PageRank leakage from bad link trades?

[LinkMaster]

I saw this reply on another thread, but didn't want to change the subject to pose a question:

Originally Posted by Mel

An anchor text link is an anchor text link and its the number of them that count not the PR of the page it comes from.

I agree, the more inbound links you have, the better. However, what if you are trading links with someone who has your link on a page with a PageRank of zero, and your link to them is on a page with a PR5? Would that be considered PageRank leakage? Would the trade be detrimental to your overall PageRank and/or search placement?

I've been wondering about this for some time now. I have been considering dropping all of my outbound links that are reciprocating with partners that have my link on pages with less than a PR1 or PR2 (all of my links pages have at least a PR4). Would this be a wise move?

[cbp]

There is no such thing as PR leakage.

Would this be a wise move?

No.

CBP

[LinkMaster]

Originally Posted by cbp

No.

Can you elaborate? I was basing my question on a PageRank calculator that showed a decrease in PageRank if a website was linked to wesbites with little or no PageRank. Are you saying that is an incorrect philosophy?

[lorien1973]

Originally Posted by LinkMaster

Can you elaborate? I was basing my question on a PageRank calculator that showed a decrease in PageRank if a website was linked to wesbites with little or no PageRank. Are you saying that is an incorrect philosophy?

Those calculators are for entertainment, nothing else. You cannot lose PR by linking to other sites. It isn't a sieve.

[cbp]

Its a total myth.

Try this an an experiment...
1. Find a page on your site with some PR (say 4) and put 300 links going out on it to PR 0,1,2 or 3 sites
2. Wait for a couple of Google PR updates to see what happens to the PR on that page (you might be waiting for a while due to there being no updates for a while).
3. Make sure the incoming links to that page stay the same.

I have done this several times.
Guess what happens to the PR of the page ..... NOTHING.

Also what about the 4.5 million sites that DMOZ links to, of which 4.49 million of them have PR less than what DMOZ have ... dosen't do them any harm.

BTW - what are you doing linking to sites based on PR? You are supposed to be linking to sites, because they are of use to vistor to your site.

CBP

[flood6]

EDIT: Wow, a lot of action in this thread while I was typing! If cbp says he has run tests that show PR is not "leaked" to pages that are linked to, I'll have to defer to him on that, I don't have any proof either way. I was just going on my understanding of the original PR equation. Here's my original post anyway:

I agree with cbp, if you're going to exchange links with the specific intent of boosting your PR, you'll want to do it with pages of similar or higher PR on their links page. The reason being, if you have a PR of 4 on your links page, you'll have no problem finding link exchange partners, and since you know you don't want too many links on one page (you'll look like a link farm) you'll want to save each page's "link real estate" for higher PR link partners.

Did that make since?

Now, there's nothing wrong with exchanging links with sites with low PR on their links pages, just realize you won't get as much of a PR boost. I'm not sure of the context that Mel's quote came from, but if you are working specifically on boosting your PR (and everything else is equal), a link from a high PR page will help you more than a low PR page.

When I use the term "PR leak", I'm talking about the inevitable passing of your page's PR to any page you link to. Unless you use javascript links or other methods to prevent the googlebot from following your links, you will be "leaking" your PR on to that linked page.

All that said, I really don't put too much stock in PR, I certainly don't ignore it, but I don't worry about it too much.

I hope that answers your question, or at least gives you my opinion on the answer.

[cbp]

When I use the term "PR leak", I'm talking about the inevitable passing of your page's PR to any page you link to.

Dead right. It is the misunderstanding of these sorts of comments that lead to the myth.

We do "leak" PR to pages linked to, but that does not mean that the "leak" reduces your PR (or as lorien1973 put it - its not like a sieve)

CBP

[cz]

N/T.,,,,

[LinkMaster]

Great stuff - thanks! Okay, back to my question. Let's say I have a PR4 links page and it has 100 outbound links on it. Half of them are being reciprocated by PR3, PR4, and PR5 pages, but the other half are being reciprocated by PR-zero pages. Would it be a good idea to dump the 50 PR-zero trades? Or does it matter?

Of course I like to offer good relevant links to my site visitors, but let's face it, I have to put food on my table too. Right now all my sites are enjoying number-one spots on Google for the most part. But a few of them are still climbing.

I've done some extensive research on this subject, and I have found that sites with a lot of inbound links and very few outbound links usually have higher PageRank than sites with the same amount of inbound links but more outbound links. This would contribute to the "PageRank leakage" theory. However, I know that Google has a PR10 and they are linked to virtually everyone. That fact seems to tear apart the idea of losing PageRank due to linking to sites with lower PageRank.

Any thoughts?

[LinkMaster]

To add some weight to the "PageRank leakage" theory, take a look at the following *images:


Image #1:





Image #2:





Image #3:





These images seem to lend credence to the "leakage" theory. If you link to a bunch of sites and don't get good reciprocal links in return, your PageRank decreases. On the other hand, if you link to sites with high PageRank and they reciprocate with links from high PageRank pages - your PageRank increases.

Again, I am still in the "guessing" mode, as many of us are (which keeps Google happy). So, I am definitely open to other opinions. That's why I posted the question. I don't want to just start deleting links to pages with low-quality reciprocal links if I don't have to. However, I don't want to give away all of my PageRank for nothing either - and that's exactly what I am doing based on the images above.

*Source of images: http://www.iprcom.com/papers/pagerank/

[cbp]

They are theoretical models that do not work in practice - see the exercise I posted above.

CBP

[LinkMaster]

Originally Posted by cbp

They are theoretical models that do not work in practice - see the exercise I posted above.

Your exercise seems to makes sense based on the fact that search engines themselves link to hundreds of millions of websites and don't get as many reciprocal links; yet, they still have high PageRank (i.e. Google's PR10 on its home page).

It's good to work through all of the rumors that are flying around the SEO forums. God knows, if you followed them all your head would be spinning!

So, are you saying that it's okay to link to as many sites as your heart desires, and it won't affect your PageRank? Based on that approach, are you implying that the only way to gain higher PageRank is to get inbound links from high PageRank pages - regardless of the number of outbound links on your site?

Along the same lines, will it hurt your site to have, say, 100 outbound links on a page? Is there any truth to this "link farm" idea that says you should only have 25 links on a page?

Thanks for clearing things up for me!

[Mel]

PageRank is based soley on the incoming links to the page, there are no terms in the equation that mention outbound links,except that the amount of PR passed is divided equally among each outbound link on the page.

Google suggests that you do not have more than 100 links on a page, but I have seen pages with 500 or more links indexed. When you get more that 100 links on a page however the spidering and indexing of the pages linked to can be slow, however.

Its not the number of links that makes it a link farm, what if your menu system has 100 links, would that make every page in your site a link farm?

[LinkMaster]

:)

[rlrouse]

1 - A link on your web page does not lower the PageRank of the linking page.

2 - A reciprocal link back to one of the pages on your site might increase the PageRank of the page being linked to or it might not (depending on the actual PageRank of the page linking back).

3 - Any PageRank gain of the page linked to in #2 above might increase the PageRank of other pages on your site (depending on your site's linking structure).

4 - #1 and #2 are completely independent of one another. CBP is right - linking out does not lower PageRank but a reciprocal link might increase it a bit.

[OBR]

When your home page links to your internal pages, you are giving some of your pagerank to your internal pages. Normally your internal pages would transfer pagerank back to the home page by linking back to it.

If you decide to add some external links on your home page, isn't less pagerank being transferred to your internal pages?

[cbp]

isn't less pagerank being transferred to your internal pages?

1. Probably
2. Its so minisule its not worth worrying aout
3. As PR is only one small component of the whole ranking formula, it even less of a factor to worry about.

CBP

[OBR]

Originally Posted by cbp

1. Probably
2. Its so minisule its not worth worrying aout
3. As PR is only one small component of the whole ranking formula, it even less of a factor to worry about.

CBP

2. with the exponential nature of page rank, wouldn't a swap between say a PR7 and a PR2 be meaningful.
3. couldn't we say that about alsmost all of the individual issues / questions.

[smo]

What about having more internal pages with link from home page? All will have back link to home page so drain of PR of home page will be compensated and all internal pages will have better PR after some time. I need not keep all my external links in one page and keep them in pages like friends, partners, sponsors, best sites etc. This will help me in keeping number of links per page to a minimum. But will this work ?

[LinkMaster]

Here's a very interesting finding that seems to support the theory of "leakage" (but I don't know if that's really the best term for what I am about to describe).

One of my websites shows 356 backlinks on Google. I come up as #1 on about ten good search terms. My nearest competitor has 158 backlinks. However, another competitor came out of nowhere and went to #2 - and he only has 76 backlinks. Here's the clincher. He disabled the link on his homepage that used to point to his reciprocal links pages. He displayed a new location URL on his homepage, but it was NOT a hyperlink (therefore the search engines didn't crawl it). He stayed at #2 for about six weeks. Then he finally caught his mistake and made the link a hyperlink. Ten days later he dropped down to #12 after the next update.

Do the math. Show me two competing sites with an equal number of backlinks, and I'll bet that the one with fewer outbound links has a higher position on Google under the same search phrases.

[cbp]

with an equal number of backlinks

The number of backlinks shown by Google are totally irrevalent to ranking - you have jumped to a conclusion based on the bad info Google choose to show in the backlink sample.

CBP

[Mel]

Originally Posted by LinkMaster

Here's a very interesting finding that seems to support the theory of "leakage" (but I don't know if that's really the best term for what I am about to describe).

One of my websites shows 356 backlinks on Google. I come up as #1 on about ten good search terms. My nearest competitor has 158 backlinks. However, another competitor came out of nowhere and went to #2 - and he only has 76 backlinks. Here's the clincher. He disabled the link on his homepage that used to point to his reciprocal links pages. He displayed a new location URL on his homepage, but it was NOT a hyperlink (therefore the search engines didn't crawl it). He stayed at #2 for about six weeks. Then he finally caught his mistake and made the link a hyperlink. Ten days later he dropped down to #12 after the next update.

Do the math. Show me two competing sites with an equal number of backlinks, and I'll bet that the one with fewer outbound links has a higher position on Google under the same search phrases.

Its good to remember that search engines don't rank sites they rank pages only.

[LinkMaster]

Originally Posted by cbp

The number of backlinks shown by Google are totally irrevalent to ranking - you have jumped to a conclusion based on the bad info Google choose to show in the backlink sample.

I would beg to differ. All the research that I have done, along with my own experience in getting 28 number-one positions on some very competitive phrases - all point to link popularity being the most important factor in placement on Google.

[LinkMaster]

Originally Posted by Mel

Its good to remember that search engines don't rank sites they rank pages only.

That's what I always thought too - but how do you account for my example? Also, I have seen sites add a lot of pages of content and then get ranked higher because of it.

Isn't it fair to say that none of us really know everything that Google does to rank a page? Isn't that why we are all here posting? If we knew all of their secrets, we would be too busy making millions as SEO consultants to be here posting.

I have found some things to work very well for me, yet others here disagree with them. I guess it all just adds to the mystery that keeps Google at the top.

[flood6]

Originally Posted by LinkMaster

I would beg to differ. All the research that I have done, along with my own experience in getting 28 number-one positions on some very competitive phrases - all point to link popularity being the most important factor in placement on Google.

He was saying that the "link:" operator used to find google backlinks will never show all the backlinks to a site. It doesn't even seem to show a proportionate number.

I just use it to get a general idea. If a backward links check shows 3 links, it is safe to assume they don't have too many backward links. On the other hand, if the check shows thousands, it is safe to assume they are doing fairly well with link popularity.

[LinkMaster]

Originally Posted by flood6

He was saying that the "link:" operator used to find google backlinks will never show all the backlinks to a site. It doesn't even seem to show a proportionate number.

Oh, I see. Sorry, I misunderstood - my bad. I agree, a Google "link:" search doesn't begin to show all of the backlinks for a page, but it gives a basic idea. All in all, when you do a backlinks search, the #1 site on a search term usually shows the most backlinks out of the resulting sites.