[RayB]

Originally Posted by RedEvo

Yup, and I'm saying in my view that's a good thing

d

I agree with you

They are great for traffic - and that's how it should be

[awebapart]

Part of trying to understand SEO is trying to get your head around how search engines work, almost reverse engineering search engines, and to reverse engineer you have to know how to engineer, i.e. develop software in the first place.

Speaking as a software developer, who has also thought about the SEO implications of signature links, my opinion is that it is almost impossible for search engines to directly work out what are and what are not forum signatures, especially since there are so many different forum systems out there with different html markup, many bespoke too (e.g. 4networking, startups, shell-livewire, etc).

So my opinion is that the clever search engines do not downgrade or ignore signature links from a forum website because they recognise the signature part of even recognise that the website is a forum. My personal opinion is that the clever search engines downgrade/ignore signature links because they see them as areas of duplicate content, small areas of content and links that are duplicated on many pages (and duplicated on the same page sometimes).

Of course this is just my opinion, me trying to second guess how the search engines work, and I could be wrong.

Even if I'm wrong on this, another SEO issue is the relevancy of the page the link comes from. If you have links in your post to your printing website, but you are posting more on pages not relevant to printing, then there wont be much relevancy there, so little value for SERPs. One thing I've found with most business forums is that there is surprisingly little discussion about printing and all the issues you might face when preparing documents for print (resolution, fonts, CMYK, etc), whereas mention web design, graphic design or SEO and everyone wants to start new threads, ask questions or join in the conversation on these subjects! What's the matter, does everyone know everything about print? (I certainly don't)

[Peter W]

I like the theory of diminishing returns. The first few links from a site might carry a bit of weight, but after that, every extra link carries less and less until their effect is negligible.

[PremierPDS]

Thanks - great idea for my first ever blog. How to prepare a document for print in plain ye olde English for those not in the know.

[RayB]

Originally Posted by awebapart.com

Speaking as a software developer, who has also thought about the SEO implications of signature links, my opinion is that it is almost impossible for search engines to directly work out what are and what are not forum signatures, especially since there are so many different forum systems out there with different html markup, many bespoke too (e.g. 4networking, startups, shell-livewire, etc).

As always Paul, a brilliant and thought provoking post. However - I'd suggest it's simple for SE's to work out Forums - I'm certain they have algo's to deal with the popular scripts and the rest can be dealt with by looking at repetition - which is what a sig is ultimately (IMO) .....

[batface]

I'm not sure that the removal of the links will devalue a site and lower it in SERPS, unless they form the major % of all links. The development of the links will assist, but I do agree they are not seen as of 'great' value to Google, but there IS value there. Once removed the site already has authority, other valuable links and will probably not lose any ground.

What I am trying to say is I believe that adding those links has had a positive effect on SERPS, but the removal none. Supposing the source of the links did not exist anymore, are you to be penalised for that? Unless almost all links derived from that site I don't thinks so.

It's a long winded and dangerous test that i'm not sure I want to try!!

Mick

[awebapart]

Originally Posted by RayB

I'd suggest it's simple for SE's to work out Forums - I'm certain they have algo's to deal with the popular scripts and the rest can be dealt with by looking at repetition - which is what a sig is ultimately (IMO)

From a software engineering point of view, it just wouldn't make sense coding a search engine, which is supposed to index any website, with special case scenario code for particular types of websites. You would be effectively web scraping different website coding styles and continually chasing your tail, reacting as site software vendors (e.g. the developers of vbulletin) change their code, e.g. looking for tell-tale signs (e.g.<!-- sig --> and <!-- / sig --> tags in this site's implementation of vbulletin's case), but even in the case of vbulletin, what about different versions of vbulletin, and different custom implementations (what if the website owner decided to take out the sig comment tags and other comment tags for performance reasons). Where would you stop? Would you write special code for the different forum systems, the different ecommerce systems, popular websites like amazon, bbc, wikipedia, etc?

In software engineering, like Occam's Razor principle, if there is a simple generic solution to a problem, that is the one that is usually chosen, and repetition/duplicate content is the simple solution in this case. If the google developers are smart, they will be investing their time in smarter generic code that solves many problems and works on all websites, rather than special case code that reads only certain websites.

The special case code google is working on for their search engine would be for special case web formats, like pdf, flash, etc. There is a separate part of google which does focus on usenet discussion boards called google groups but this adds a web format to a technology which is not in standard web page format.

That is not to say that other systems do write special code for extracting information from forums, e.g. forum web scraping with boardreader.com

From a marketing point of view I also think there would be an outcry if it was discovered that google wasn't treating all websites equally, with the same algorithm, as this would damage their claims for search results relevancy. You can't claim relevancy if you treat different sites differently with different algorithms, since you aren't then comparing like with like.

My view, which could still be wrong, is that there is only one generic algorithm used, at least that's how I would design the search engine.

[RedEvo]

As a matter of principle Google manually check search results. When they see 'anomalies' they tweak the algo to rectify it - rather than manually adjusting results. Now, logically, if a page is hitting number one, but the page is rubbish, and the reason it's hitting number one is because the page has a shed load of sig links, you might conclude they tweak the algo to give these links less weight.

As the algo has to 'rate' links as part of Google's search quality I can only imagine they are able to differentiate content links from sig links and programmatically this isn't too much of a challenge.

This is just my take on it as we are all pretty much guessing

d

[RayB]

Originally Posted by RedEvo

This is just my take on it as we are all pretty much guessing

d

Absolutely, but I suppose we can improve our guesswork through observation

[IP2]

If you put my signature into Google, it returns their home page - now how cool is that!?