How to prevent PR drain?

[godsakes]

To channel more PR into my internal pages & linking partners. I'm think about preventing my affiliate links from passing on PR, what's the best way of going about this?

I've know javascript links tend to be the most common, but can google now index them and will there be compatibility issuses?

I've noticed other sites have their affiliate links point towards an 'exit' page which then uses a meta redirect to the advertised site. Could internal links on these 'exit' pages redirect PR?

More recently I've noticed some kind of php trickery on links can this be done on ASP?

Thanks in advance

[Geir]

Another member has this site in the signature:
http://seo-toys.com/link-hide-seo-toy/via-toy.shtml

I haven't tried it myself yet, but it seems to be exactly what you are looking for.

:) Geir

[Owlcroft]

Yes, (thank you, Geir), I'd reckon it is.

It ("Via" (http://seo-toys.com/link-hide-toy/via-toy.shtml) ) is a very simple php script that you place in some convenient directory, then block in your robots.txt file.

There is fuller explanation and documentation at the site.

[dsr771]

Is it accurate that a slight page rank is lost for outgoing links?

It does seem true that only so much PR can be given from a page to other pages. But does the page with the links going out actually loose PR of its own?

Maybe that answers my question. :)

Thanks

[Bizcut88]

I use clicktracker on some of my sites.

http://www.focalmedia.net/clicktracker.html

[Bernard]

You can also use the free affiliate link cloaker (http://www.seo-help.com/html-tools/profit-protector.html) and disallow the html pages it produces with robots.txt.

[vlead]

I prefer JavaScript links for all outbound links. There were rumours that Google was able to parse JavaScript but it is technically very risky (for Google).

[Owlcroft]

Is it accurate that a slight page rank is lost for outgoing links?
Read this online paper (http://www.webworkshop.net/pagerank.html) and draw your own conclusions.

[godsakes]

thanks everyone,

I don't use PHP so I guess the javascript options are my best, I'm giving the 'affiliate link cloaker' a try.

Bernard: You can also use the free affiliate link cloaker and disallow the html pages it produces with robots.txt.

I noticed it produces some kind of exit page with a javascript unicode redirect, any disadvantages over a meta refresh?

I'm sticking these files in an 'exit' directory is this the correct robots.txt instructions?

User-agent: *
Disallow: /exit/

thanks again

[Bernard]

Yes, the robots.txt snippet looks correct. The Javascript is actually not a redirect. It frames the target site with an empty frameset. It does not work for visitors who have Javascript turned off BTW. You can always add a <noscript> section if you like.

[Owlcroft]

I don't use PHP so I guess the javascript options are my best
JavaScript is not your best bet.

The "Via" toy mentioned above is quite self-contained: you put it in a handy directory of your choice, you put one line in your robots.txt file, and you write any normal HTML links you want to block just a little differently, pointing to the toy. It's all documented and easy (and free).

There are other ways, too (CSS, for one), but this is a simple, bite-sized little toy.

[Help Desk]

Page Rank Drain or Leakage is a myth. If such thing occurred there would never be high pageranking sites.

[mopacfan]

Page Rank Drain or Leakage is a myth. If such thing occurred there would never be high pageranking sites.
Thank you ThinkBling, I was just going say the samething, well not exactly, but it would have been very similar :)

[godsakes]

Prehaps 'drain' is bad word,

'Channelling' is what I'm aiming for.

But before this turns into a moral debate, let me say that I'm no fan of SEO I wish google or another search engine simply just finds good content (and my site with it). But it doesn't so tough, SEO I must and every little helps.

[Owlcroft]

Page Rank Drain or Leakage is a myth.
Sigh. As the Count famously said, "One more once."

Read this online paper (http://www.webworkshop.net/pagerank.html) and draw your own conclusions.
When you feel prepared to dispute its conclusions with a matching depth of data and mathematical analysis of that data, please, by all means, do so here. I have yet to have that oft-repeated challenge answered, which is suggestive.

I am not an expert: like Will Rogers, I only know what I read in the papers. If you want to challenge what's in online SEO research papers, fine, I'll read your paper too.

I'm waiting for the latest edition . . . .

[aspen]

Page Rank Drain or Leakage is a myth. If such thing occurred there would never be high pageranking sites.
Its not a myth, it is mathematics.

If you have a pizza and 3 friends and you all must eat, each of you gets 25% of the pizza. If you have a pizza and 7 friends and all must each, each of you gets 12.5% of the pizza. Adding friends = less food for everyone.

Links and PageRank are the same way. Every link on a page gets a share of the page's PageRank. Of the links on the page some are links to other pages of yours, others are links to someone else's pages. If you add more links to someone else's pages your pages get less PR. If you add more links to your pages, the other person's pages get less PR. If you remove all links to someone else's pages your pages share all the PR.

So less outgoing links = more PageRank flowing through your site.

Not a myth, mathematics.

[Help Desk]

...When you feel prepared to dispute its conclusions...

Here's a well formed post from compar concerning linking truths (http://forums.digitalpoint.com/showthread.php?t=800).

Here's another with actual mathmatical calculations to a PageRank calculation (http://forums.digitalpoint.com/showthread.php?t=856).

Also read the following post that I will make in regards to aspen's post as I break it down in simpler terms.

[Help Desk]

...If you have a pizza and 3 friends and you all must eat, each of you gets 25% of the pizza. If you have a pizza and 7 friends and all must each, each of you gets 12.5% of the pizza. Adding friends = less food for everyone...

Sometimes we look so hard at the details, we lose sight of the big picture.

PageRank is Google's measure of the importance of a page. PageRank is not to be confused with SERPs (Search Engine Result Pages). What makes something more important? There are numerous things. What is a good indicator? “Reference” is. A person on the mind of millions is more important than a person on the mind of hundreds. A good example is the president. He is on many more people's minds than my bling (http://www.thinkbling.com/forums) self.

Now if the president talks of a specific country, does he lose importance or pass/leak his reference off to somebody else? Never. Having him reference something else might actually make his reference increase.

However, if he references 2 things, each individual topic might not get the same focus as when he references just 1. Is each topic's focus halved? No, but it is lessened.

To compare, lets take site A with 1 link versus and site B with 100 unique links (duplicate links probably don't carry additional weight). The link from site A should have more emphasis than any 1 link from site B. However if the above type of algorithm holds, the total weight of all links from B would be greater that of A.

I did not choose to find mathematical evidence on the web as every article must be taken with a grain of salt as it is impossible to do any of these tests in a closed environment. All that one can do when you pose a page-rank question is to think about the larger algorithm and its intentions. Page-rank leakage just doesn’t make sense.

[aspen]

A page does not directly lose PR for linking out. Other pages on the same site just get less.

If you don't think its good search engine design to implement this type of algorithm you are more than welcome to your opinion. However it seems to me that you haven't thought about the larger algorithm.

Have you read Sergey and Larry's thesis or that PageRank explained paper? Do you understand the mathematics of it?

The average PR of all pages on the Internet must = 1. When Google is calculating PageRank every iteration of the equation MUST lower the average PR of the Internet, or they will never reach 1.

So, in order for PageRank to decrease with every iteration of the algorithm, two things must happen.

1. Links have to divide the PageRank of a page equally. So, to use your analogy, if the president talks about two topics each would have to be halved in focus. They have to be. This is so PageRank doesn't increase.

2. There has to be a dampening factor, or PageRank won't decrease.

Now, here is reasoning for #1.

If the PageRank on page X given to outside link A does not exactly equal the PageRank removed from internal link B then if X links to A and A links back to X (or A links to a middleman which then links back to X). Then PageRank will increase in an infinite loop. For each iteration of the algorithm global PageRank will increase. They will never reach the normalized sum of 1. They will be caught it what is basically a feedback loop, like when you put a microphone next to a speaker.

Its mathematics. The equation is unbalanced. The PageRank passed on cannot exceed the PageRank possessed.

If you want to believe that outgoing links do not detract from the PageRank that is passed to internal links, I can't stop you. I've dealt with such believers for years and for some people no matter how much evidence you show them they just can't change.

The problem with such people, in general, is that they do not understand why search engines exist. They think it is unfair, undemocratic, or unethical to hide links or to refuse to link you to someone. They think that implementing this algorithm is unfair to webmasters and does not promote a utopian Internet.

In reality search engines do not exist for webmasters, they exist for consumers, they exist to provide relevant results, not to provide a fair and balanced battleground for webmasters. They do not care if their ranking algorithm is fair, only if it is relevant.

So, don't worry about whether an algorithm is fair to webmasters, worry about whether it produces relevant results.

In truth, outgoing links do not hurt that much. If the link is on a subpage and is outnumbers by internal links the PR loss is very very small. However, if you put a plethora of outgoing links on your homepage, or if every page of your site has some, then you can use significant PageRank from them. If the place you're linking to links back to you though you'll likely get most of it back.

Also, by adding outgoing links, someone might find your site more useful, and that might inspire them to link to you, so you gain in that way.

So you can argue whether it is ethical or even needed to engage in so-called "PageRank Hoarding" but, do not argue against the mathematical certainty that all links on a page must share the total PageRank of the page and that the total weight of the links on a page cannot exceed it's total PageRank.

[compar]

I think there is one last step that is being missed in this discussion. It is true that if you double the number of OBL from a page each target page including the internal pages will only receive half the PR they previously did. See my article and chart which is published in the SitePro News (http://www.sitepronews.com/) today.

The point that is being missed is that this means that each internal page has that much less to pass back to the home page. So excessive outlinking from a page does diminish the pages PR, all other things being equal.

However the people who are paranoid about this are not usually talking about doubling the number of outbound links. They are obsessing about even one or two OBLs.

This type of paranoia is ridiculous for two reasons.

1. PR has almost nothing to do with SERP placement today.

2. There is a growing body of evidence, that sites with a substantial number of outbound links, are getting higher SERP placement.

So forgetting that hiding reciprocal links is selfish, and bordering on dishonest, it may in fact be working against the page. I think this is what is known as ironic justice, and I invite all my competitors to continue to hide all their outbound links. Just don't come crying when your page doesn't show up in the SERPs.

[aspen]

The point that is being missed is that this means that each internal page has that much less to pass back to the home page. So excessive outlinking from a page does diminish the pages PR, all other things being equal.


Absolutely. But when dealing with people who don't believe it I find its better to introduce them to the concept one step at a time.


2. There is a growing body of evidence, that sites with a substantial number of outbound links, are getting higher SERP placement.


Whether or not search engines count outgoing links, or if they even should, is a whole nother discussion. If you find a site with many outgoing links ranked well it is a coincidence. The outgoing links and the ranking are not related.

There isn't a shred. People have been saying for years that outgoing links help and they don't. It is a mathematical certainty. As certain as 2+2=4. This type of speculation is just a result of people making uneducated observations in an uncontrolled environment.

Not only that but it makes no sense. If all one had to do was link to a popular authority site everyone would do it and the ranking would be meaningless. It would basically allow people to vote for themselves.

Plus it doesn't increase relevance. Search engines are not in business for webmasters, they are in business for users. Search engines want relevant results. The most relevant result is an authority site, not a hub site. If you think of it logically you would see this.

A search engine wants to provide access to information. If there is a hub site and an authority site which one should the search engine show to the user? Do you really think the hub site should rank better because it links to the authority site? If the authority site is so good why not rank it better? Why tell the user "We know where to find the information you need, but check out this site instead, where you can find a link to the information."?

SERPs are hub sites. A SERP provides links to relevant pages. A hub site provides links to relevant pages. They fulfill the same purpose. Except the search engine has done more work analyzing and ranking the linked to pages, and undoubtedly has more pages to offer.

The only search engine that uses outgoing links is Teoma, and they only use them to create a seperate result set off to the side. This type of implementation is the only one I can see any search engine using as it is the only one that makes sense.

Incoming links are a measure of usefulness as judged by other webmasters or just average joes with a geocities site. If outgoing links make a site more useful, as in the case of DMOZ, it will have incoming links that reflect that.

So why would Google, or any search engine, bother guessing on if outgoing links made a site more useful or not when it can just poll the globe on a site's usefulness?

Sometimes javascript menus make sites more useful, should Google give bonuses for those? Of course not, if they make a site more useful it will be reflected in the incoming links.

The more useful a site is, the more incoming links it has. So by measuring incoming links you're in effect measuring the usefulness of a site's outgoing links (along with everything else).

You'll be hard pressed to find any well known search engine professional who would stand behind the claim of outgoing links helping. Danny Sullivan has called the idea absurd in one of his newsletters. Dan Theis feels the same way. (http://www.seroundtable.com/archives/000289.html#comments)

If all of that isn't enough for you, there is the raw proof.

http://www.searchenginelabs.com/

The last experiment on the page. 3 trials, the pages with outgoing links do not rank better than those without them.

[Help Desk]

Can you sum up what your point in a shorter post? Your point is being lost. You might be making things much more complex then they should be.

[T0PS3O]

It's perfectly clear to me what he means, and makes sense too.

[compar]

Whether or not search engines count outgoing links, or if they even should, is a whole nother discussion. If you find a site with many outgoing links ranked well it is a coincidence. The outgoing links and the ranking are not related.

There isn't a shred. People have been saying for years that outgoing links help and they don't.
Aspen,
You don't seem to be very upto date. Maybe all that stuff that you have known all these years has changed. There is such a thing as being stuck in the past. Learning is a daily ongoing endeavor. Whenever I hear somebody say, as you did two or three posts ago, that you have known it all for years, I get nervous.

There has been a lot of evidence coming out of the recent SEO Conference in London England that one of the Google representatives at the conference confirmed that OBLs were being used in SERP placement determination.

I don't have a link to this information off hand, but maybe someone else does. Or I'll find it, later in the day, and post it for you.

[Help Desk]

It's perfectly clear to me what he means, and makes sense too.
Let's take a step back. Why would out bound links ever hurt? How does hurting a site by providing out bound links give a better indication?

I'm not asking for a re-iteration of the formula everybody thinks is Google's. I'm asking a broader question.

[GuyFromChicago]

There has been a lot of evidence coming out of the recent SEO Conference in London England that one of the Google representatives at the conference confirmed that OBLs were being used in SERP placement determination.


I believe this to be true, and think it makes perfect sense.

An OBL to a "authority" site is most likely very useful to your site's visitor. Why wouldn't Google (or the other SE's for that matter) give a page some more importance in the serps if it links to useful info?

Even if it's not true as of today, like the whole "themed" linking debate, I think it will be in the future.

Regardless, an OBL to a site you visitor would consider useful is a good idea. Put on your PR blinders and build a site/page that you visitor will find usefull.

[T0PS3O]

Why would out bound links ever hurt?

It won't unless you're building a linkfarm and get penalized.

It will only if you are PR-PaRanoid and OBLs do dilute your precious PR.

[Help Desk]

It won't unless you're building a linkfarm and get penalized.

It will only if you are PR-PaRanoid and OBLs do dilute your precious PR.

You do realize you just contradicted yourself in the same post right?

It won't...
It will...

I'm sure it was just do to hurrying. Are you saying that OBL's reduce you pagerank and that doesn't hurt?

[aspen]

Aspen,
You don't seem to be very upto date. Maybe all that stuff that you have known all these years has changed. There is such a thing as being stuck in the past. Learning is a daily ongoing endeavor. Whenever I hear somebody say, as you did two or three posts ago, that you have known it all for years, I get nervous.

There has been a lot of evidence coming out of the recent SEO Conference in London England that one of the Google representatives at the conference confirmed that OBLs were being used in SERP placement determination.

I don't have a link to this information off hand, but maybe someone else does. Or I'll find it, later in the day, and post it for you.
It is you who is incorrect. There is no direct evidence to support this. There is plenty of evidence to refute it, and no one can come up with a plausible explanation of why this would be done in the first place. Your best evidence is some rumors at a conference. I can show you a controlled experiment where outgoing links are shown to have no effect on a page's ranking.


Why wouldn't Google (or the other SE's for that matter) give a page some more importance in the serps if it links to useful info?


Reread my longer post above. Then read what I'm about to write.

How does Google know if outgoing links add to the usefulness of a page? By comparing the topics of the two pages in question? Being on topic doesn't make something useful. Useful is subjective by nature. So how in the world does a search engine measure something that is subjective?

Can you tell if a page is useful by analyzing its content? Absolutely not. A computer cannot tell if a page is useful by analyzing content, it can only tell if a page covers a certain topic. Usefulness is subjective.

Sergey Brin and Larry Page had to answer these questions. They came up with the idea that you could measure a site's usefulness by measuring it's incoming links. They theorized that the more useful a site is, the greater chance someone will link to it, so incoming links could be a measure of usefulness.

If outgoing links add to the usefulness of a page then that increase in usefulness should be reflected in their incoming links. As such measuring incoming links is all you need to do.

If Google, or any other search engine, used outgoing links directly as a measure of usefulness then sites where the outgoing links did not increase usefulness, such as sites that sell text links in the footer, would recieve bonuses they did not deserve. Search engine relevance would be hurt, and it'd be a big step back.

People have been saying for years that Google either uses outgoing links or they will. It is a persistent rumor. Also for years I have been saying the same things I have been saying now. No one has ever offered any proof that shows me to be incorrect. The pages I have with 0 external outgoing links continue to rank well. The controlled experiments people run continue to show that outgoing links do not help.

Until one of those things changes I am going to believe, recommend, and advise that outgoing links do not help you in search engines.

I don't tell people they shouldn't link out, I don't tell people outgoing links will hurt them. What I tell them is that outgoing links do not help, and that making them will sacrifice some of the PR you could be sending elsewhere. However that PR in the typical situation is small and usually not worth worrying about, but in extreme circumstances outgoing links can have a substantial negative effect on your rankings.

[Help Desk]

...How does Google know if outgoing links add to the usefulness of a page? By comparing the topics of the two pages in question? Being on topic doesn't make something useful. Useful is subjective by nature. So how in the world does a search engine measure something that is subjective?...

This is not a PageRank issue. PageRank is based entirely off of the PageRank of incoming links. It is a general importance indicator.

However, this is a big SERPs issue. If 5 sites all rank equally for a search term or phrase but only one of them has a link to the other 4 then it is easy to see that this site is a better repository for information.

[aspen]

This is not a PageRank issue. PageRank is based entirely off of the PageRank of incoming links. It is a general importance indicator.

You admit PageRank measures general importance. But does general importance not encompass usefulness?

However, this is a big SERPs issue. If 5 sites all rank equally for a search term or phrase but only one of them has a link to the other 4 then it is easy to see that this site is a better repository for information.

If the site is a better repository why does it not have more incoming links than the others? Obviously the world as a whole doesn't think its any better, or it would have more incoming links. So why would Google second guess the opinion of the Internet community?

[Help Desk]

You admit PageRank measures general importance. But does general importance not encompass usefulness?
No. If you are looking for a way to use PHP to start your own Amazon.com commision earning store, my site is extremely useful. If you are looking for a way to cure cancer, my site is not usefully at all. PageRank is a general indicator and not the ultimate one.

Go to Amazon.com, it will forward you to a sub-page. This sub-page has a page-rank of zero. However this page is extremely usefull if you want to buy something from them.

If the site is a better repository why does it not have more incoming links than the others? Obviously the world as a whole doesn't think its any better, or it would have more incoming links. So why would Google second guess the opinion of the Internet community?
Links are just indicators and are not a fool-proof way of finding the best results. Try searching for a specific coupon on Google. You will get a bunch of krap.

If a search engine only uses one indicator (IBLs) to search, then the results can be more easily skewed. If you are using 2 indicators (OBLs and IBLs) the results can be better. For example the US treasury does not raise rates based solely on the number of new homes sold per month.

Now if the weight of my IBLs is calculated to be 1,000,000 and yours is 1,200,000 but I have reference to the top 10 sites of the search and yours doesn't, then my page would probably be more targetted to the search.


On a side note, when it comes right down to it, none here really are looking for a high pagerank or top-ten search results placement. We are just looking for targeted traffic.

[aspen]

No. If you are looking for a way to use PHP to start your own Amazon.com commision earning store, my site is extremely useful. If you are looking for a way to cure cancer, my site is not usefully at all. PageRank is a general indicator and not the ultimate one.


Incoming links are a measure of usefulness, they do not tell you the topic of content.


Now if the weight of my IBLs is calculated to be 1,000,000 and yours is 1,200,000 but I have reference to the top 10 sites of the search and yours doesn't, then my page would probably be more targetted to the search.


Maybe your site might be better, but why would Google both to guess when it can just poll the Internet community? Look at backlinks, look at anchor text. Looks like more people find site A a good reference than site B. Site A is probably better, who knows, maybe site B links to site A and most of the visitors to site B end up at A anyways. So, how can B be more useful than A because it links to A?

In anycase I have no wish to argue with you about how a utopian search engine should be run. I only wish to discuss how search engines are run, and I've not seen any empircal evidence to support your points yet.

[compar]

It will only if you are PR-PaRanoid and OBLs do dilute your precious PR.
Right. But PR-PaRanoia is ridiculous. How many times do we have to tell these paranoid manics that PR is almost meaningless.

If you look at my chart you will see that the range of values for any given PR ranking is large. So the chance that a few backlinks -- and that is all we are talking about in these debates. Someone wants to hide their affiliate links. How many affiliate links can you have off on one page?

So the chance that a few backlink would actually drop you a whole PR number is very small. And if it did so what?

[Owlcroft]

The simple reality is that no one outside Google knows for certain how SERPs are calculated. Period, the end.

To say that this or that does or does not "make sense" is to assume that what makes sense to you or to me or to him or to her must necessarily make sense to Google, and vice versa; that is an assumption so fraught with unrealities as to be risable.

"Experiments" are always a good idea, but I wonder how many of the experimenters could calculate a Standard Deviation if their lives depended on it. For even a sniff of meaningfulness, experiments require a substantial (and controlled) sampling. I don't see that in those I've heard tell of.

Google is a seriously imperfect institution. Take some topic you know well and Google it and ask yourself whether, if a good friend asked you to provide a list of 10 or 20 web sites on the topic, Google's top 10 or 20 are the ones you'd give. Google thrives only because it fits Churchill's remark about democracy as a system of government: the worst possible except for all the others that have been tried.

Considering the overhead of getting into the game on a serious footing, Google is probably safe for a while, as its two chiefest competitors are excellent shots, if one's own foot is a favored target.

[T0PS3O]

You do realize you just contradicted yourself in the same post right?

It won't...
It will...

I'm sure it was just do to hurrying. Are you saying that OBL's reduce you pagerank and that doesn't hurt?

I didn't contradict myself. The answers were just tailor made for the reader's level of interest on PR rank.

Spelled out:

Outbound links won't hurt IMO because I don't care the least about PR, I link to whoever I want as long as I think it is likely to increase turnover and or profits or benefits the general usefullness/customer experience.

It will hurt however if you are one of those PR freaks who can't sleep because they are so worried about the current dance and that it might drop, but only if indeed G's secret algo DOES dilute PR thus dropping your PR thus affecting you personally or professionally because you will be sad about the loss of PR.

So choose your answer, both are true.

[Help Desk]

Incoming links are a measure of usefulness, they do not tell you the topic of content....

How so? The anchor text does as well as the words surrounding it in the sentence and on the rest of the page.

[ResaleBroker]

Another member has this site in the signature:
http://seo-toys.com/link-hide-seo-toy/via-toy.shtml

:) Geir

Thanks for the tool link! My site is a good candidate for this since I provide a lot of outbound links to outside resources; gov't pages, etc.