Finding the Most Powerful Links
Given that many large brands and mainstream media sites are trying to leverage their brand strength by adding interactive content to their sites and every SEO blog in the world (and some from distant universes) have posts about leveraging social media and buidling trust with link baiting, it is probably a pretty safe bet to think that Google is going to be forced away from trusting core domain related trust...and it is going to have to get even better at filtering link quality as well. You know Digg spaming is mainstream when there are many different business models based on spamming Digg.
Other social media sites are not behind the curve in getting spammed to bits. I recently noticed spam software for mass submission of videos to video hosting sites, and I see del.icio.us and Technorati pages ranking everywhere, and when I look at Del.icio.us I run into tags like this
Wow. Garbage.
When you look in Google's search results for long tail queries in consumer finance or other profitable verticals you see many sites rank which are various flavors of forums, user accounts, xss exploits, and other social spam. In spite of Yahoo! being the most visited website compare Google's recent stock performance to Yahoo!'s. Given that content as a business model does not scale well, traditional monopoly based content providers are going to have to work hard to get users to create / add value to / organize their content. As they do, many of these types of sites will make it easier and easier to leverage them directly (easy to rank content host) and indirectly (indirect traffic and direct link authority) to spam Google.
The brief history of SEO (well as best I know it) sorta goes something like
matching file names
page titles and meta tags
keyword density
full page analysis
link count
pagerank
anchor text
manual and algorithmic link filtering
duplicate content detection and semantic analysis
delaying rankings
delaying indexing
and now we are up to site related trust...which is getting spammed to bits and will only get worse
Anything that has been greatly trusted has been abused. The difference between the current abuse and past abuse is that in the past it was typically smaller individuals screwing with Google. Now Google has become a large enough force that they are actually undermining many of the business models of the providers of the content they are relying on.
Going forward, especially as Google, small content providers, unlimited choice, and easier access to the web marginalize the business models of many of the sites Google currently trust those sites are going to rely on users to help foot the bill. Google will give some content providers a backdoor deal, but most will have to look to user interaction to add value. That user interaction will be spamville. Thus I think rather than just trusting core domain levels I think Google is going to have to reduce their weighting on domain trust and place more on how well the individual page is integrated into the site and integrated into the web as a whole.
If everything Google trusts gets abused (it eventually does) and they are currently trusting raw domain related trust too much (they are) it shouldn't be surprising if their next move is to start getting even more selective with what they are willing to index or rank, and what links they will place weight on.
Jim Boykin recently announced the launch of a his Strongest Subpages Tool. Why the need for it?
If you’re getting a link from a page that no other site links to (beyond that site), what is the true trust of that page?
However, if you get a link from a subpage, that has lots of links to it, and your link is on that page, there’s outside trust flowing to that page.
If you’re getting links from pages that only has internal links to it, I doubt there’s much value in it.
Jim's tool has been pretty popular, so if you have trouble accessing it don't forget that you can do similar with SEO for Firefox. Jut search Yahoo! for site:seobook.com -adsfasdtfgs, where adsfasdtfgs is some random gibberish text. That will show you how well a page is integrated into the web on many levels...page level .edu links, external inbound links to that page, etc. etc. etc. You can also go the the Yahoo! Search settings page and return 100 results per search.
Comments
Thanks Aaron. I wasn't think you were saying that user created content was the key, though I agree it can be useful.
I think I was with you on this. That there are those under the impression that user created content is the key (or just more important than it is), and by placing too much emphasis on content created not by themselves, but by anyone who wants to contribute it can lead to an erosion of quaily. And that Google is doing things to further this impression.
At least I think that's what you were saying.
SEO Specialist wrote:
If schools would have enough funding, they would not look into options like that. The old rule that schools are to be a commercial free zone can be torn apart anyway. The media industry did already a good job turning kids into walking bill boards.
Based on that premise am I not seeing anything wrong with schools receiving funds via affiliations. The problem is (as it is always) who schools affiliate with and what they promote.
Hi Aaron, well basically as I see it Google has a major problem with adding value to sites mainly on the number of external links. As you mentioned every webmaster get gets tons of links, so its getting to easy..
One option that might be used is the proper code used on the site. But that can't be used, since majority of websites are using non valid codes. I agree that as seo once should focus on deep linking as you mentioned, but then again, it's to easy, most of the webmasters are doing that.
Best regards
A more comprehensive link graph makes search more relevant, not less...so just the number of links out there is not an issue that hurts Google, it is an issue that helps Google.
What hurts Google is that they are undermining the investment that goes into quality content and causing many of those businesses to fight back by having consumers play a more active roll in their sites.
Valid code is a nearly useless measure of information quality IMHO. In many cases I even think it is an indication of misplaced focus, false ideals, being naive, and lower quality content.
I agree with you that valid code is no indication of better informational quality, but I disagree that having valid code is misplaced focus or lower quality. It's just a sign of someone who wants to write valid code. I think validation and seo are two separate issues.
One of the things about the historical list in the post is that as you move forward in time each gets a little harder to spam. Not impossible, but harder.
No matter what system is in place there will be people trying to game it. It's simply a part of human nature for some people to want to beat the system. Even if playing within the system is easier and has a greater return there will be people trying to game it.
I want to make sure i'm following, but is the point that as Google keeps evolving the way to succeed will be to rely more on your visitors for content and thus your content gets watered down and loses quality?
Hi Steven
I was not saying that user participation is the key to creating great content...just that many business models that are growing irrelevant are looking to user participation to solve the problem of paying their costs in a world of near infinite choice.
I think most strong brands have a good bit of their founder or company ideals in them. Interaction is a way to create virtually free content and brand evangelists if it is done well though.
Kind of seems like the same old story. One person takes advantage of the system and therefore the entire system is tightened for the whole group.
Instead of distrusting everyone by default couldn't there be another way to integrate user-defined relevance with indexing? I don't know, honestly your post sounds almost hopeless as if things are only going to get worse and we'll all end up in an eternal sandbox.
Thinks will keep getting more complex to manipulate if you do it completely from an algorithmic perspective and do not want to spend time learning or money to help boost you along, but somebody has to rank...may as well be me, etc.
"I think rather than just trusting core domain levels I think Google is going to have to reduce their weighting on domain trust and place more on how well the individual page is integrated into the site and integrated into the web as a whole."
It looks pretty much as a page rank, doesn't it? I assume it may work to rank web pages within a trusted domain, but not among domains.
Aaron, you work with large customers. Some of them are probably public US companies. You know, it is hardly possible to get a legitimate link from such a website, because of privacy considerations. They make links only to the websites they work with. Do you mind, that a link from a poorly linked page of a highly trusted domain should weight more than a link from a highly linked web page of poorly trusted domain? I suppose the corporate domain content is very difficult to manipulate with. Therefore it can be considered a kind of reliable reference source. Isn’t it a good search basement Google is looking for? If it is, why you predict Google will undermine the values of trusted domains in favor of well linked web pages? Look on the last Matt Cutts post: http://www.mattcutts.com/blog/how-google-handles-hacked-sites/. It is an evidence that Google invests to improve the quality of trusted domains rather than to diminish their value in search algo.
I've been questioning the continued viability of a citation-based ranking system for some time now. Well, since the linkflation that occurred during the green bar madness days.
Link paths, link trust and link structure analysis will always have a role to play, but link values need to be reassessed. When links are treated as currency, and the checks and balances that are in place to prevent counterfeit currency from being produced are ineffective the value of a link is suspect. This problem is exacerbated by the fact that ANYONE can print currency.
We can dispense with the notation that the current system is a citation-based system in the academic sense. The 'papers' aren't being peer-reviewed, and citations from non-specialists carry little weight in academia. The current system isn't a voting system either. Votes aren't counted equally and people can vote as many times as they like.
Reciprocal link values discounted? Does that make sense? Reciprocal links are at the core of natural linking structures.
No follow? Maybe, but ONLY for user-generated content. Any other application of nofollow is disingenuous. I like to take that one step further and just say that unmonitored user-generated content is mostly shit. If you don't have the resources to monitor it, don't allow for its creation.
Search technology will get better, and SEOs will learn that blindly following the arbitrary 'get more links' dictum isn't necessarily in their best interests.
Some are. Some are not ;)
If they put you in your topical clique they are probably great. If not I think there is a poor risk to reward ratio to getting too aggressive with it.
That's just it. If the engines can determine 'topical cliques' (link relevancy), then why bother with the FUD about paid links? Why bother with nofollow? A link from my blog or your blog to some casino is off-topic and if the engines can determine that...
The link in my name above goes to a most recent result of their trust in domain names: a viagra spammer hacked a bunch of .edu (yeah right G doesn't trust them more...) and put his own viagra pages on them. Guess what? He's got most of the top 10 spots for the query buy viagra with that.
We're just a step away from schools monetizing research with affiliate links...
On a related note, the post linked to also offers some legitimate tips on reputation management. Most large companies want to control their reputation online and the link offers ideas on how to own many of the top 10 spots in the SERPs.
Add new comment