Google Searchpage Analysis

Discussion in 'Google' started by kdb003, Feb 2, 2007.

  1. #1
    I did some analysis on some Google search page factors. I examined things like the length of the title/description of the search result, the number of bold tags in the title/description, the depth of the url(ie google.com would have depth 0 and google.com/adsense would have depth 1), etc...

    You can see the graphs here and my full analysis here.

    If anyone is wondering the queries I used were from the AOL search data that they released. It is the first 800 or so from this file.
     
    kdb003, Feb 2, 2007 IP
  2. kdb003

    kdb003 Active Member

    Messages:
    150
    Likes Received:
    5
    Best Answers:
    0
    Trophy Points:
    73
    #2
    I'll do a little summary of what factors correlated with better results:

    The more bold tags in the title the better the ranking
    The shorter the title/description the better the ranking
    Not have ... in the title/description correlated with better ranking
    The smaller the file size of the page the better the ranking

    The depth of the page and the number of bolds in the description did not show any clear correlation one way or the other.
     
    kdb003, Feb 3, 2007 IP
  3. Vigor

    Vigor Peon

    Messages:
    693
    Likes Received:
    12
    Best Answers:
    0
    Trophy Points:
    0
    #3
    Interesting. Thanks for sharing it. However i cant see it as conclusive as one factor is compared to a multitude of variables that make up the ranking. You just dont know which variable is causing the rank. Never the less, some interesting trends.
     
    Vigor, Feb 3, 2007 IP
  4. kdb003

    kdb003 Active Member

    Messages:
    150
    Likes Received:
    5
    Best Answers:
    0
    Trophy Points:
    73
    #4
    I am a little confused as to what you are saying. Rank is obviously derived from many different factors. These statistics help determine some things which correlate well with rank. While correlation does not necessarily imply causation, it might and it is worth experimenting with further.

    I am not sure if it is well accepted or not that click through rate on search results affect the results, but these statistics provide a strong argument for that.
     
    kdb003, Feb 5, 2007 IP
  5. thegypsy

    thegypsy Peon

    Messages:
    1,348
    Likes Received:
    109
    Best Answers:
    0
    Trophy Points:
    0
    #5
    DooD I see very little that couldn't be evaluated a zillion other ways.
    I would also say it's some strange choices in test matter....

    You can't figure out what 'competitive terms' are, and U are doing SEO experiments? Back to the drawing board ol chum

    Have you considered studying search engineering?
     
    thegypsy, Feb 5, 2007 IP