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

[Experiment] Can Google detect quality of article?

Discussion in 'Black Hat SEO' started by satyawrat, Apr 12, 2010.

  1. satyawrat

    satyawrat Jr. VIP Jr. VIP

    Joined:
    Jul 8, 2009
    Messages:
    926
    Likes Received:
    1,182
    Occupation:
    Hustler
    Location:
    Gurgaon
    Home Page:
    This experiment is aimed to find out whether Google is able to detect the quality of an article.
    Am using {|} method to spin articles.
    Here is what I am doing:
    I wrote an article of 300 words. Which had my main keyword in it 5 times, all the 5 times it was linking to my main site. [My main site ranks #10 for this keyword]

    Then I created spinner article, but with a twist. Only those sentences which had my keyword in them got quality rewrites.

    I am using Speech Recognition program for doing the articles. I created 10 sentence variation for those 5 sentences and for the rest i did 5 variations.

    Now where does the quality go down?
    Except those 5 sentences, the variations are nothing but me rapping into the speech recognition software. hehehehe....

    Now the speech recognition software isn't capable of understanding rap. So here is a demo of the spinned article:
    I will make atleast 16 articles which are 60% unqiue from each other and then blast them with SeNUke....
    Lets see how it goes... will keep you updated.
     
  2. Talsin

    Talsin Regular Member

    Joined:
    Mar 18, 2010
    Messages:
    385
    Likes Received:
    107
    Hmm interesting test.

    I would guess that they cant, the way they weight a given page is how they decide its quality surely, so they rely on other people to say whether or not its worthwhile.

    But the trouble with making spammy rubbish is that you'll only get whatever you put around it in terms of linking rather than getting some possible organic linking (yea right haha I know most spun shit is never gonna get any organic links), but also you risk it not lasting long term because of human moderators coming across it (on both the sites your hosting and from google - though no idea how good team google asia editors are).
     
  3. aftershock2020

    aftershock2020 Senior Member

    Joined:
    Oct 19, 2007
    Messages:
    981
    Likes Received:
    477
    Yes and No...

    No, they don't read and quality control your articles physically.

    However, they do use algorythms to search for content related words, phrases, associated terms of the given target niche you or the author posted the content under and which keywords you/they may have attached to it when submitting it to the source directory or posting it on the page within your article archive.

    It hasn't changed since the late 90's in the way social media works. IF your site is about dogs and you have a bunch of crappy articles on there, that talk about ' Dog, dog, dog, dog ', you are going to be in trouble, as duplicate content isn't an issue if you past an article from someone else's site/source but keyword stuffing still very much is.

    Google and every other top SE scans for the follow elements from your articles:

    Keywords, title relevance to kw, body content relevance to kw ( associated keywords and terms/phrases in relation. Example: Keyword - Italian food. Related words - spagetti, pasta, fork, sauce, etc. ), link relevance to kw, TOPIC relevance to kw ( Yes, where you place the article in the directory has a bearing on its success in quality and ranking. ).

    Hope this helps.
     
  4. Talsin

    Talsin Regular Member

    Joined:
    Mar 18, 2010
    Messages:
    385
    Likes Received:
    107
    The above is true but that doesnt mean you cant use related words to still pretty much just spit out garbage that doesnt mean anything to a human, similar to the old keyword stuffing (lets call it LSI stuffing with a coating of spam) and maybe get away with it.

    I cant see it being a valid long term strategy though, and the risks of being slapped for it must be reasonably high.
     
  5. aftershock2020

    aftershock2020 Senior Member

    Joined:
    Oct 19, 2007
    Messages:
    981
    Likes Received:
    477
    This is true but the thing is that the article directories that you are posting to are social sites, which meansthey keep their readership, traffics and fan following by the quality of their sites, which extends to google. You won't get crap articles accepted due to it being a waste of time in dealng with the top directories. Ezine@articles will tell you that they openly and deliberately use human monitors on their site to read the articles.

    There are some that don't read properly and allow things to slip through the cracks but due to the media moving more to that ' human ' element to further detour and prevent article spinning all together.

    Something to be aware of...

    1. If you can't get the article submitted in the directories, it doesn't do you any good.

    2. Once someone wastes their time reading a crappy spin, your account is going to be rated the lowest possible, if not reported to the admin of the source site to be removed for spinning/spamming. This is a real and serious matter these days due to sites wanting to protect their reputations and good standings as authority sites.

    Food for thought, even though it is a little off/side topic.
     
  6. socksiong

    socksiong Newbie

    Joined:
    May 10, 2009
    Messages:
    13
    Likes Received:
    0
    quite interesting. but how many competitors does your keyword phrase have? it might be good if we have 2 extreme test environments. 1 with less than 500 competitors and another with a few million.
     
  7. Disposable

    Disposable Registered Member

    Joined:
    Apr 15, 2009
    Messages:
    64
    Likes Received:
    60
    There are a number of problems with the design of your experiment. I do this stuff for a living, so please understand that my comments are trying to be helpful to you. I love experiments and hate to see enthusiasm and effort wasted in getting results that are not actionable.

    1) Your sample size is too small. 16 articles for a single keyword is not enough. You need more articles spread across more keywords to get any meaningful results.

    2) The measurement is unclear. Are you planning to measure the ranking of the control article versus the spun versions as a measure of "quality"? The problem is that it takes varying amounts of time for articles to be indexed and ranked. In addition, there are other factors that affect ranking beyond "quality" if indeed Google is able to determine quality at all.

    3) Ranking is determined by several factors. I agree with others that some LSI keywords must be included in the gibberish text of some articles to allow Google to better understand what the article is about and what keyword(s) it should be ranked for. In addition, you need to control for the power of each article directory that you use to get indexed and ranked.

    4) How will the results be used? Let's say hypothetically that the test shows that quality does not matter for ranking purposes. Each place that this crap content is posted has some mechanism for getting is labeled as spam whether article directories, splogs, etc. The only use that I can see is as disposable links that appear in the index for a short while until they are inevitably deleted. This is of limited use.

    These are just some quick thoughts about your experiment. There are certainly other issues that have not been discussed yet. Ultimately I think that the results of your experiment will spark a lot of discussion/debate, but do little to provide the concrete answers that you seek.

    Regardless of the outcome, keep testing . . .
     
  8. Megalomaniac Midget

    Megalomaniac Midget Power Member Premium Member

    Joined:
    Oct 1, 2009
    Messages:
    688
    Likes Received:
    1,063
    Occupation:
    Bullshit Artist
    Home Page:
    Instead of fucking around with half-assed experiments i would suggest that if your serious about SEO is to start reading the numerous patents that google have filed, plus the numerous academic papers. This question is old news.

    The answer to your question is yes, they can detect the quality. The exact mix of the secret sauce noone knows, but they are using the flesch-kincaid readability test, Flesch-Kincaid Grade Level & Automated Readability Index.

    run your spinned content thru this for shits & giggles
    Code:
    http://www.online-utility.org/english/readability_test_and_improve.jsp
    If your scaling up for world domination heres some code you can hook into other systems
    Code:
    http://code.google.com/p/php-text-statistics/
     
    • Thanks Thanks x 2