I used to visit some fortunetellers just for fun. Now it's my turn to do some predictions, but this time it's not about anybody's life. It is about search engines and SEO.

One of my SEO predictions would be about Google's usage of toolbar and analytics for ranking. Google didn't buy Urchin and come up with google analytics just to give up some free service to people. I believe there is a deeper reason for this.

Just imagine, if you have data on how long people stay at your website, and other information such as your entry and exit pages, these could be very good indicators of "relevancy" of your website and individual web pages.

So let's say you created a web page about "red widgets" and out of 1,000 visits to your site, 800 of them visited your "red widgets" page. This could be used as an indicator that your "red widgets" page is relevant for the keyword "red widgets".

Such techniques might be used by search engines in the future to detect and quantify the "relevancy" of a web page. However, the above example is a measure of the "red widgets" in relation to the whole website only. But can the search engines compare this "red widget" web page against other "red widget" web pages?

For that, let us expand this a little further. Assuming there are only 3 websites that has a webpage about "red widgets".

1.red widget webpage in website 1 was visited 9 out of 10 visits to that site, (meaning 1 visited another web page in that same domain).

2. red widget webpage in website 2 was visited 1000 out of 100,000 and

3. red widget webpage in website 3 was visited 800 out of 1000.

If I am a search engine, a good ranking would be

"search results"
1. red widget website 3
2. red widget website 2
3. red widget website 1

Notice that website#2 has 1,000 red widget visitors while website #3 has only 800 visitors. BUT if you look at the probability, website#3 has a probability of 80% while website#2 has only a probability of 1%.

Now you might say that website #1 has a probability of 90% since 9 out of 10 were "red widget" visitors but the number of visitors to their website is so low that it might not be "relevant" enough compared to website #2 and #3.

Google has the processing power and the engineers to develop such algorithm. I don't see any reason why they wouldn't use this in the future as a factor for ranking.

What do you think?