Do Relevant Links Matter?

[lesaurus]

I agree with what Micallef, but at the same time I think SEs can only give you so much weight from relevant blogs. Simply because they can not accurately determine what a relevant link is.

Of course a marketing website linking to a marketing website is relevant, but what about linking from a Playstation 3 website to a Call of Duty site? Or a Health blog to an Acne Treatment site? Just because the niches don't include the same keywords and key phrases doesn't mean they're irrelevant. And similarly 'healthy ageing' and 'domain ageing' include two of the same keywords, yet are completely irrelevant. There is no way they can accurately keep track of the relevance of keywords.

Of course they can. That is the sole purpose of search engines, to measure relevance. They have ever growing methods to measure and track keywords. Their algorithm is THE most important asset that they have, the corner stone of thier business. There are lot of comments here that are written in such a matter of fact way as if they KNOW this algorithm. You don't know SHIT we don't know SHIT, we can only guess and test and even then it will probably change.
I am not going to tell you the answer coz I don't know, but the way I think of things is by trying to think like the owner of a search engine would. How would I determine the relevance of a backlink when assesing the relevance of a site. If you can answer this then you are closer to understanding the algo.
Posted via Mobile Device

 

[biks]

I've seen many cases of using articles (ezine) to get a link. You just write some bullshit article and do a non-related link in the bio box. I've seen plenty of tiny niche sites use this and get high SERPs. The deal is you don't want your own ezine article to rank higher than your little niche site.

 

[oldenstylehats]

There are lot of comments here that are written in such a matter of fact way as if they KNOW this algorithm. You don't know SHIT we don't know SHIT, we can only guess and test and even then it will probably change.

As you've pointed out, there isn't any point in trying to understand the algorithm of any specific search engine because it will inevitably change. However, anyone who has studied modern semantic indexing theory at an academic level can make plausible inferences as to how relevancy is most likely calculated.

In other words, some of us do know SHIT.

 

[megler]

As you've pointed out, there isn't any point in trying to understand the algorithm of any specific search engine because it will inevitably change. However, anyone who has studied modern semantic indexing theory at an academic level can make plausible inferences as to how relevancy is most likely calculated.

In other words, some of us do know SHIT.

I love threads like this. Damn guys, I just want to know the best way to link my baseball site to my screwdriver site for best results...

 

[oldenstylehats]

I love threads like this. Damn guys, I just want to know the best way to link my baseball site to my screwdriver site for best results...

Write an article on your baseball site that somehow relates to baseball to screwdrivers and embed a link for the screwdriver site in it.

 

[megler]

Write an article on your baseball site that somehow relates baseball to screwdrivers and embed a link for the screwdriver site in it.

k. I'll make a field in the database that will accommodate that so I can rotate links in and out by just changing the id. thanks!

 

[yowies]

And..on a side note..internal links, anchor text or otherwise.. pointing back to a main page on the same site are also relevent.
I have seen online newspaper sites for instance which had massive amounts of links showing in Yahoo and sometimes google and they were old pages pointing back to the root domain.

 

[Micallef]

How much can we know about Google's relevance assessment abilities?

Well, after ranking #1 several times using only autospammed, non relevant rubbish links, I know that either:

a) They don't want to fully judge the relevance of links quickly and effectively

or

b) They can't fully judge the relevance of links quickly and effectively


I don't think it's a.

This doesn't go against what has been said earlier. Relevance is best by far and Google is very good at picking up and rewarding positive relevance, but for reasons unknown, Google will count irrelevant links (which, to be honest, should be cancelled out if google is to provide the best results) towards even a #1 ranking - if there are enough.

 

[oldenstylehats]

Right, there isn't a question whether irrelevant links matter, they do. The question is to what degree they contribute. Once again, it is a spectrum. Google has more than enough collective computational power to apply any number of different relevance/similarity algorithms, all of which would be able to perform the task at hand. The result is generally an index value between 0 and 1, in which nothing would ever receive a score of 0 or 1.

If you're mathematically inclined and want more information, check out this book.

 

[biks]

a) They don't want to fully judge the relevance of links quickly and effectively
or
b) They can't fully judge the relevance of links quickly and effectively

Ya know, I was thinking about this exact same thing. It seems it would increase the load on everything if they did exactly the same analysis of EVERY page that is out there. Then is every 5 word phrase compared to every other 5 word phrase out there to see if it's duplicate content?

Then how does Google store the "relevance data" of a site? Another meta data file that we don't know about? Or is Google just one giant database and everything is queried on the fly? It sure looks like it's just grouping webpages together by only using their links between them.

 

[oldenstylehats]

Ya know, I was thinking about this exact same thing. It seems it would increase the load on everything if they did exactly the same analysis of EVERY page that is out there. Then is every 5 word phrase compared to every other 5 word phrase out there to see if it's duplicate content?

Then how does Google store the "relevance data" of a site? Another meta data file that we don't know about? Or is Google just one giant database and everything is queried on the fly? It sure looks like it's just grouping webpages together by only using their links between them.

If you really want the answer, read the book I referenced in my post. That is how that part of the process works. It isn't like Google runs on a few servers. They have clusters of clusters of clusters. Computational power isn't the problem.

 

[BlackxHat]

Very informative. What do you guys think of software like xrumer? Use it or not?

 

[rinrin]

I am so glad to stumble upon this thread. Very useful info that I can apply to my business. Thanks guys for sharing your insights.

 

[d3t0x]

I think xrumer, if used correctly, can be an awesome tool without relevant backlinks.

 

[biks]

re: Information Retrieval: Algorithms and Heuristics (The Information Retrieval Series)(2nd Edition) (Paperback) (oldenstylehats book suggestion from above)

Ouch. My math anxiety is setting in. Does it have lots of pictures in it?

 

[gorang]

I think, it does matter. But if you're getting really high PR links from non-relevant sites, you will likely see a really good increase as well.

 

[g111k]

To be honest I think it can be summarized and put quite simply:

a) Relevant links within relevant context has some mythical, minimal benefit in terms of your SERP... as many have pointed out, it will take less of these "high quality, hard to find" links to rank well, as compared to irrelevant links within irrelevant context.

b) On the other hand, relevant links within relevant context will have obvious, significant benefit in terms of the quality of traffic that will be channeled to your site. I don't think we need to debate on how well the traffic will convert when clicking on a "auto insurance rates" within an auto insurance related article... compared to a "online gambling" link on a dog training site.

 

[goawayplease]

Links convey three important things for off-page optimization: Trust, Popularity, and Authority.

The reason EDU and GOV links are so valuable is in part because of their Authority. Links coming from pages of topical relevance can likewise increase topical Authority to a page.

Building a slightly relevant link can be as simple as this: Do you need the best Viagra? We've got it!

Google's algorithms look at the words around the link to try and create some sense of context for the page it's about to visit. It's a step towards LSI.

 

[Rhab]

I use to get so caught up over issues like this and things like dupe content. I would drive myself crazy over debating it, learning and researching.

Bottom line, create as many links as it takes to overcome the competition. Relevant or not, do more than the competition and relevancy all of a sudden becomes irrelevant =p

 

[Rhab]

I use to get so caught up over issues like this and things like dupe content. I would drive myself crazy over debating it, learning and researching.

Bottom line, create as many links as it takes to overcome the competition. Relevant or not, do more than the competition and relevancy all of a sudden becomes irrelevant =p