Google Track outbound links automatically

Google analytics help center has a great article on how to manually track outbound links, but it's not obtrusive and puts allot of burden on the developer.

So I thought I would write a little script to do it automatically and unobtrusive. I am using the wonderful library mootools, but it's simple enough that you don't need it, and can be rewritten very easily.

modified

Based on remy's suggestion, doing a test for window.location.host instead of el.href.test(/^http:\/\//)

da code

//no tracker then quit
if(!pageTracker) return;

//suggested by remy
var host = window.location.host;

//go thru each link
$$('a').each(function(el)
{
  //if rel tag is google and track if based on what
  //is inside bracker
  //ex: google[download/thisfile]
  if(el.rel && el.rel.test(/^google/i))
  {
    el.addEvent('click',function()
    {
      var track_id = this.rel.match(/^google\[(.*)\]$/);
      if(track_id)
      {
        //we have a match use array[1] for tracking
        pageTracker._trackPageview(track_id[1]);
      }

    });
  }
  
  //if has http, must be external link, track as 
  //outbound/domainname
  //else if(el.href && el.href.test(/^http:\/\//))
  else if(el.host != host)  
  {
    el.addEvent('click',function()
    {
      var track_id = this.href.match(/^http:\/\/([^/]+)/);
      if(track_id)
      {
        //we have a match use array[1] for tracking
        pageTracker._trackPageview('outgoing/'+track_id[1]);
      }

    });
  }
});

code breakdown

Call when document is ready. Go through all "a" tags. Again I am using mootools to grab all links with $$('a'), but you can use document.getElementsByTagName('a') to grab all the hrefs on a page. If we want tracking, based on REL tag, or external link, AND the google tracker (pageTracker) exist, then add onclick dynamically to the link.

First if checks to see there is a rel tag like rel="google[download/thisfile]" then track link as download/thisfile.

Second if checks to see href starts with http://anything that has a different window.location.host, and with a little regex gets the domain and tracks it as outgoing/DOMAIN. note: that for my site, anything that starts with an "http://" is an external link.

source