Forum Moderators: DixonJones
1. Using referral data from the browser. You can find these in your log files as long as your server is tracking full log data. This is unreliable as "back buttons" can create falsely high numbers and browsers not volunteering the referral data can create falsely low numbers. It does, however, have the advantage of not needing anyone to do anything.
2. using tracking URLs. This is the most popular I think. You ask every link to add "?referrer=theirID" onto theit link and you then track the landing page urls (the urls are different because "theirID" is unique to the referrer). This is quite accurate because if a person clicks on the link, then around the site and then reloads the page, the page will reload without the "theirID" extension, so the only real error is if a person clicks from the referring site to yours several times. Most tracking packages will be able to identify these occurances as "the same visitort session2 although every package is different in this regard. The disadvantage of this is the POTENTIAL for search engines to treat the various urls as dupe content. The worst case I had of this was when the referring (tracking) url was listed in Yahoo at the expense of the page without the tracking url - so anyone coming from Yahoo got tracked as a referrer from somewhere else. That is rare though I think.
3. Redirecting through a database. Sounds like this is similar to what you have now. Most affiliate programs track this way. If you want search engine benefits, then you need the redirect to deliver what's called a 301 message, then Google et al (in theory) works it out. You'll need technical help from the redirection server to make this work, but look up something like "server response header checkers" to find a tool that will be able to tell you if your tracking links are currently giving a 301 to the main landing page.