What's the "Normal" Amount of External Links that Most Sites Have?

[ak]

[Randy Ray]

OK so if I want to make a site that's wholly funded by selling links for SEO, and I sell 20 links per post, that is normal?

I hear what you're saying, but I don't think that the number of links per post is the signal that's going to cause the devaluation. Suppose I run a blog that's not at all funded by selling links for SEO, but I post 20 links per post every day just to stuff that I find interesting? And I include insightful analysis into why those pages are interesting? Intent is hard to judge by a single signal.

It's more likely a combination of signals, and the number of links on page is probably the least important signal of all. If I were programming an algorithm that was hoping to determine intent, I'd look at multiple factors, and it would be a lot more complex than just the number of links per page.

For example, it'd be easy for Google to make a list of known linksellers and linkbrokers, then cross-index that with a list of sites that have a large percentage of links from those sources. You could set it at a percentage--so for example, if I run spammypokerlinkbuyer.com, and 80% of my backlinks come from known linkseller networks, then there's a good chance that I'm buying links too, but still, I'd want more of a signal than just that. You might also look at how unusual the anchor text is. I'm sure with Google's resources, it wouldn't be hard to create a list of keyword phrases that are more often targeted by linkbuyers than keyword phrases that aren't often targeted by linkbuyers. Now I've got two signals to look at. I'm sure the data-crunchers at Google could easily come up with a large number of behaviors or signals that correlate to linkbuying and linkselling, like the velocity of link growth, the percentage of anchor text that targets a specific phrase, the percentage of links to a site's homepage as compared to the percentage of links pointing toward internal pages, etc. None of those things by themselves are damning, but if too many signals correlate with the patterns of linkbuyers and linksellers, then boom, your site goes on the list.

That's just conjecture--I don't know any Google engineers--but I think it's a lot more similar to the way the algorithm behaves than it is to think that a single arbitrary signal like the number of links on the page determines whether or not the site is selling links.

[Newjabber]

What's the "Normal" Amount of External Links that Most Sites Have?

I don´t think there is a close answer to this, and i woulden´t spend too much time on it. Use common sence both when buyin and selling links and use many other different link/traffic building methods. I don´t think there are a million google spy´s manually sandboxing smaller sites for selling a few links, but if you are only linking to bad sites and getting links from auto blogs, sits with duplicated content etc. then you might trigger something.
IMO the keypoint is to try different strategies and do testing. I have sometimes found that when trying to "please" google with a pure whitehat site (i guess that would be the closest explanation) then that site might get bumped anyway or not do good in ranking and some less quality sites do great traffic and rankings wise.

You can get some indications of what works at the moment, but things change, so do testing and don´t be afraid to try new stuff to stay ahead. Google is a impossible constant changing riddle.

[addz123]

I hear what you're saying, but I don't think that the number of links per post is the signal that's going to cause the devaluation. Suppose I run a blog that's not at all funded by selling links for SEO, but I post 20 links per post every day just to stuff that I find interesting? And I include insightful analysis into why those pages are interesting? Intent is hard to judge by a single signal.

I agree with what you're saying. I run my own blog and I'd happily link out to lots of random sources in each post, especially affiliate websites which would be relevant to my posts and provide examples to users of whatever I'm talking about.

That's the reason I asked the question - because it's obviously not simple to spot link sellers purely based on link signals.

Content is a good metric for Google to monitor sites with, but I'm seeing an awful lot of link sellers using well written content, forget content being spun or anything.

It made me wonder how well Google can pick up on these link selling sites in comparison to high quality sites that link out liberally. I'm also seeing a lot of "mass" link buyers working for big firms, so a lot of the links themselves are going to big brand sites or even operators (i.e. good sources to link out to), as opposed to thin sites.

[parttimepoker]

Content is a good metric for Google to monitor sites with, but I'm seeing an awful lot of link sellers using well written content, forget content being spun or anything.

It made me wonder how well Google can pick up on these link selling sites in comparison to high quality sites that link out liberally. I'm also seeing a lot of "mass" link buyers working for big firms, so a lot of the links themselves are going to big brand sites or even operators (i.e. good sources to link out to), as opposed to thin sites.

Well, I mean - Google only cares to a point about link sales. I know they make a lot of noise about link buying and what not, but the pure fact of the matter is that in many genres, natural links just don't exist (or are super-scarce). Our industry is definitely one example.

At some point, there's a convergence between the sites that are high quality and make for good search results and sites that have the budget and organization to buy a bunch of quality links over a long period of time.

Google cares primarily about quality search results. Link buying is not always at odds with that goal. Google cares far more, for example, about poor content. That much should be evident from Panda; I can't think of any update in recent memory that has specifically knocked back sites guilty of buying links (and G has specifically said they don't drop sites in the SERPs for selling links).

They'd have a hard time doing so for a lot of queries; the first 2-3 pages of almost all gambling-related queries would vanish. What would remain, for many queries, would represent a worse search experience for the user.

That's not to say G doesn't penalize some sites for buying links, they obviously do. But I think three things indicate that people shouldn't worry too much about buying and selling links: a) distinguishing paid links from organic links seems quite difficult for humans, let alone algorithms, b) link buying and quality content are not mutually exclusive and c) the focus on so many other factors that don't have to do with link buying in the updates of the last few years.