated in the important thread:

The Ongoing PageRank Massacre Hits Big Players

There were also related threads.

Originally Posted by mjtaylor

The question, does it apply to the whole site? can be answered by looking at whether any of the pages of a given site have PR. If they do they can pass it ...

Limits? I think if Google saw one link sold it would remove the PR of the page where it's sold ... if the link selling is site wide ... then bye bye PR period ...

They *can* if they are *caught* ... or if they fit the pattern that has been worked into the algorithm to address the issue.

<quote>
Limits? I think if Google saw one link sold it would remove the PR of the page where it's sold ... if the link selling is site wide ... then bye bye PR period ...
</quote>
So the conclusion is.

If you wan't to sell links on your page:

1. Set up a professional link broker business. Did you watch the second video in my first post? Note what Rand Fishkin say's in the time interval 06.20 - 04.35. Build a marketplace for paid reviews. There has to be a solution to this problem to make it white hat. Related
   - SEOmoz | A Solution to the Paid Links Debate - Sponsored Editorials
   - SEOmoz | The Art of Buying Links Under the Radar
   - SEOmoz | Matt Cutts on Nofollow, Links-Per-Page and the Value of Directories
2. Hide the link selling as good as you can. An argument for unlegal and / or unethical business.
3. Is a page with Google Adsense treated the same as a page with affiliate links? Competition has to be fair.

[incrediblehelp]

kgun I am thinking the whole website, but that is just a guess.

[kgun]

Originally Posted by wige

My understanding of this method (based on comments from Google that seem to be intentionally non-specific) is as follows:

My bolding.

Interesting and you seem to agree with me that this important subject is far from settled.

Originally Posted by wige

Google knows and acknowledges that although the SEO community stays on top of every tidbit of information coming from the Googleplex, the average webmaster/small business person is relatively oblivious to the hows and whys of rankings. As a result, most systems err on the side of caution - punish a site, not a server; punish a page, not a site; punish a link, not a page.

My bolding again.

This is what I mean is still diffuse.

Originally Posted by wige

As the Googlebot reads a page, it breaks the content down, trying to figure out what the different parts of the page are. For instance, the bot will look for a footer so that can be ignored from the keyword calculations. Most sites have visual separations between different types of content - think of your own site, a visitor can immediately spot "this is navigation, this is the logo, these are ads, this is the footer". These distinctions are carried over into the code. Googlebot tries to find the same elements to determine how to rank different things. Wording that is found in the header, footer and navigation are discounted relative to the value of the content. In the same way, links that are separated from the content are discounted (or completely ignored) when the value of outbound links is calculated. In most cases, the bot will "ignore" sections of the page if it detects possible paid links. The idea that links within content are worth more because the surrounding text imparts more relevance has already been discussed, this idea simply expands on that concept.

Good explanation.

That is why separting the three layers of the web is so important:

1. Content in desrciptive (semantic) tags. XML is much better than HTML. XHTML is a bridge between HTML and XML, since an XHTML document has to be well-formed.
2. Presentation or styling in CSS / XSLT stylesheet files.
3. Behaviour in JS files.

There are additional layers for XML-driven sites, like link-bases (files), Document Type Declaration (grammar files), sematic links related to link-bases and roles of links (arch roles) etc.

Originally Posted by wige

Nofollow enters the equation because if the spider finds these links greatly outnumbering the editorial links, or in some other way these paid links trigger a red flag for the spider, it may decide that the purpose of the page is to spam the index. In that case the page could be penalized by having all pagerank removed, or even be dropped from the index. If this happens to enough pages on the site, then the site would risk a penalty. Finally, if enough sites on the server get penalized, the server itself may be penalized. Nofollow simply gives the webmaster a way to avoid the risk of too many false positives.

Do you mean that affiliate links, ideally, shall have a nofollow property? Some providers encrypt links etc. so it is not self evident that the link functions as it shall if it (some properties) is (are) changed. So there have to be a clear option to put no follow on links from the affiliate link provider or the merchant. What about the value of a link with and without a nofollow on it is the next natural question. May be not so important today, since a SeBOT will in most cases have no problem in identifying affiliate and Googel AdSense links. But as explained above there is an argument for unethical business to hide this.

Originally Posted by wige

Bear in mind, paid direct links were created specifically for the purpose of gaming search engines. If you bought a link on a web site for any reason other than to increase your IBLs, they would not be direct links. If you sold links for any purpose other than to pass on page rank, you would not use direct links.

I know that. But again, return to start, my first post. There are

1. One single link on a site.
2. One single or a few links per page (example FaceBook).
3. Sites with pages only designed for the purpose of linkselling, so it is separted from the rest of the site. A site with high traffic takes the opportunity to sell some of that traffic, regardless of the Se's policy.
4. Sites with AdSense and affiliate links all over. This Ad should be contextual and sematic aimed at the surfer.

May be some professional sites should have an option to turn off all Ad?

Do not turn this thread into an usability / accessibility discussion. Cross contry skiing is not relevant for every human being, as it is irrelevant for most people to go alone to the South Pole. Then he also needed equipment that only a few people need.

[kgun]

Originally Posted by incrediblehelp

kgun I am thinking the whole website, but that is just a guess.

When Jaan as I see as a SEO expert says this, the discussion and subject is

1. Far from settled.
2. There is no official guidelines / policy from the Se's.
3. Affiliate link providers and merchants have no policy on it.

[mjtaylor]

Originally Posted by kgun

1. Hide the link selling as good as you can. An argument for unlegal and / or unethical business.

Unethical? How so? Perhaps this question is asking for a thread of its own: Link Selling: Unethical or Just Unwise?

Cheers, MJ

[kgun]

Note the word hide.

Can it be compared to insider trading on the stock exchange? The laws on that are very strict in Norway and as far as I understand, no less strict in the USA.

The internet business is young, but rules will follow.

Internet law even younger than the business. IMO a growing and may be very profitable subbusiness of law generally (for younger lawyers).

Remeber the message and date of this post.

[mjtaylor]

I note your use of 'hide' and I acknowledge the need to hide link buying from Google if one wants to retain one's good ranking. But isn't having to hide from Google a pragmatical rather than an ethical issue?

[kgun]

Originally Posted by mjtaylor

But isn't having to hide from Google a pragmatical rather than an ethical issue?

No, I disagree, especially if

1. You hide to get a better ranking than other in the global search space.
2. It is not natural. When is it natural?
3. Other hidden goals / agendas that is used to increase your cash flow (at the price of other companies in the global search space).

But how can Google or another SE differentiate link A (paid) from link B (free) if the only information is the anchor tag and the anchor text? There is no ID, no div, no other attribute and no CSS class to differentiate the two. So I understand (and may accept) your pragmatic view, since that is (more or less) impossible. I wrote more or less in paranthesis, since advanced pattern reckognition algorithmes may be used to identify paid from free links with a probability > 1/2, that is tossing a coin.

[mjtaylor]

Originally Posted by kgun

No, I disagree, especially if

1. You hide to get a better ranking than other in the global search space.
2. It is not natural. When is it natural?
3. Other hidden goals / agendas that is used to increase your cash flow (at the price of other companies in the global search space).

What does being competitive have to do with "right or wrong"?

What does natural have to do with "right or wrong"?

[kgun]

Originally Posted by mjtaylor

Well, what's new? Google did cut the PR of quite a few major link selling sites this year - and those sites can no longer pass it ... because they don't have it.

The question, does it apply to the whole site? can be answered by looking at whether any of the pages of a given site have PR. If they do they can pass it ...

Limits? I think if Google saw one link sold it would remove the PR of the page where it's sold ... if the link selling is site wide ... then bye bye PR period ...

They *can* if they are *caught* ... or if they fit the pattern that has been worked into the algorithm to address the issue.

That was your first post.

Originally Posted by mjtaylor

What does being competitive have to do with "right or wrong"?

What does natural have to do with "right or wrong"?

When I page search "right" the first hit on this page is your above post.

What is the logic or intuition?

[Tubby]

I think when google cut the little green bar rank. it follows through the whole site.

but they seem to still send you the traffic anyway. (observations from my own sites only)

I have no idea, if the page rank down line, for off site links is effected.

[Webnauts]

May I join the discussion too?

Some statements of Google. Please read carefully:

Google works hard to ensure that it fully discounts links intended to manipulate search engine results, such excessive link exchanges and purchased links that pass PageRank. If you see a site that is buying or selling links that pass PageRank, let us know.We’ll use your information to improve our algorithmic detection of such links.

Source: Why should I report paid links to Google?

Your site's ranking in Google search results is partly based on analysis of those sites that link to you. The quantity, quality, and relevance of links count towards your rating. The sites that link to you can provide context about the subject matter of your site, and can indicate its quality and popularity. However, some webmasters engage in link exchange schemes and build partner pages exclusively for the sake of cross-linking, disregarding the quality of the links, the sources, and the long-term impact it will have on their sites. This is in violation of Google's webmaster guidelines and can negatively impact your site's ranking in search results.

Source: Link schemes

Many people who work on ranking at search engines think that selling links can lower the quality of links on the web. If you want to buy or sell a link purely for visitors or traffic and not for search engines, a simple method exists to do so (the nofollow attribute). Google’s stance on selling links is pretty clear and we’re pretty accurate at spotting them, both algorithmically and manually. Sites that sell links can lose their trust in search engines.

Source: Official Google Webmaster Central Blog: Information about buying and selling links that pass PageRank

Reading all the above, I do not see anything unfair from the part of Google. It is a fair and an ethical requirement to protect the webmasters from spamindexers or black hat SEOs, and the searchers from manipulated and irrelevant search results, and above all that to improve their search engine quality.

I have observe this topic for a while, and I have noticed that they hit the PageRank of the pages where bought or sold links have been found, but reading the bolded in red marked phrases, whole site can be negatively affected in terms of ranking results.

I am not a fan of anyone, except of myself but still I think Mr. Cutts wrote something which this time I fully agree with him: Selling links that pass PageRank

After reading the above, I do not think there is any space left to blame Google for anything. If you want to hide links, just do. But do it clever enough that you can pass the PR. If you do that, you must know that you are already wearing a BLACK HAT!!! You are not doing anything else than cloaking or whatever you want to call it. to manipulate the search results! I am against such dirty practices and I reported several sites already to Google. Who's next please?

JUST MY LAST FEW DOLLARS...

[kgun]

Yes, and this thread was hijacked, since the thread in the SEO forum was reduced to a discussion about Google.

This is the Google sub forum.

But who cares?

Question:
SEO experts or Google experts?

[mjtaylor]

I did not mean to hijack this thread, which is why I started the ethics question as a separate thread ... my apologies if I have done something untoward.

[kgun]

No problem.

[kgun]

This short post is inspired by an authoritative book "Microeconomic analysis" chapter 24, third edition by Hal Varian ISBN 0-393-96026-9

Economists have tools that dates back to Pigou (1920) to answer these questions. The subject is "externalties" in production and consumption.

When the actions of one agent directly (read sold or bought links) affect the environment of another agent (read Google's PageRank algorithme or competitors that do not buy links), we call this an externalty. In the case of an externalty in production, the production of one company (Google's PageRank) is affected by the actions or production by another company (e.g. sold links).

In general, market equilibria will be inefficient in the presence of externalties. This naturally leads to an examination of various measures to restore an efficient market equilibrium, like disallowing pagerank to be passed on sold links.

[kgun]

Two updated links that may be of interest to you:

How to implement 'rel=nofollow' - google-blog.dirson.com (E.g. how it is done on phpBB).

Directories To Avoid that Don't Pass PR (Page Rank)

[AVC]

Well, we all know this is the case today for sure, this is an old thread so the facts are established.

Matt Cutts said many times that if a webmaster sold links without no follow tags that Google would remove the power to pass PR from those pages.

[kgun]

Originally Posted by AVC

Matt Cutts said many times that if a webmaster sold links without no follow tags that Google would remove the power to pass PR from those pages.

How easy is it to identify that?

Read post 37 and 38 too.

[AVC]

It is real easy, if Google identifies you as a link seller in a contextual link brokers network or if you running interlinked directory link farms, all Google has to do is get a spam report on a page a few times or if they have a directory tagged as a link farm they simply de-index it and give it a PR 0

[kgun]

No, I don't agree. It is not easy if it is done behind the scenes and (partly) randomized. Reread post 35 - 38 in the above mentioned thread if you do not take my point.

In my opinion it is an agenda for black hat SEO.

I have seen it time again and again. This only hurt serious webmasters while the black hatters make money.

Exercise:
There are not millions, but billions of billons of links out there. How should

1. GoogleBOT identify links pulled from a database with (random) rel="nofollow" on them?
2. How could that be manually handled?

Before they are taken, they may have earned enough to set up 100 new sites in different geographical areas. How maniy identities can you have in the faceless cyberspace?

[AVC]

It is easy for Google due to their sophisticated link mapping software technology that shows them all reciprocal, triangular and interlinking relationships, they know more about linking than webmasters and black hat scam artists combined.

Google builds profiles on webmasters, they know all the sites you control, own or are associated with also, so if you are involved with link farms, link selling or buying beware.

[kgun]

Originally Posted by AVC

It is easy for Google due to their sophisticated link mapping software technology that shows them all reciprocal, triangular and interlinking relationships, they know more about linking than webmasters and black hat scam artists combined.

But note that I use the word random / randomize. There are no pattern in randomness aside from the inability to identify any pattern. An other method is to use own bots to identify your own linking patterns and use a link pattern database in your own behind the scene paid hidden link agenda.

Don't underestimate this potential.

My final remark. A search engine do not own the webs link structure, link base and related information. The search engine own their alogorithms. As an economist, a search engine should be very careful with

• Methods that favours scammers.
• Unfair competition.

So like the discussion in the long WPW thread, can inbound links really hurt you

Originally Posted by kgun

And Google should be very careful with reported links:

1. They think are bought (sold by other companies). If they are 95 % (some scientists require 99 % confidence level) or more sure they should only give no pagerank is my personal view.
2. Irrelevant links. See my example above with fish and tomatoes.
3. Suspect links.
4. Scam or spam links.

the same point could be made here regarding "The ability to pass pagerank. Can it be lost?"

Unless they use strict confindence levels in their algorithmes they may end up penalizing the wrong sites.