Is Jason's experiment flawed?

Alexa is a site that ranks websites’ popularlity by analysing sites visited by people who download their toolbar. However, its stats have long been thought to be dubious at best.

This morning I see Jason Calacanis is running a little experiment to see how easy it is to game Alexa’s results.

The only thing that confuses me is that he is asking for people’s help in this experiment. Surely the more people who help him, the less it shows that Alexa can be easily gamed – or am I missing something?

9 thoughts on “Is Jason's experiment flawed?”

  1. It is very easy to game from what I’ve seen other people do. The smaller the market too, the easier to game. The more people you have doing this, the better but even a small amount of people, say less than a dozen can get you way up the chart. Want to be number one on Alexa Ireland Tom? 🙂

  2. What is so ridiculous is that the Alexa toolbar only works on IE and as I look at my personal site stats nearly 57% of the traffic is coming from non IE browsers. Yet the traffic indications on Alexa is off to a factor of at least 10.

  3. Yeah, I had the same question. If you can game it with only three machines as he claims, why is he asking for everyone to do this?

  4. I agree Jason’s experiment is invalid. Alexa only counts 1 page view per page, each day, for each IP. I blogged about it at the link above.

  5. I agree with you Tom. Jason tried to contradict his own words. It seemed that he too tried the same trick that others do in order to improve their Alexa rank! Bad it is.

  6. McCAbe: Alexa rank can be important, for example, when selling advertising or maybe a whole website. There are few other ways for the buyer to verify how many visitors a site has.

Comments are closed.