Does PR Sculpting Work?

[janeth]

Originally Posted by activeco

But what surprised me the most is the clear lack of understanding of basics of PR distribution, among other things.
While less knowledge about some other important internet mechanisms (such as networking protocols, in the first place) can be understood, having in mind that most SEO's are actually mainly (web)designers slightly traversing into the underground world of networks, it is a bit shocking that people specialized in pleasing search engines actually do not know the most fundamental facts of Google's PR calculation.

I was also surprised by this.

[Peter (IMC)]

And, you don't get to keep and count the PR that you have to pass to others; i.e., you don't get to vote for yourself.

You do get to vote for yourself, but not through links themselves. More pages in your site will add PR to your site. (but 10 pages more isn't going to make much diference. This difference is only noticable when comparing sites with few pages to sites with tens of thousands of pages.)

PR sculpting is possible with nofollow, (but it's a waste of time as far as I'm concerned.) I agree with the way Janeth describes how PR works. Every page has a single PR point (not toolbar PR point but real PR point) to start with (before calculating the distribution of PR) and thus the total amount of PR available in Google's index is at most the number of pages they have indexed. Then after an iterative process the distribution of PR through links is determined.

So that answers the question: "But, how is the initial value created?"


There is no "proof" needed to determine wether or not PR sculpting with the nofollow works or not. It does work because if you deny a link to pass PR then the distribution will be different. The more important question is whether or not this has any significant effect on positions and traffic to your website. In my opinion it has very little effect simply because PR is just one of the many ranking factors and what's more important, the amount of PR you can channel away from certain pages to other pages is so small that you´re talking about a change in real PR of less than a %.

In case you have a really big site of tens of thousands of pages (which is actually what Greg Boser is talking about, Webnauts) then you can significantly sculpt PR in the site. And I can imagine that in those specific cases it can have a positive effect on positions for some keywords. But I have to say that I suspect that there is also a negative effect on the number of keywords the site is found for which in turn means less traffic to the site. Which brings me back to my original point that using nofollow to sculpt PR is a waste of time.

[activeco]

Originally Posted by Peter (IMC)

In case you have a really big site of tens of thousands of pages (which is actually what Greg Boser is talking about, Webnauts) then you can significantly sculpt PR in the site. And I can imagine that in those specific cases it can have a positive effect on positions for some keywords. But I have to say that I suspect that there is also a negative effect on the number of keywords the site is found for which in turn means less traffic to the site.

What do you mean by that?
I don't see any relationship with nofollow and number of keywords on the page.

[janeth]

Originally Posted by activeco

What do you mean by that?
I don't see any relationship with nofollow and number of keywords on the page.

He's not talking about the number of keywords on a page. He's talking about a site with over 10,000 pages and that sculpting PR would have an affect on getting your site to rank. If it was that big.

But then went on to say that your efforts would be better spent on something else.

[activeco]

Originally Posted by janeth

He's not talking about the number of keywords on a page. He's talking about a site with over 10,000 pages and that sculpting PR would have an affect on getting your site to rank. If it was that big.

Well OK, he's talking about number of keywords sitewide.

But I have to say that I suspect that there is also a negative effect on the number of keywords the site is found for which in turn means less traffic to the site.

I see the logic, but I can't agree.
The whole idea is to transfer the traffic to the more important pages.
It doesn't matter if you lose 1,000 visitors coming to your 'privacy' page with a keyword "IP address", when you get another 1,000 going to your shopping cart page with a keyword "buy <product> now".

[janeth]

Originally Posted by activeco

I see the logic, but I can't agree.
The whole idea is to transfer the traffic to the more important pages.
It doesn't matter if you lose 1,000 visitors coming to your 'privacy' page with a keyword "IP address", when you get another 1,000 going to your shopping cart page with a keyword "buy <product> now".

Not traffic. Your site visitor can go anywhere they want on the site. What we would be doing is controlling where Google would put the internal ranking into the site.

Another words if I have a web design site and fix the site so all my internal power is moved to one page that talks about custom website design. Then I rank #1 for custom website design but loose all the ranking for the rest of the site, was it worth loosing all my ranking power to the other pages?

And what will the affect of all those pages linking back to my home page be now that I am not longer linking to them?

Maybe my home page no longer ranks for web designers, which is more important than custom website designs.

And now that Google is no longer visiting my privacy page or contact me page is it possible that they could no longer she me as a real business and I could drop in rankings because of this?

Not saying not to use it but would have to wonder if it was worth it in the long run.

[activeco]

Originally Posted by janeth

Not traffic. Your site visitor can go anywhere they want on the site. What we would be doing is controlling where Google would put the internal ranking into the site.

Not exactly. In most cases your internal links have the same (or arguably slightly less) weight as the external links. In case of lot of internal links one can be able to spread them in the best possible way and indeed help the PR of the important pages.

Another words if I have a web design site and fix the site so all my internal power is moved to one page that talks about custom website design. Then I rank #1 for custom website design but loose all the ranking for the rest of the site, was it worth loosing all my ranking power to the other pages?

True, so you won't do that obviously.

And what will the affect of all those pages linking back to my home page be now that I am not longer linking to them?

Why would you do that? You have to link properly ALL the pages.
The key is how many times and from what places.

And now that Google is no longer visiting my privacy page or contact me page is it possible that they could no longer she me as a real business and I could drop in rankings because of this?

See above.

[janeth]

You could get the same affect out of building a nice looking site with flash navigation or java and then linking to your main pages using text links inside your copy.

It would be a lot of work and would take some thanking but the result would be better in the long run. Each link would be surrounded by text. Then you could get links to all the internal pages from outside sites.

You could even go to free hosting sites build a site about a certain topic, submit that site to the directories and then link the new site you just built back to the internal pages on your main site.

Every internal page would have links coming from outside sites and would be linking to your main pages through the text links on those pages.

[activeco]

Originally Posted by janeth

You could get the same affect out of building a nice looking site with flash navigation or java and then linking to your main pages using text links inside your copy.

Sure, there are many ways for the proper internal link design.
But, what is good with rel=nofollow (among other things) is that you can use it on old, existing sites and very easy earn some points.

[janeth]

Originally Posted by activeco

Sure, there are many ways for the proper internal link design.
But, what is good with rel=nofollow (among other things) is that you can use it on old, existing sites and very easy earn some points.

Yes and I guess it would be easy to change back if you decided not to use it.

[Peter (IMC)]

Originally Posted by activeco

What do you mean by that?
I don't see any relationship with nofollow and number of keywords on the page.

I wasn't talking about the number of keywords on the page. SEO has the effect that the number of keyword phrases a site is found for, increases. There are so many variations on a basic keyword phrase and longer keyword phrases that they are giving a site many more visitors than those few main keywords you put in your title and other places.

When you nofollow a bunch of pages and also noindex them as described by Greg Boser then you´re going to lose a whole bunch of reasons to get positions for those types of keyword phrases. I can imagine that this has a negative effect on the total number of visitors.

Generaly when we optimize a site the number of keywords the site is found for goes up 10 times. (like being found for 400 different keyword phrases per month before optimization to like 4000 different keyword phrases after optimization.) The nofollow (the way Greg Boser suggests to do in big sites) may have as a result that the total number of visitors goes down.

The idea that the more important pages have to get more visitors by sculpting PR using nofollow doesn't make that much sense to me. You can argue that a contact page is less important, but it won't be found that often anyway.

It doesn't matter if you lose 1,000 visitors coming to your 'privacy' page with a keyword "IP address", when you get another 1,000 going to your shopping cart page with a keyword "buy <product> now".

Here you are assuming that reducing the PR of that privacy page will make those positions for "ip address" dissapear and that by the added PR to those product pages they will gain positions. In my oppinion that's a wrong assumption. Google doesn't work in that way. What's needed is a good internal link structure to get those pages to rank higher. Anchor texts of internal links for example have a much greater effect so it's better to focus on that than to focus on nofollowing links with hardly any effect.

Again, in huge sites there may be some effect, but then you'd be targeting higher level category pages and reduce the PR of the deeper content pages (That's Greg Boser's suggestion again). And in most comercial sites that would mean reducing the PR of your product pages. Doesn't seem that smart to do. Perhaps sites like article websites could benefit from something like this but for most sites, PR sculpting is a big waste of time.

[activeco]

Originally Posted by Peter (IMC)

When you nofollow a bunch of pages and also noindex them as described by Greg Boser then you´re going to lose a whole bunch of reasons to get positions for those types of keyword phrases. I can imagine that this has a negative effect on the total number of visitors.

Of course. Suggesting noidex-ing white-hat regular pages is at least irresponsible. I don't advocate that.

The idea that the more important pages have to get more visitors by sculpting PR using nofollow doesn't make that much sense to me. You can argue that a contact page is less important, but it won't be found that often anyway.
...
Here you are assuming that reducing the PR of that privacy page will make those positions for "ip address" dissapear and that by the added PR to those product pages they will gain positions. In my oppinion that's a wrong assumption.

Added PR does gain positions. It's hard to find anyone today disagreeing with that.
The simple fact (again) is that the distribution of the PR from a particular page depends on # of receiving links so instead of: "PR juice per link = Transfering PR / X links" you have "PR juice per link = Transfering PR / (X-#nofollow's)".
The PR gain can be enough/substantial to boost important pages in serps and produce more traffic to those pages.
The traffic to insignificant pages is reduced but they are still there, less linked, indexed, having the same content.

[incrediblehelp]

Activeco have a gander:

Why your nofollow testing sucks - SEO Theory - SEO Theory and Analysis Blog

[janeth]

Originally Posted by incrediblehelp

Activeco have a gander:

Why your nofollow testing sucks - SEO Theory - SEO Theory and Analysis Blog

A really long article to say he does not believe it was properly tested.

[crankydave]

Originally Posted by janeth

A really long article to say he does not believe it was properly tested.

And that it is basically impossible to properly test because external influences cannot be eliminated. So everyone who says it works can't prove it and everyone who says it doesn't work can't prove it because neither can conduct a "test" in a "sterile" environment.

Dave

[activeco]

Originally Posted by janeth

A really long article to say he does not believe it was properly tested.

Very true, neat and short statement. Exactly what one misses on his blog. I think his articles were written mainly for search engines.
The guy is backpedalling in a way from his previous post where he insulted everyone not agreeing with his beliefs.
The arguments were ranging from 'stupid search engine optimizers and total morons' to the more sophisticated ones such as 'incompetent asses and idiots'.

This time his article was more moderate with the simple heading "Why your nofollow testing sucks", where he took a long journey to say that he still didn't believe it.

Interestingly, in another article he states the following: "A good Web site only needs its own internal linkage to help most pages rank for something useful. A bad Web site relies on external links."

Let's play fair here; I understand it was in another context. But he should be aware that the statement means more than he originally thought.

Anyway I don't understand why so much dust around this question.
There is nothing new here, good internal linking was always a kind of "artistic knowledge", with or without rel='nofollow'.
You can dissipate your PR and link extensively all around or you can link smart.
It's your choice.

[janeth]

Originally Posted by activeco

Anyway I don't understand why so much dust around this question.
There is nothing new here, good internal linking was always a kind of "artistic knowledge", with or without rel='nofollow'.
You can dissipate your PR and link extensively all around or you can link smart.
It's your choice.

It is an old trick with a new name and a new tool. Guess there has not been a whole lot to talk about in the world of SEO so they decided to make a big deal about an old trick.

[incrediblehelp]

Originally Posted by crankydave

And that it is basically impossible to properly test because external influences cannot be eliminated. So everyone who says it works can't prove it and everyone who says it doesn't work can't prove it because neither can conduct a "test" in a "sterile" environment.

I guess this is the best way to look at it.

[deepsand]

Originally Posted by Peter (IMC)

You do get to vote for yourself, but not through links themselves. More pages in your site will add PR to your site. (but 10 pages more isn't going to make much diference. This difference is only noticable when comparing sites with few pages to sites with tens of thousands of pages.)

But, that's irrelevant to the issue under discussion.

Originally Posted by Peter (IMC)

PR sculpting is possible with nofollow, (but it's a waste of time as far as I'm concerned.) I agree with the way Janeth describes how PR works. Every page has a single PR point (not toolbar PR point but real PR point) to start with (before calculating the distribution of PR) and thus the total amount of PR available in Google's index is at most the number of pages they have indexed. Then after an iterative process the distribution of PR through links is determined.

So that answers the question: "But, how is the initial value created?"

In the method as originally developed at Standford University by Larry Page, the intial value value is 1, with the 1st approximation being 1 divided by the total number of pages being evaluated; the damping factor is variable. Furthermore, there is the implicit assumption that each page stands alone, i.e. that it is not related to any other page by virtue of, for example, being part of a set of pages all belonging to the same domain.

In the variation implemented by Google, the damping factor is commonly assumed to be 0.85; what it actually is we cannot say. More importantly, though, we can say that the assumption re. each page being wholly unrelated to, i.e. independent of, all other pages is not present. The very fact that Google saw a need for the "nofollow" attribute serves as proof of such.

Lastly, when Google first evaluates a given site, or later re-evaluates it, it does not reset all intial values to 1, or some proportionate value, and re-evaluate all pages. In this regard, Google's implementation varies significantly from the theoretical model presented in the origianl paper. Therefore, the question re. initial values, in Google's implementation, remains unanswered.

If you're interested in studying the mathematics here involved, you may find PageRank - Wikipedia, the free encyclopedia useful.

Originally Posted by Peter (IMC)

There is no "proof" needed to determine wether or not PR sculpting with the nofollow works or not. It does work because if you deny a link to pass PR then the distribution will be different. The more important question is whether or not this has any significant effect on positions and traffic to your website. In my opinion it has very little effect simply because PR is just one of the many ranking factors and what's more important, the amount of PR you can channel away from certain pages to other pages is so small that you´re talking about a change in real PR of less than a %.

In case you have a really big site of tens of thousands of pages (which is actually what Greg Boser is talking about, Webnauts) then you can significantly sculpt PR in the site. And I can imagine that in those specific cases it can have a positive effect on positions for some keywords. But I have to say that I suspect that there is also a negative effect on the number of keywords the site is found for which in turn means less traffic to the site. Which brings me back to my original point that using nofollow to sculpt PR is a waste of time.

Since the issue at hand deals specifically with intra-site links, which are not addressed by the original published work, as above noted, and Google has not told us hold they deal with such, we can but speculate and offer opinions. Therefore, to say that no proof is required, simply because one believes that "sculpting" works will not serve to settle the issue.

Lastly, and most importantly, precisely how Google aggregates the PR of the individual pages within a site when determining SERP remains a closely guarded secret, so that we cannot with any degree of certainty say what the results of tinkering with the "nofollow" attribute of intra-site links might be.

[deepsand]

Originally Posted by crankydave

And that it is basically impossible to properly test because external influences cannot be eliminated. So everyone who says it works can't prove it and everyone who says it doesn't work can't prove it because neither can conduct a "test" in a "sterile" environment.

Dave

Precisely what I've been saying all along.

We don't know, and, more importantly, we can't find out on our own.

[activeco]

Originally Posted by deepsand

We don't know, and, more importantly, we can't find out on our own.

You just destroyed complete SEO business.

[deepsand]

Originally Posted by activeco

You just destroyed complete SEO business.

Oops.

Good thing I didn't give up my day job!

[crankydave]

Originally Posted by deepsand

Precisely what I've been saying all along.

We don't know, and, more importantly, we can't find out on our own.

Just to be clear, you quoted my impression of the linked blog post and is not neccessarily my opinion.

Dave

[deepsand]

Originally Posted by crankydave

Just to be clear, you quoted my impression of the linked blog post and is not neccessarily my opinion.

Dave

Plausible deniability?

Seriously, though, can you see any way to construct a test such that all variables but the one to here be tested can be controlled or their effects nulled?

Theoretically, one would construct 2 completely identical sites, right down to the date of domain registration, length of term of registration, owner, host, date(s) of being indexed, and any and all other externalities related to the these 2 sites.

By doing the above the effects of any others sites should be equally felt by these 2 sites, thus effectively nulling such.

One site would be the control. Once having confirmed that both are ranked equally by all of the SEs of interest, one would then tinkering with the intra-site links of the test site, one at a time, observing any changes in the SE's results, followed by 2-link alterations, etc..

The problem as I see it is how to meet the initial conditions re. externalities relating to the control and test sites, not to mention the potential problem of "duplicate content." Merely ensuring that both sites get indexed and re-indexed at the same time, and by the same data center, seems to be a show stopper.

[janeth]

I go fishing and I put a worm on my hook throw it into the water and catch a big cat fish and repeat the process and catch another and continue to do this for a couple hours. Then someone walks up and says, "are there catfish in the pond" and I say, "yes".

Then they say, "can you prove it".

I say, "yes look at all I've caught".

They say, "yes but can you prove there are more".

I throw my hook in and catch another.

But they explain that they don't want to waste their time and would like for me to prove there are cat fish in the pond before they go fishing. I could have just caught the last one.

We continue to do this for the next couple hours. Me catching fish and them asking me to prove there are fish in the pond.

At the end of the day I've caught enough fish to pay my bills and buy a new car. They happily tell me they are glad they didn't give up their day job because I was unable to prove there were fish in the pond. Then they head to the store to by my fish.

[Peter (IMC)]

One site would be the control. Once having confirmed that both are ranked equally by all of the SEs of interest

Interesting. So they rank 2 sites in the same position.....

So everyone who says it works can't prove it and everyone who says it doesn't work can't prove it because neither can conduct a "test" in a "sterile" environment.

This logic is way too easy and applies more to God than anything else. It basically means that it works for those who believe it works and it doesn't work for those who don't believe it works. When it comes down to what you believe it is pretty much always true because you will see what you´re looking for anyway.

Added PR does gain positions. It's hard to find anyone today disagreeing with that.

I'm one that doesn't agree with that the way that you wrote it down. PR is only as useful as the link text is related to the target page. So simply channeling PR to other pages regardless of link text is not going to help much.


If you want to do a test on if PR sculpting works do the following test:

• Create a 100 page site and make sure there are exactly 99 links in every page. (make sure to use not existing words so that you don't have any interference from words in other sites)
• Each link links to each of the other 99 pages.
• Make sure you've got backlinks from other sites to only one of the pages (we will call this page the home page) in order to get at least enough PR in the site.
• Try to get that page to to show up as a PR 2.
• Expected result: You will find that all the other pages will have the exact same toolbar PR (most likely PR1).
• Now in each page, except for the home page, nofollow all the links except for the link to the homepage.
• Wait until toolbar PR changes.
• Expected Result: Your internal pages will drop in PR and your home page may go up a point.

You can test if it works, but it will take you probably about a year.

Testing these things is possible but it's not the right way to go about it. Using common sense is much better and faster. For example, we know that the distance between the earth and the sun is about 149476000 kilometers (92900000 miles), but nobody ever measured it directly by measuring it with a ruler. (which wouldn't even be of much use anyway I guess considering that gravity has a direct effect on the length of the ruler, depending on from where you observe the ruler. )


To get back to the original question of the thread:

Re: Does PR Sculpting Work?

I guess you have to be more specific. It seems that the question is too open to be answered. For me it was about if it is possible to get PR to go up in some pages by using the nofollow. But you could also have meant if it works for getting more visitors to your site. The first question I would answer with: "Yes!" The second I would answer with: "Better not waste your time on using nofollow to sculpt PR and use that time to use techniques that get you much better results."

[Peter (IMC)]

At the end of the day I've caught enough fish to pay my bills and buy a new car. They happily tell me they are glad they didn't give up their day job because I was unable to prove there were fish in the pond. Then they head to the store to by my fish.

lol, Now that's typical "Janeth" logic...

[incrediblehelp]

Originally Posted by Peter (IMC)

To get back to the original question of the thread:

I guess you have to be more specific. It seems that the question is too open to be answered. For me it was about if it is possible to get PR to go up in some pages by using the nofollow. But you could also have meant if it works for getting more visitors to your site. The first question I would answer with: "Yes!" The second I would answer with: "Better not waste your time on using nofollow to sculpt PR and use that time to use techniques that get you much better results."

OK. Does PR Sculpting internal navigational links help internal pages rank and get more traffic?

[activeco]

Originally Posted by Peter (IMC)

I'm one that doesn't agree with that the way that you wrote it down. PR is only as useful as the link text is related to the target page. So simply channeling PR to other pages regardless of link text is not going to help much.

If the PR is transfered, it is transfered to other pages regardless of anything else. Once a page receives the PR, it is used as a ranking factor for all of its content, not just particular keywords.
There could be filters and penalties but that is beyond this thread.

If you want to do a test on if PR sculpting works do the following test:
...

Testing these things is possible but it's not the right way to go about it. Using common sense is much better and faster.

Uhm...agreed.

For me it was about if it is possible to get PR to go up in some pages by using the nofollow.
...
The first question I would answer with: "Yes!"

OK, so according to the common sense you can get higher PR on some pages. Agreed.

But you could also have meant if it works for getting more visitors to your site. The second I would answer with: "Better not waste your time on using nofollow to sculpt PR and use that time to use techniques that get you much better results."

But you don't agree that more PR means more visitors, right?

[janeth]

Originally Posted by incrediblehelp

OK. Does PR Sculpting internal navigational links help internal pages rank and get more traffic?

If you had a site with over 10,000 pages and a year to play with it. I think so.

[Webnauts]

Originally Posted by Peter (IMC)

The second I would answer with: "Better not waste your time on using nofollow to sculpt PR and use that time to use techniques that get you much better results."

I fully agree. But who am I=

Adam Lasnik confirms that you are fully right too: Video: Interview with Adam Lasnik of Google | Online Marketing Blog

[Peter (IMC)]

But you don't agree that more PR means more visitors, right?

Not the way you put it there. You make it look like that more PR per definition means more visitors and that is not correct. PR is not a stand alone factor. It works together with other factors and I believe Google is using some form of a Topic Sensitive PageRank (TSPR) algorithm. Which basically means that the amount of PR used to determine positions depends on the keyword search for (for example, PR used to rank your page about horses may be 0 when the query is cows.)

I'm not sure if they do it exactly like that or if they use some sideway logic to determine link relevance seperately and not directly in PR it self. But it basically comes down to the same thing.

So just blindly transfering PR to other pages is not as useful as it may seem.

[deepsand]

Originally Posted by janeth

I go fishing and I put a worm on my hook throw it into the water and catch a big cat fish and repeat the process and catch another and continue to do this for a couple hours. Then someone walks up and says, "are there catfish in the pond" and I say, "yes".

Then they say, "can you prove it".

I say, "yes look at all I've caught".

They say, "yes but can you prove there are more".

I throw my hook in and catch another.

But they explain that they don't want to waste their time and would like for me to prove there are cat fish in the pond before they go fishing. I could have just caught the last one.

We continue to do this for the next couple hours. Me catching fish and them asking me to prove there are fish in the pond.

At the end of the day I've caught enough fish to pay my bills and buy a new car. They happily tell me they are glad they didn't give up their day job because I was unable to prove there were fish in the pond. Then they head to the store to by my fish.

But, with fishing you can prove that the rod is attached to the line, which in turn is attached to the hook, which is attached to the bait, and the the fish is ultimately attached to the hook. Thus, you have an empirically demonstrable unbroken link between cause and effect.

However, with "nofollow" and SERP, you cannot observe anything between your conjectured cause and any effect.

[activeco]

Originally Posted by Peter (IMC)

You make it look like that more PR per definition means more visitors and that is not correct. PR is not a stand alone factor.

I see.
Actually my mistake was to bring the "more PR = more traffic" statement as I thought of it as a general consensus. It was indeed not the question about the benefits of the PR raising on some pages, but simply 'was it possible to do using "nofollow"'.
The thread regarding your definition of PR could be interesting and I would suggest you to start the one.

[deepsand]

Originally Posted by activeco

I see.
Actually my mistake was to bring the "more PR = more traffic" statement as I thought of it as a general consensus. It was indeed not the question about the benefits of the PR raising on some pages, but simply 'was it possible to do using "nofollow"'.
The thread regarding your definition of PR could be interesting and I would suggest you to start the one.

Re. the definition of Page Rank, you might find PageRank - Wikipedia, the free encyclopedia useful.

Note, however, that such does not contain details re. Google's implementation of such. Of particular interest is the question of what damping values Google uses for intra-site links as opposed to inter-site ones.

[activeco]

Originally Posted by deepsand

Re. the definition of Page Rank, you might find PageRank - Wikipedia, the free encyclopedia useful.

OK, I am dealing with the Google PR for at least seven years. Several own developed scripts based on it gave me more insight into a little things normally you don't think of.
I am pretty sure that the Google's PR distribution from one page does not pass (anymore) the same weight to all the outbound links equally, as well as the possible discount for internal linking, but that's not the topic either.
Anyway could you tell me what you don't agree with (if so), regarding my PR understanding and what part of that Wikipedia page I should pay a special attention to?

[activeco]

As a bonus let me provide the public here with the related quote straight from the official horse's mouth.
It was Cutts' reply to Martinez' article "SEO Myth: You can control the flow of PageRank on your site".

Originally Posted by Matt Cutts

Hmm. Michael says "SEO Myth: You can control the flow of PageRank on your site" and then later "You cannot control the flow of PageRank on your Web site but you can control your own internal link flow. Link flow is not PageRank."

To me, the second statement (choosing how to link within your site) clearly does control the flow of PageRank on your site.

I agree with much of Michael's ideas, but his habit of flatly asserting things can undermine his arguments.

[deepsand]

Originally Posted by activeco

I am pretty sure that the Google's PR distribution from one page does not pass (anymore) the same weight to all the outbound links equally,

Speaking from my background in IT, were I implementing any version of the Page Rank algorithm, I would surely tailor it to the specific application. In the case of using it, or any similar algorithm, in an SE, I would indeed seek to adjust the weight assigned to any outbound link(s) based on the characteristics of the target and, to the extent possible, dynamically based on the apparent intent of the searcher.

Originally Posted by activeco

as well as the possible discount for internal linking, but that's not the topic either.

Also, I would most definitely treat intra-site links differently that inter-site links. A simple way, though not necessarily the best one, would be to use different damping factors for each.

And, since Google may in fact have done such, which I suspect given the circumspect manner in which M. Cutts has addressed the issue, this particular point is very much on-topic.

Originally Posted by activeco

Anyway could you tell me what you don't agree with (if so), regarding my PR understanding and what part of that Wikipedia page I should pay a special attention to?

I was not taking issue with anything that you stated; rather, having noticed your mention of discussing a definition of Page Rank, and not knowing if you'd read the original paper or any abstracts of it, I offered the wiki link as a starting point. If, after a thorough review of your posts, I find anything worthy of disagreeable mention, I'll address such then.

[deepsand]

Originally Posted by activeco

As a bonus let me provide the public here with the related quote straight from the official horse's mouth.
It was Cutts' reply to Martinez' article "SEO Myth: You can control the flow of PageRank on your site".


Originally Posted by Matt Cutts
Hmm. Michael says "SEO Myth: You can control the flow of PageRank on your site" and then later "You cannot control the flow of PageRank on your Web site but you can control your own internal link flow. Link flow is not PageRank."

To me, the second statement (choosing how to link within your site) clearly does control the flow of PageRank on your site.

Frankly, I find Cutt's remark puzzling, in that he takes no issue with the 1st portion that says "You cannot control the flow of PageRank on your Web site," changes the wording of the 2nd portion such that it contradicts the 1st, and does not challenge the final statement that "Link flow is not PageRank."

[activeco]

Originally Posted by deepsand

Frankly, I find Cutt's remark puzzling, in that he takes no issue with the 1st portion that says "You cannot control the flow of PageRank on your Web site," changes the wording of the 2nd portion such that it contradicts the 1st, and does not challenge the final statement that "Link flow is not PageRank."

Well, Matt assumes that everyone thinks simple and efficient.
He just thinks that it is the common knowledge that not all link flow is necessarily the PageRank flow, but that all PR flow goes through the links, even internal ones.
Strange, isn't it?

[deepsand]

Originally Posted by activeco

Well, Matt assumes that everyone thinks simple and efficient.
He just thinks that it is the common knowledge that not all link flow is necessarily the PageRank flow, but that all PR flow goes through the links, even internal ones.
Strange, isn't it?

Strange indeed.

For, if one interprets his remark to mean that all PR flow is Link flow, but not all Link flow is PR flow, then he would seem to be supporting Michael's claim that you cannot control internal PR flow. And, were this the case, why even make his comment?

In view of other comments of his, my take is that Cutts is given to making comments that are deliberately obfuscatory.

[activeco]

Originally Posted by deepsand

In view of other comments of his, my take is that Cutts is given to making comments that are deliberately obfuscatory.

Only if you want to see them in that way or if they are contrary to your assumptions.
My impression of him is that he's honest, knowledgeable and very straight, up to the point. Of course he's limited with the Google's trade secret, but in most cases he can express himself clearly without breaking it.

[deepsand]

Originally Posted by activeco

Only if you want to see them in that way or if they are contrary to your assumptions.
My impression of him is that he's honest, knowledgeable and very straight, up to the point. Of course he's limited with the Google's trade secret, but in most cases he can express himself clearly without breaking it.

I've no doubts regarding his knowledgeability, and, make no assumptions re. his statements.

Rather, I see his bent toward obfuscation as being a natural result of his commitment to faithfully discharging his obligations with respect to maintaining confidentiality.

My observation in this respect is, therefore, not a condemnation of his remarks, but rather an advisory that one should be exceedingly careful about believing that they wholly and accurately understand all that they read into his statements.

My personal preference would be that, when unable to speak plainly in a manner that fully and accurately addresses the issue at hand, he simply refrain from commenting.

[Peter (IMC)]

Originally Posted by deepsand

Frankly, I find Cutt's remark puzzling, in that he takes no issue with the 1st portion that says "You cannot control the flow of PageRank on your Web site," changes the wording of the 2nd portion such that it contradicts the 1st, and does not challenge the final statement that "Link flow is not PageRank."

This is an example you making problems that aren't there. Cutts talks about 2 statements of which he put each in between quotes. You all the sudden see 3 statements. Obviously Matt says that he agrees with the second statement which includes what you split of as a third statement.

This way nobody is ever right about anything in your book.

[kevsta]

Originally Posted by Webnauts

One thing I still did not understand about the whole PageRank Sculpting story.

As we all know, Google is against PageRank manipulation. Or?
And now their employees are telling us how to manipulate the PageRank of our sites, using i.e the nofollow attribute.

Did I miss something?

Just thinking out loud.

this has had me pondering for a little while now too.

and great thread guys, best discussion have seen on this anywhere.

[janeth]

Originally Posted by kevsta

this has had me pondering for a little while now too.

and great thread guys, best discussion have seen on this anywhere.

There is no way that Google can stop people from controlling links within their website.

We can use other things like flash and java instead of the nofollow tag. This has been around for a long time, the only thing that is new is using the nofollow tag. And the fact that you can come into a site without having to make a lot of design changes to keep the bots from following a link.

[kevsta]

Originally Posted by janeth

There is no way that Google can stop people from controlling links within their website.

We can use other things like flash and java instead of the nofollow tag. This has been around for a long time, the only thing that is new is using the nofollow tag. And the fact that you can come into a site without having to make a lot of design changes to keep the bots from following a link.

no for sure, but to go from saying that one must not do anything purely for the benefit of search engines, to saying "oh go on then it's ok to use nofollow to play around with our PR it may help you rank better" seems quite like a u turn on a fundamental "given" ?

btw i am utterly convinced there is merit with this too. no its not very predictable, & its certainly not a sterile environment within which to theorise or test, but sometimes a suck it and see approach can unearth things no amount of debate or procrastination ever will.

am not anywhere near confident enough to advise or use on clients sites or anything thats not my own as yet, but i wouldnt rule it out at all in the future.

[deepsand]

Originally Posted by Peter (IMC)

This is an example you making problems that aren't there. Cutts talks about 2 statements of which he put each in between quotes. You all the sudden see 3 statements. Obviously Matt says that he agrees with the second statement which includes what you split of as a third statement.

This way nobody is ever right about anything in your book.

Michaels's 1st sentence is compound, consisting of 2 separate statements. Together with his 2nd sentence, there are 3 statements in total.

Elementary, Dr. Watson.

[crankydave]

Michael calls controlling PR within a site a myth. Says you cannot control PR flow only link flow.

To me, Matt's reply to Michael is pretty fundamental, since links are the only way you can "flow" PR. Controlling where your links point is controlling where PR flows.

I also understand the point Michael is making. Since it cannot be known how much PR a particular link passes, if it passes PR at all, you cannot "control" what cannot be known.

All that is really is known is that PR is preserved and cannot be left on the table. So, preventing it from going one place means it has to go somewhere else. While you can control where the links point, you can't control, on the granular level, how much of it goes where. You can only point.

Dave

[incrediblehelp]

I was on a call with Matt Cutt's, Adam Lasnik and other Googler's and I brought this topic up at the end of the call. Basically Matt said if you site navigation is fundamentally sound (which it should be in the first place) he said you would be wise spending your time in other optimization techniques.

Take that as you will.

Update on PR Sculpting Question from Google Webmaster Help Groups Call and Matt Cutts - Search Marketing Blog from Cincinnati, Ohio