1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

How does google know if an article make sense?

Discussion in 'Black Hat SEO' started by IMShane, Oct 8, 2011.

  1. IMShane

    IMShane Junior Member

    Joined:
    Sep 20, 2011
    Messages:
    131
    Likes Received:
    23
    We all know that a site/page needs ton of backlinks from unique content to rank higher in google SERP.

    Using spinning tools we can automatically generate lots of "unique" spun articles from the original one, especially when you remove the original words / restructure the paragraphs / remove several sentences etc. But as we all know, these "unique" articles hardly make sense if it's generated automatically.

    So my question is, will backlinks from these articles help in rankings? Is google smart enough to determine if an article make sense? How does google know if one article is of good quality?
     
  2. Kastez

    Kastez Jr. VIP Jr. VIP Premium Member

    Joined:
    Sep 16, 2009
    Messages:
    1,044
    Likes Received:
    164
    Location:
    Everywhere
    - So my question is, will backlinks from these articles help in rankings?
    - yes
    - Is google smart enough to determine if an article make sense?
    - still nope, but people with browser can help google see quality of your content.
    - How does google know if one article is of good quality?
    - thousands factors.
     
  3. NapsteR

    NapsteR Jr. VIP Jr. VIP

    Joined:
    Mar 2, 2011
    Messages:
    2,779
    Likes Received:
    2,374
    Occupation:
    Full Time IMer
    Location:
    http://www.seophd.com
  4. keinehabe

    keinehabe Supreme Member

    Joined:
    Nov 4, 2008
    Messages:
    1,207
    Likes Received:
    472
    Gender:
    Male
    Occupation:
    -= CEO =-
    Location:
    Heaven
    Home Page:
    that's exactly what's panda lol , they virtually '' learn '' what's good content and what's machine generated , it's not about how good is the content , they are measuring loads of other factors ( time spend by peoples on various pages / clicks / where peoples go after they saw one page ( hit back / click on external links / click on internal links ... and so on )

    my two cents goes like this :
    no matter what content do you have actually , ranks are given by the backlinks ( and I have some serious doubts about this ) I even lost a bunch of domains testing this crap ... if you have some crap content spun/ copied / rewrote ... and you are outranking a trust domain for some of their natural keywords simple result is they will take off from index the bad content leaving the trusted domains up even with worst or no damn links at all :D ...

    other scenario is like this : following the circles of links , actually any dumb robot can figure a pattern for any given site/keyword .... google is enough smart no doubt about this and they virtually learning how the links come and go for specific page/target ....
    finding the pattern used by any of us is easy like hell to trace everything and to extract '' my / yours / hes '' content then just compare with other(s) similar one ... hope makes sense what I wanted to say ....

    short long story , the machines can't '' read '' & '' learn '' like humans but they can be trained to act like ones if specific set of instructions are given :)


    actually this is pretty easy to accomplish , just use the thesaurus of any language . MS have built in the correct forms of spelling words , and it's just comparing them , even MS can be made to '' learn '' new words and new spellings ( check out write any misspelled word , right click and hit update / ignore ... and so on play with options ) ....
     
    • Thanks Thanks x 1
    Last edited: Oct 8, 2011
  5. pcboy

    pcboy Registered Member

    Joined:
    Feb 20, 2009
    Messages:
    60
    Likes Received:
    53
    Location:
    Largest democracy
    bounce rate and google has outsourced its search results to be evaluated by humans.

    For every search query, google gives its raters 10 results on two windows, raters have to rate which side the results are better, which sites are authority sites and spammy ones.(They might use this data to update algorithm as well).. hope this helps
     
  6. madoctopus

    madoctopus Supreme Member

    Joined:
    Apr 4, 2010
    Messages:
    1,249
    Likes Received:
    3,498
    Occupation:
    Full time IM
    if you understand how the google index is built and what google can use to figure out stuff and what kind of stuff it can figure out you can build horrible content that gets indexed and looks ok to google. Explaining all that is a bit over my mood at this moment but I can tell you it has to do with n-grams and possibly correct grammar. It might also use other stuff I don't expect it to, but I was able to get even complete garbage indexed and passing link juice. If you're ok going full blackhat then fine but if a human reads that crap and that human is a competitor or from google you are screwed.
     
  7. spasovski

    spasovski Regular Member

    Joined:
    Mar 21, 2011
    Messages:
    394
    Likes Received:
    240
    Occupation:
    Web designer and Internet marketeer.
    Well, that's why algorithms are created developed.
    The new Panda update is capable of learning and self developing further more in content quality detection including many other factors that help your ranking.

    If I created an algorithm for google to recognize spun content, I would simply use reverse engineering. When you spin content, you use synonyms, so the same meaning is in the same line of text but the words are different using synonyms. So the algorithm would check all the synonyms in that line and it will save all versions of that same meaning. That database for the algorithm would be TBS, SpinnerChief's and other spinners thesaurus files :), which leaves it with the easy part just the check the web for other sites using the same meaning. It will definitely use more factors than only checking the synonyms but this is only a simple explanation of how that could work.
     
  8. humvee9

    humvee9 Power Member

    Joined:
    Jul 1, 2010
    Messages:
    703
    Likes Received:
    220
    I guess the've diffrent definitions of "making sense" for article type, and several hundred
    parameters based on which the article is decided to what level of sense it makes to
    a human reader. Super Complex Algos, Google Bots and manual review is the least i can think of of now.. gotta get ajob at google to understand How it works :p