Determining Another Site's Traffic Measurements?

Question

Hi.

I have a conceptual question.

I am wondering how companies such as Alexa Internet determine a given site's (not my own) overall traffic and traffic for each unique page. I would appreciate a technical response - if you were to design this feature (i am sure it is complicated but hypothetically...) how would you go about it?

Thanks in advance.

Answer 1

One way is to be hooked into one or more core routers. From there you could perform deep packet inspection to see where traffic is going, what pages are visited, etc.

Another way is to have people install a browser toolbar which records where they go and submits that information back to you. I think this is how Alexa works.

A third way is to have web site owners install a bit of javascript which performs analytics and submits that data back to you. This is how Google does it.

A fourth way is to buy that data from companies that do one of the above.

Answer 2

Alexa estimates website traffic by extrapolating the data from the browsing sessions of the subset of the Internet population who use the Alexa toolbar or browser extensions. This isn't a truly random sample, so questions are raised over the accuracy of such data: http://en.wikipedia.org/wiki/Alexa%5FInternet#Accuracy%5Fof%5Franking%5Fby%5Fthe%5FAlexa%5FToolbar

Installing the Alexa toolbar modifies the browser user-agent, so you can estimate the % of visitors to your site who are contributing data to Alexa by scanning your server logs for requests with the appropriate user-agent strings.