A bit of a review post here, but we’re going to use this when developing backlinks. We need to be able to judge if a site is a place we want to get a link from.
Bear in mind when reading this, much of this isn’t ‘factual’. I really don’t know how Google ranks sites. Some of this stuff I’ve developed from reading. Some stuff I’ve developed from my own sites. But this stuff does work and I believe it will continue to work. So even if some of this stuff isn’t perfectly correct right now, following it does work somehow - and I expect will continue to get you ranked in the future. FWIW, the very first site I ranked was about 5 years ago. I built links for 3 months and have not touched it since. It has remained ranking reasonably well for the last 5 years - untouched.
Anyay, at this point I’ll repeat my mantra. You are known by your backlinks.
Pagerank:
Pagerank is a Google algorithm that assigns a number between 0 and 10. A pagerank (PR) of 10 is high.
Pagerank is also an iterative algorithm in that it cycles around in a loop. The more you calculate it, the more accurate it becomes. The actual calculation isn’t overly important though.
The way it works is this. Lets say I somehow have a pagerank 5 web page. If I link to you, I lose a fraction of my pagerank and you gain can some. So maybe my pagerank on that page becomes 4.8 (though keep in mind, pagerank that we see is only ever an integer) and your pagerank becomes a 3.
Your pagerank isn’t linear either. It takes a lot more to go from a pagerank 2 to a 3 than it does to go from 1 to 2. It’s generally assumed that it’s logarithmic, it takes roughly 8 times the links to go from 2 to 3 than it does to go from 1 to 2.
Now lets say you get another backlink from a pagerank 5 web page. Now your page may become a pagerank 4. And so on - the more backlinks you have from higher PR web pages, the higher your PR becomes. And if you link out to someone, you will pass a wee bit of PR to them. (It’s understood that this pagerank bleed does lower your pagerank. However you should assume the bleed is so mild that you really shouldn’t worry about it. Linking out has far bigger concerns and reasons than worring about going from a PR 5.85555 ot a PR 5.85554.
So, pagerank flows through links. The more, higher PR backlinks we have, the higher our PR is. And it’s harder to move up from a PR 5 to a 6 than it is from a 4 to a 5.
Google used to use PR heavily in their ranking. A high PR meant good rankings. I do not believe this is the case anymore - I believe the PR described above is hardly used at all in rankings. In other words, a PR 3 can easily beat a PR 5 site.
What you do use PR for is an indicator of ‘volume’ of backlinks. Higher PR pages and sites generally have more backlinks. Are they good backlinks? Is this a good place for you to get a link? PR does NOT answer those questions anymore.
Relevance:
Relevance of a page is based on two things. Text on the page, and relevance of sites or pages that link to that page.
Now, lets consider something. I believe Google right now is primarily ‘page’ based and not ’site’ based. So getting links from relevant ontopic pages, on an offtopic website works great right now. Will this continue to work? Ideally I think Google would like to use site specific relevance instead of page specific relevance but they run into problems with ‘general’ sites like news sites and the like that have pages on many topics. The reason this is important is if you develop links from ‘pages’ that are important but are basically fudged on a site that really shouldn’t be linking to you - well, that’ll work great right now. It may or may not work great later.
So for relevance, look at the page and make sure it uses words similiar to what’s on your site. Then check the backlinks of the site and make sure they have backlinks from places that are relevant as well.
Trustrank:
This is another conceptual guideline. It may not be exactly what Google does, but approaching this from the right perspective does work now, worked previously, and I expect it to continue to work.
Trustrank is like pagerank. And it works very similiar to pagerank. Trustrank flows from page to page and site to site via links. Get a log of high trustrank links to your pages, your webpages will get a high trustrank.
There are three differences though between TR and PR. First, we can’t see trustrank (pagerank has plenty of little toolbars we can install in our browser to see the PR of any page).
Secondly, rather than being iterative, TR only flows downhill. It’s been speculated (and speculated correctly IMO) that Google dictated some sites as ‘trusted’ and gave them TR. All trustrank flows down from those sites. We do not know what those initial sites are.
Thirdly, unlike PR, I believe Google uses TR heavily in it’s algorithm. Get a high trustrank, and you’ll rank. How do you get a high trustrank? Get lots of links from sites with high trustrank.
Since we can’t see trustrank, how do we measure it? Well, we can’t. All we can do is assume. But that’s OK - it’s very easy to do a gutcheck of a trusted site. Let’s look at some examples. Goverment of Canada site. Trusted? Of course. Highly trusted? Sure is. A university. Trusted? Of course. Not quite as high a TR as the Government of Canada, but probaby not a slouch either. What about a directory site that links out to anyone and everyone for a $15 fee. Trusted? No, not much at all.
In short, just have a look at the site. If it seems highly trusted, assume it is. If it seems shaky, assume it’s low trustrank.
Age:
It’s been assumed that Google has some sort of aging algorithm in it’s backlink checking. Older links are better. This was done as a spam prevention measure. Most web spammers are not willing to wait 6 months or a year to see if backlinks work. So aging cuts back on that. It also places distance between developing links and their affect on your ranking making it more difficult to tell if a specific link helps your site. Six months from now and your site moves up or down - what caused it? Can’t tell if it’s a link or not - and Google wants that obscurity.
Now that you’ve read all that, here’s how you judge backlinks. You want old, trusted, relevant backlinks. When you’re looking at sites to get backlinks from, check their backlinks. If they have old, trusted, relevant backlinks, then that makes them a good link for you to get. Such a link means you know have trusted relevant backlinks that will eventually become old - and you’ll rank.
{ 2 comments… read them below or add one }
So a backlink from a university subpage could have a PR of 0 and a high Trust Rank? Would any site at a utoronto.ca (i.e. something.geography.utoronto.ca) have the same trust rank as the main page (theoretically?)
No, it wouldn’t have the same trustrank as the main page. Remember trustrank flows through links.
The reason a subpage on a university web page would have trust would be because we assume the main page has huge trust, and that there’s a short path between the home page and the sub page.
Let’s say it’s a professor’s page. Then from the home page we go:
main page>departments>geography>faculty and staff>professor gadget>professor gadget’s web page.
So the trust can flow down through the links. That’s a shorter path to a trusted site than say a directory link that may be 30 or 40 links away from anything resembling a trusted page.
Be careful though. People like links from universities because they seem to help - and I think it’s because universities are near the top of the heap in trust rank - they’re very close to whatever sites Google decided were trusted, maybe even initial trusted sites. BUT - if it’s an orphaned page, like a student page, that doesn’t have any links from the actual university site, then my suspicion is that it has little trust. Maybe there’s some inherent trust just being on the domain, but having links pointing at the page makes a large difference I think.
You can check backlinks to a specific page instead of the whole domain using the searchstatus plugin as well, to see what’s linking to a given page. Make sure you use ‘page’ links, not ‘page external only’ so that you also see links from within the university’s site. That’ll tell you if it’s an orphaned, unlinked page.
In terms of the PR0, that really, really wouldn’t concern me. The only problem with a PR0 is that it means there’s pretty much no links to the page. And we want links to the page so that something - PR, TrustRank, whatever, can flow through that page to us. That being said, I really don’t care if the page I’m getting a link from is a PR 1 or a PR4. I look for trust, relevance and age when I’m going after links - pagerank as I noted is just a volume indicator and means the site may have more trust. But low PR does not mean low trust or relevance.
My PR isn’t very high. And I don’t have high PR links - many of the links I work to develop are 1’s and 2’s.
Gotta think of Trustrank as a graph like pagerank - but a seperate graph that’s not related to PR. And it’s the trustrank graph that makes us rank. Low PR pages can be high trustrank pages.
You must log in to post a comment.