MODERATOR NOTE: This was split from Can Inbound Links Hurt You? (http://forums.searchenginewatch.com/showthread.php?t=3033), and while some comments are Google-specific, the thread's been moved to the link building area as this is broadly applicable to that topic.

*some links may not help you, but none will hurt you*

So you'd see run of site links as a benefit to be actively pursued?

some links may not help you, but none will hurt you
So you'd see run of site links as a benefit to be actively pursued?

moi?
The only reason I would go for a sitewide is for the traffic. If I had an advertisement on another site that had high click-thru and conversion rates, then I would probably want it viewed by as many different site visitors as possible. We'd have to do a CBA on the increased traffic versus the increased cost of that traffic to make that determination, and theoretical speculations about the effect this would have on the SERPs would stand aside.

My belief, and my recommendation to those who actually still wonder about such things, is that one should only get one single link per domain, which is the optimum number of links to get per domain. Why? Because 1 link from 1000 different domains has roughly 1000 times the value of 1000 links from a single domain.

All too often we see posts from webmasaters complaining that they have a PR7 site with 10,000 backlinks and some PR4 site with only 500 backlinks is kicking their serpy butt. "Why", they ask. Investigation quickly reveals the 10,000 backlinks are coming from just a handful of domains, with thousands of links per domain, whereas the 500 backlinks are all coming from different domains, 1 link per domain.

Multiple links from a domain, such as site-wide, may count for more than a single link, but not much more. Site-wides are deeply discounted on a per-link basis. The only reason to take a discounted link is for the traffic actually going through that link. If you are in the game for topologically accentuated (anchor text) link-pop, just get one link per domain and move on to the next one.

I don't make the rules, I just try to figure out what they are and make the most of them.

Because 1 link from 1000 different domains has roughly 1000 times the value of 1000 links from a single domain. At last, someone has posted some common sense instead of common belief!

Why? Because 1 link from 1000 different domains has roughly 1000 times the value of 1000 links from a single domain.

Interesting speculation, but can you provide some evidence that this is in fact the case?

The fact that some domains have 10,000 links and are being outranked by a domain with only 500 links is easily explained by the one with 500 links having more relevant anchor text links than the one with 10,000 links. IMO its not the number of links that are important, or even the number of anchor text links, its the number of relevant anchor text links that are going to get you rankings.

relevant
how do you define that relevancy?

In the context of search, IMO relevancy is simply relevant to the keywords in the search, If you search for dried prunes you don't expect to get pages back about peaches.

Similarly having 500 anchor text links containing dried prunes as the anchor text is not going to help you much in a rankinq quest for peaches.

What I am saying Aaron is that its not how many total links you have thats important, its not even how many anchor text links you have thats important, its how many anchor text links that target the term which the search in question addresses that is important.

If you are asking indirectly if I think the links have to be from relevant sites or authority sites or whatever, then no, I am not saying that simply because I see no evidence whatever that you will get a better anchor text ranking for computers if the anchor text link comes from a computer related page or from an authority site.

If you are asking indirectly if I think the links have to be from relevant sites or authority sites or whatever, then no, I am not saying that simply because I see no evidence whatever that you will get a better anchor text ranking for computers if the anchor text link comes from a computer related page or from an authority site.
I was just trying to get your take a bit better :)
cheers

No Problem Aaron, though we seem to be getting far off topic.

The Universal Rule of Thumb Truth:
Incoming links cannot adversely affect your rankings.

The Universal Rule of Thumb Exception Truth:
In rare circumstances incoming links can hurt you.

neuron - Good, well-reasoned, thoughtful posts.

I'm more negative about run of site links than you are, as I think they tend to raise a flag. Were Google ever to go after paid links, run of site links could be one of the first places they'd look. Sites getting them might also be scrutinized more closely. This is assuming that there are such things as algorithmic flags.

Since there's such good input on this thread, let me ask my paranoid question about another linking situation... which is, can hidden links from other sites hurt you?

One of the sites I optimize receives links that make me worry because they're hidden links. The linking pages are always fake directory pages, generally with Toolbar PR0 or 1, with visible affiliate links and various advertising network links (generally not AdSense).

In addition to the visible advertising links, though, there are hidden links to the top sites scraped from serps for the search this page is "optimized" for.

No telling why they're doing this... whether they've read somewhere that linking to top ranking sites helps... or whether they're trying to sabotage us. I get nervous when I see (in the source code) my page title in an anchor with the first paragraph of my page, sometimes repeated half a dozen times in hidden text. While many of these disappear as fast as they appear, they're around at least long enough for Google to show them as backlinks.

Any thoughts about where these are these among the rare, potentially harmful inbounds, or whether I should just be happy for the anchor text while it lasts? Policing this kind of stuff, I should mention, when you get near the top in a competitive area, can drain a lot of energy.

sure. Define "almost" to me first.


Reggy, I would suggest reading this from Google (http://www.google.com/webmasters/facts.html) in context with what lots0 has said;

Fiction: A competitor can ruin a site's ranking somehow or have another site removed from Google's index.

Fact: There is almost nothing a competitor can do to harm your ranking or have your site removed from our index. Your rank and your inclusion are dependent on factors under your control as a webmaster, including content choices and site design.

*Interesting speculation, but can you provide some evidence that this is in fact the case?*

I think I might, though it may not be popular here.

While looking at some of the directories that advertise on the internet.commerce network, I thought I'd look at how some of the other advertisers were doing.

I picked three at random, chosen for keyphrase and not being a Y! shop, price comparison site etc.

1.garage heater :
On G, site was no. 38 showing 18,900 links, 139,000 in Y!
However, for those KWs, site no. 6 showed just 45 links in G, 102 in Y!
(Y! position 1)

2. Flower delivery :
Site was no. 4 on G, showing 31,200 links, 494,000 in Y!
Again, site no.3 showed just 337 links, Y! 5,130.
(Y! position > 80)

3. Health insurance:
Couldn't find the site in the first 50 in either G or Y!, G showed 9,540 links, Y! 84,300.
On G, site no. 5 for those KWs showed 404 links, 1,840 in Y!

With the numbers involved, and bearing in mind we're not just talking links here, but targeted anchor text, both G and Y! seem somewhat unimpressed by them.

What is also interesting, these links don't just come from a single domain, they come from a recognised network of domains.

There is an answer for all this and the answer is the same as "how do search engines deal with spam?"

We should all be long familiar with page spam. Matt had some nice examples of this in Vegas during his PowerPoint presentation. Pages are generated that are basically the same except that keywords are changed from page to page, placing keywords in the URL, the Title, H1 headings,and at various places in the page itself. Page spam has become more sophisticated over the years, as his examples showed, but it started out as "doorway" pages with gibberish interspersed with keywords. How do the search engines deal with this?

Link Spam is a similar phenomanon. Some webmaster will have half a dozen domains at his command so he loads some pages with hundreds of links to another domain, or perhaps he puts site-wide links up, one link per page to every other page on all his other websites. This happens. I'm not making this up. In fact, I've done this myself. In the past I've even seen some success with such methods, just as I've had some success with page spam, but it will continue to be a less and less successful practice as the search engines are aware of these tactics and continue to take measures against them.

One way the SE's could deal with link spam is to just count a single link from domain to domain. Certainly the SE's will deal with it by reducing the impact of thousands of links from one domain to another, but how much will they reduce it? If they are interested in stopping link spam in its tracks, then the ideal reduction would be to a single link per domain.

But you protest: "My links are valuable relevant links. They use relevant anchor text to point to relevant content. My links aren't spam."

And the SE's say: "Very good, we're sure the site visitors will find your thousands of links very valuable. Nevertheless, due to the exponential growth in iterative calculations needed to properly calculate link popularity, and your unreasonable demand on those resources, we're still just going to count 1 link for our calculations.

If you don't agree with this, please contact your favorite search engine and explain to them why you think your thousands of cross-links should be included in the calculations. Maybe if you throw in a couple of dozen PCs to help do all the calcuations you are trying to impose on them they'll listen to you. However, you still don't get my vote.

I may very well be wrong about this, but from my not-well-documented observations the algorithmic identification of multiple links from domain to domain does not happen early in the calculated process and may not take place until after an initial complete algorithmic calculation has been made. That is, the reduction from many links to one link for the purpose of the web-wide calculation does not happen until after the first full calculation. Symptomatic of this is that sites receiving link spam seem to initially rank high in the search engines only later to be reduced in rank as the many links per domain are reduced to one. Also symptomatic of this is the cry of "penalty" from those link spammers who declare that "My site was ranked high, and now I think they've discovered I did some heavy cross-linking and I have lost all my rankings, they've penalized me."

My take on it is that despite the fact that Google has more than 500 engineers working on their engine, they do not have enough engineers to deal with your link-spammy site. Rather, they would deal with all the link-spammy sites in one fell swoop by way of reducing 1000's of spam links to 1 link per domain. The fact that this reduction does not take place early in the calcuation thus leads many to believe they have been singled out as somehow important enough to get some special attention.

Again, you can't go wrong if you just get one link per domain. Yes, it's hard work, yes it takes a lot of time, and yes such links are unlikely to be discounted at all.

I have never met a webmaster who has not at some point in time gotten some lame-brained idea (I would be included in this lot) of generating page or link spam. Yet, the internet and it's rewards are much like they are in life. Having six of my friends point at me with their thousands of fingers while exclaiming "he's the man! He's the man!!" is not going to make me an important person. But if thousands of people are pointing at me, then perhaps I have done something worthy of note.

sure. Define "almost" to me first. I'm not sure why you are asking me to explain Google's use of the word but here is the meaning as per Dictionary.com (http://dictionary.reference.com/search?q=almost&r=67)

Slightly short of; not quite; nearly: almost time to go; was almost asleep; had almost finished.

neuron, I agree with your thoughts and observations. To me (and the silent minority/majority) it's very likely that one link per domain is all we get credit for. Google give a BIG hint at this when thay say (bolding added by me) "PageRank relies on the uniquely democratic nature of the web by using its vast link structure as an indicator of an individual page's value.". Having one site cast more than one vote is not democratic by any stretch of the imagination.

Unfortunately, this is not what many want to hear so they simply wont.

I would have bolded "uniquely democratic", though the "democrat" says it all.

I realize most people probably are aware of this to a certain extent. However, in Vegas I saw some rather stunned, dawning realization expressions when these ideas were mentioned, on what I had otherwise assumed to be rather astute and web-studied faces.

A side-effect of misunderstandings about this is that people think site-wides can hurt them. Site-wides do not hurt you, but they don't help you much either, and really, site-wides only help you if you are getting traffic through them.

You can see the depth of this misunderstanding, or perhaps better the depth of the abuse of this misunderstanding when you visit your favorite text link broker and see that he is selling site-wide text links. If the broker understood this and was not abusing the lack of knowledge of this sellers and buyers, then he would explain these matters on his website, and urge them to trade just one link per domain. Then, not only would such services provide more impact, but more impact for the dollar. I would even think that the text-link broker who first came forward with such offerings might even be at a competitive advantage. For starters, he would lose the buyer's worry that site-wides could somehow hurt his site, and with it the incidents of sites momentarily surging in ranks because of the site-wides only to plummet back down as the site-wides are consolidated to one.

Glengara are you saying that you manually checked those 139,000 links and verified that all are relevant (to the search term ) anchor text links?

Neuron if what you are postulating were in fact happening then all but one of the internal links in a site would not count. I can show you dozens of sites that rank well on nothing more than internal anchor text links.

I fail to see what mentioning page spam and well known individuals names has to do with the subject under discussion unless Tim in fact stood up on his hind legs in front of people and said "Google only counts one link from each domain"

Did he, or are you making broad interpretations of what he said on one subject and applying it to another?

While you have blithly assumed that all text link brokers are blindly selling sitewide links, I think if you took the time to study the situation, you might find that this is not the case at all, and you might also find that those who are in the business of selling links perhaps understand what is happening in the link world better than the rest of us. They have the data from millions of links and the results of those links to know best what works and what doesn't.

I do agree that in the past Google has devalued links from certain domains that were known to be selling links, but so far as I can determine this had nothing to do with sitewide links.

I can show you dozens of sites that rank well on nothing more than internal anchor text links. I guess you well have to define "rank well" and what term you are saying they "rank well" for. Also, the pages must blank and/or make no mention of the anchor text if it is to be taken seriously as any sort of evidence. Then you have the task to prove that it's all (and not one) the internal pages are causing the "rank well" situation.

Without this, your speculations does nothing to disprove those that say otherwise, which BTW (in absense of any absolute proof) is the more likely situation based on common sense and the fact that Google state;

PageRank relies on the uniquely democratic nature of the web by using its vast link structure as an indicator of an individual page's value.

I guess you well have to define "rank well" and what term you are saying they "rank well" for. Also, the pages must blank and/or make no mention of the anchor text if it is to be taken seriously as any sort of evidence. Then you have the task to prove that it's all (and not one) the internal pages are causing the "rank well" situation.
when I search for "viagra" or "online casino" I realize that anchor text is driving the rankings there. I do not need to see a blank page to know what is going on there.

PageRank relies on the uniquely democratic nature of the web by using its vast link structure as an indicator of an individual page's value.
to take one of Google's branding statements and say that thier technology follows the idealology behind their branding statement is naive. That is a Google branding statement...nothing more and nothing less.
Google Branding Strategy (http://www.v7n.com/google-branding-strategy.php)

when I search for "viagra" or "online casino" I realize that anchor text is driving the rankings there. I do not need to see a blank page to know what is going on there. Please read the thread in context if you must answer on someones behalf. You are waaaay off.

to take one of Google's branding statements and say that thier technology follows the idealology behind their branding statement is naive I didn't say that they striclty adhered to this statement. What I did say was, "in the absense of any abolute proof..."

To say it not true because it's on Google.com is really naive. Following your logic (or rather lack of) this statement from Google is also false;

Google's mission is to organize the world's information and make it universally accessible and useful.

Google's mission is to organize the world's information and make it universally accessible and useful.

should read:
Google's mission is to organize the world's information and make it universally accessible and useful, while profiting by delivering targeted advertisements near the information.

They have that text on the side of their black stealth helicopter :rolleyes:

should read:
Google's mission is to organize the world's information and make it universally accessible and useful, while profiting by delivering targeted advertisements near the information.

Given the fact that 98% of thier income is from "contextual advertising" (isn't it nice how they can dream up this kind of obfuscation rather that just saying PPC ads?) could it be that this is the real focus of Google and all those engineers are not working on ways to build a better organic search, but instead are working on ways to make Adwords even more profitable?

Given the fact that 98% of thier income is from "contextual advertising" (isn't it nice how they can dream up this kind of obfuscation rather that just saying PPC ads?) could it be that this is the real focus of Google and all those engineers are not working on ways to build a better organic search, but instead are working on ways to make Adwords even more profitable?
the illusion of a much better search technology is necissary to build the network to make AdWords more profitable though.

So I guess the Public Relations people work on the illusion, which the engineers work on Adwords?

illusion Never knew Google did illusions.

How dare Google profit from running the most popular SE on the planet. How dare they send me so much paying traffic via free inclusion and AdWords. :rolleyes:

seobook, if you have a site, I take it you have a robots.text file telling Google not to index any pages of your site?

*Glengara are you saying that you manually checked those 139,000 links and verified that all are relevant (to the search term ) anchor text links?*

Thought I'd leave that for someone who is either unfamiliar with the internet.commerce network, or wishes to desperately clutch at straws ;-)

I think I might, though it may not be popular here.

While looking at some of the directories that advertise on the internet.commerce network, I thought I'd look at how some of the other advertisers were doing.

Thought I'd clear up this reference -- it's about the little internet.commerce links that show up in the left-hand side of the page.

These are links Jupitermedia, which publishes this site, runs across all the sites it operates.

I've posted about this program before, but here's some additional background:

1) It predates Google and the heavy use of link analysis by search engines. IE, it wasn't designed as a way for people to build up links for search ranking purposes.

2) Jupiter tells me it is specifically not sold as a means of getting better rankings on Google.

I don't handle the program -- it's an ad thing, and I deal with only editorial. However, I can say anyone considering it should measure against my threegolden rules of link building (http://searchenginewatch.com/sereport/article.php/3422121), the key one in this case being:

Rule 2 - Buy links if visitors that come solely from the link will justify the cost

Personally, I'm not surprised if the sites in those boxes aren't ranking well for those terms. It's fairly easy for a search engine to see that the same link, with the same anchor text, may appear on hundreds of pages and decide to discount the weight of those links. Not ban! Not hurt! Just not give the links as much credit as if there were hundreds of links to a site with variety to anchor text and domain locations.

Because 1 link from 1000 different domains has roughly 1000 times the value of 1000 links from a single domain.At last, someone has posted some common sense instead of common belief! No - it's part of common belief. Or - at least - the idea of targeting links from a wide spread of IP ranges, rather than just links from one domain, has been a constant mantra at v7n:


Getting keyword rich inbound links from many different sites (from various IP ranges) is the single best way to improve your Google rankings.

- Aaron Wall
Best selling author, SEO book
www.seobook.com (http://www.seobook.com/)



Link building is indeed the way to top rankings. Getting links with the right anchor text is the key to SEO. But - here's the secret - it's not the number of links per se, but the number of unique domains and IP's that are linking to you.

- John Scott
Creator of Bluefind Directory (http://www.bluefind.com/) and the v7n webmaster community (http://www.v7n.com/).

It is also a foundation policy in a lot of commercial link building work.

However, link building isn't always about numbers within a site, or numbers across an IP range. Links are not all created equally.

But as to the original question - "Do Run Of Site Links Work? (http://showthread.php?p=25688#post25688)" - sure they do. But they are not necessarily going to be very effective in isolation.

The linking benefits from a single site are generally regarded as limited, but can build up.

* Not ban! Not hurt! Just not give the links as much credit..*

I'd agree with that, with one proviso.

IMO, G sees RoS links as a potential "links scheme", and which particular view is taken of them seems to depend on whether other "links schemes" are in play.