View Single Post
Old 03-17-2010, 06:07 PM  
gideongallery
Confirmed User
 
Join Date: Aug 2003
Posts: 7,082
Quote:
Originally Posted by Jdoughs View Post
You cannot 'statistically sample' 15% of the internet, to tell what sites (in the remaining 85%) should rank.

They do not poll viewers about what shows they DID NOT watch, or HAVE NEVER heard of, and if a TV show isnt on, or isnt available to the viewers, they do not penalize it for not being 'viewed'.

What you are saying, and how you are explaining it in this context, is just not correct.
you sir are a complete moron

we are talking sites that ALL appear in google
so there is no 85% that is not there you moron

we are talking about sampling a portion of the population that goes to google and searches and then returns to google to search again and useing their results to approximate that action for the pages and to determine relevency for the future searches

weather that person goes to the embedded search box on the side or types google.com and then enters the search term again or hits the back button it doesn't matter

that behavior of web users can be statistically sampled for that page and a priority of who keeps the user the longest could be factored into the ranking (highest ranking going to those that keep the searcher longer)


assuming that if 15% bounce back to google 15% will do the same thing if they have a different search engine set in the search box IS a statistically valid comparison.

and ranking that page higher IN GOOGLE is valid too.

like i said if you don't believe it is revelent do the split test yourself, put one 2 minute clip v 10 minute clip. optimize the pages exactly the same, with only the length of the video being the difference and see which one ranks higher.
__________________

“When crimes occur through the mail, you don’t shut the post office down,” Steve Wozniak
gideongallery is offline   Share thread on Digg Share thread on Twitter Share thread on Reddit Share thread on Facebook Reply With Quote