Experiments

  1. Bit.ly test has been stopped

    January 6th, 2010
    By Rob Hughes

    A few months ago there was a tweet (and I believe a blog post) by Dave Naylor on the exploitation or URL shorteners in general, specifically bit.ly (as that was the example he used).

    On seeing this I decided to sit down and see what I could find out about further ways to ‘influence’ these URL shortners to send me more traffic and to possibly impact on crawl times and as such rankings.

    On 24th September 2009 I decided on how I should do this and implemented it straight away.

    I’m not going to give exact details out there of what I did because there are probably URL shorteners out there that are allowing this, it would be silly for me to give away this note without trying to test it a bit further :)

    (In fact I know there are URL shortners out there that still allow this because I just found one, but that’s a story for another day).

    The graph below shows the instant clicks that I received from testing this (without mentioning this test anyway, not even verbally):

    bitl.y traffic exploit stopped

    Unfortunately, as can be seen on the 23rd December 2009 all this traffic instantaneously stopped!

    Not ideal as at this point I wasn’t checking it as much as I should have done and had no idea that this had happened and as such wasn’t able to do an instant check of my logs and traffic to see if any impact was made (hence why I’m not in the best of moods today because I have just ruined a lot of the benefits of the test)..

    In case you think that this test was more of a “blackhat SEO exploitation” rather than an actual test – it wasn’t – what more is there to say about it. It was purely a test to see if increased traffic and crawlability would affect my rankings in one way or another.

    From a quick (very unscientific) look at my statistics, etc. I would have to say that this has worked in the way it was supposed to, but, without a more in detail look, which I’ll be doing at some point this week, it’s hard to come up with any hard evidence.

    I don’t normally want to share methods I’ve used to try and test things like this with the public because, as everybody knows, anything you find and then share reduces your competitive advantage… in this case though, I’m not really giving away any details that people can’t work out for themselves – I just wanted to post something up that gave an example of some of the tests that I am running in the background.

    **EDIT**
    I decided to go back and search for the exact posts on Dave Naylors blog so that I could add in the links that he deserves :)

  2. Ranking Algorithm (Introduction)

    December 3rd, 2009
    By Rob Hughes

    So recently I’ve been doing a lot of work with link location networks at Summit Media, and from this I’ve been having a look at information I can get quickly and what effects it has on rankings – obviously we already have a scoring factor that we use for tasks like this but I’ve decided to start looking at this in a lot more detail.

    I’ve come up with a list of around 50 factors that I think I can get information for relatively quickly and easily which I will be starting my study with.

    I’ve started small and picked a high volume search term and are basing my initial analysis on just the top 20 results for this one term.
    Clearly there are a lot of issues with this but I think as an initial study to get some sort of basic information before I take things a step further this should work quite well.

    Not really got anything to share majorly at the minute, just thought I’d put up an introduction of the new project I’ve set myself…

    Here are a couple of example graphs that show some of the very top level analysis that I’m starting with :)

    Graph of relationship between Page Title Length and Google Rank

    Graph of relationship between Page Title Length and Google Rank

    Graph of relationship between Meta Description Length and Google Rank

    Graph of relationship between Meta Description Length and Google Rank

    Clearly both of these have pretty poor relationships, not necessarily a bad thing – all information is good :)

links

Other

blogroll

social