kutilo

There is an old .PDF file that contains a slide show from a presentation by Sergey Brin showing how Google used to work. I say "used to work" because we know that Big Daddy changed some things, we have reason to believe that Big Daddy may have added some things, and we have no way of knowing exactly how today's Google differs from the Google described in this Circa 2000 lecture.

web.cs.wpi.edu/~cs4241/2000d/class11.pdf

Some disreputable SEO gurus say things like "the most powerful [of ranking factors for Google] is link text". It takes a real idiot to something as stupid as that.

There is no way to quantify how much any one factor influences rankings for specific queries. One of the most commonly cited examples of "the power of link text" is the infamous miserable failure, where pages clearly not intended to be ranked for that expression have been sneak attacked by third parties who link-bombed George Bush, Jimmy Carter, Michael Moore, and others.

That's like sneaking up on your sleeping mother-in-law, pouring a bucket of paint on her head, and then telling your wife, "Honey, your mom just does not know how to put on make-up."

This type of incompetent analysis of search results is one of the chief reasons why people so often scratch their heads over what happens at Google. The incompetence is further burdened by mistaking how Google finds documents in its various index structures with how it ranks the documents it finds.

The finding happens first. The ranking happens second.

Even the new Google is still doing it that way.

So, when you type a query into Google, you can know that Google assembles a collection of documents which match your query according to pre-determined criteria. That collection may be drawn from what Google calls the "short barrel" (sometimes called the "short index") -- a document list comprised of data drawn from page titles and anchor text. The collection may be drawn from what Google calls the "full barrel", which consists of all the data they have about a document. Or the collection may be drawn from both.

As far as the short barrel is concerned, you can control both factors: page title and anchor text. You can create as many links to your own pages as you want. They can be on your pages or not on your pages. It doesn't matter. I managed to get into the top five positions on a popular query that was related to another query I had optimized by changing the anchor text for ten links on my own pages. All that did was create a new entry in Google's "short barrel" for the popular query. It didn't increase my relevance score or PageRank.

Merely getting a page selected in the up to 1,000 listings that will be sorted and presented to the user is important. But once you're there, other factors take over. That is, if all you're relying on is link anchor text to prove your relevance, you have to have an awful lot of inbound links to hit the top. Otherwise, you're stuck in the long tail of the sorted results.

Any idiot can take Google's technical information out of context. Several idiots do that on regular occasion. But when it comes to ranking well on Google, getting into the selected 1,000 listings is only the first step. After that, you have to compete with all the factors in play.

Selecting query results is not the same as sorting them. There is no indication in any of the technical literature that being found in the "short barrel" gives you any sort of relevance boost. It only means you have a better chance of being included in the 1,000 results.

One of the most important aspects to Google's relevance algorithm is that they weight each "hit" (each factor) by the number of occurrences. This is why simple repetition of keywords on pages works so well. This is why getting tons of inbound links works so well. This is why doing lots of anything over and over works so well.

You can bludgeon your way to the top in more than one way. Some factors, such as URLs and page titles, can and do have limits. How many times can you repeat a keyword in your URL without tripping a spam filter? We don't know. But maybe apple-apple-apple.com works slightly better than apple.com because you get a weighted 3-count with the first domain versus a weighted 1-count with the second.


Well, I'll come back to this when I get more time.
_________________
Grand Canyon Hotels austin apartment