Posted 21 March 2005 - 07:03 PM
There's a lot of good stuff in this thread - at least it's on the brink of being really good. But, there's some stuff that's just plain wrong.
A big part of the problem is that there are a lot of terms flying around and technology has advanced since those terms were created. It happens all the time, and you'll know that what I say is true the next time you dial your telephone.
As technology advances, people like to reuse old terms - for simplicities sake, probably, though it really only goes on to confuse things. This thread started out talking about Link Popularity and updating PR. The nature of what these things are remain exactly the same as they always have been - in simple terms, they are a measure of inbound links to page A multiplied by a factor of the link popularity of the linking page - page B.
In the olden days, link popularity and PR by themselves were a big factor in SE rankings. If you had enough links from enough popular pages, you would rank well - plain and simple.
Now it's different, but the importance of link popularity has, if anything, increased. It's just that how it affects the serps is different.
I can't possibly begin to explain this all here, but I'll try my best to give a general overview.
There are a few things you need to get your arms around first, though. The first thing is that Google (and all the search engines) aren't looking at pages as just "whole pages" anymore. They are breaking pages down into elements. These elements are everything from "Navigation" "Titles" "Descriptions" "Paragraphs" and, oh - I dunno, lots of different things - different pages have different elements that are important, so.... If you want to see search engines breaking things down into elements, look at Froogle and see how products that aren't fed to froogle (i.e. They are crawled by the spiders) have a price, a description, a product name, and a picture. Sure, it's not always 100% accurate, but it gets it pretty darned close - and it does it by breaking the page down into elements. You can also see this at Google news - the news spider hits a page, grabs the headline, a photo, and quite often, it even pulls out the proper "teaser" paragraph or sentence to show on the page, too. (AllTheWeb used to isolate specific descriptive sentences separate from the blurb - but it stopped showing those once Yahoo took over).
Once you understand that a page has elements and search engines can identify those elements, we need to look at how the links work. And this, right here, is the bulk behind what's causing the "Sandbox Effect" that folks are talking about all over the place.
On a simple and basic level, a link is a link is a link. Get all the links you can. But, on a deeper level, links are even more important than they used to be. The bulk of the concept here is called "LocalRank" and it basically works the same way as Pagerank does, except that its value is based upon the specific search term being used. If someone searches for something, the first 100 (or 1000? I dunno - they never said) documents are feed into a box and run through the link database. So, a page that has inbound links from other pages within that subset of pages gets a higher LR value than a page that doesn't have any of those pages linking to it. (Note also, that I may have a link from a New York Times article, and that helps my overall PR, but if that specific article doesn't come up somewhere in the search, it doesn't help my LR score, even though the Times is an authority).
Now, we also have the other thing that I've been talking about for a while (well over a year, anyway), but that tends to be given the brush off by most. It comes from the technology purchased from Applied Semantics a good while back. Bascially, it's a means of determining the actual "meaning" of something - it's theme or even its specific topic - by looking at how that document relates to other documents.
This part is very complicated, so I'm going to simple-it-up as best I can. Basically, you start with a "seed" where you "know" what a set of pages is about. It's presumed that Google uses the DMOZ as the seed, but that may not be the case. The Google Directory has a list of web pages (it looks at each link in the directory as a link to a page, not the site - which is a bit different from the way DMOZ does it). In this case, it actually helps your site to have a deeper listing in the directory than a higher level one, since the deeper you go, the more specific the topic.
So, now you've got a bunch of pages that you know what they are about and a tree of subjects that you can climb up and down (the DMOZs struture itself). Then you break those pages down into elements and see all the pages that those pages link out to (both internal and external) and which page element those links are from and you also see what pages are linking into those pages (both internal and external).
Here you're sort of making data soup. All of this is compared against various real search terms that bring up a page where the topic is known and compared against pages it links to and has links from that are also in those results. Elements of the page are compared to see which parts are talking about the same thing (i.e. using common words, phrases terminology - especially in the proximity of the link).
Blah Blah. Stir Stir. Pretty soon, through the WAY pages are linked to each other and how nearby text relates to other text in related pages, you've got Applied Semantics and you can, in a way, Know What A Page Is Saying.
So, now back to link popularity - it's all still important, but it's different. None of these newer things are going to give much value to a link you get from a random page of links (i.e. a link exchange page). It WILL help to have links from a links page that has an identifiable topic and one where the topic has been determined by Google, but it's hard to know which ones have.
Outbound links are now critical to your success, as well. I can lie in my content, so a SE can't place all its ranking weight on content because of that. I can't control who links to me nor can I always even understand WHY someone has linked to me, so using inbound links to my page isn't perfect. But, even though it would seem to be the contrary - it's my outbound links that are the hardest to spoof and lie about. Let's look at it...
I have a movie site. I'm making a list of the people and films that won oscars this year. Google knows, over time, that my site is about movies, but now it's looking at this new page. The content seems to be about movies, but what's it about specifically? Well, that's hard to say. It's got lots of words on the page. But, there's a link to the Oscar.com winnders page in there. Google knows that that page is SPECIFICALLY about the Oscars, and not just movies in general. There are common elements on both of our pages - both list some of the same actors and movies. Those common elements allow Google to know that my new page is not just about movies, but it's about Oscars, as well.
Now, knowing what I know, I COULD try to spoof it and link out to something not relevant to the page, but that would help me rank for that subject. Well, it just won't work because there aren't any common terms on the pages. I COULD try to spoof it by creating a doorway site and link to that page and have that doorway page share the same subject - but that dummy site I made has no Semantic Authority (there's a term I just coined for this post - look for it in next year's SEO Glossary). It takes a while (seems like a year or longer) for your site to gain any Semantic Authority. Semantic Authority is pretty topic sensitive too - I can't just have a broad site to use for boosting rankings of my other sites that I'm going to make a year from now - it'll never get authority. So, basically, it's spoof-proof.
In the end, this Applied Semantics technology, though very confusing (sorry I'm not better at explaining things) is probably the biggest advancement in SE tech since, well - ever.
I hope this clears up some of the questions. Yes, links are as important as ever - probably even more so. Link popularity (as originally defined) isn't so important, but it's still the seed to all of this.