Preface: My employer does a lot with marketing campaigns, and a regular request from our clients is long-term ROI. Google analytics has been great, and Google Adwords/Conversion tracking has been decent. It's difficult to explain to a client, though, that they are unable to track conversions that occur for more than 30 days. If a user comes directly from our marketing campaign, we want to see how much that user spent over the lifetime of their account, not just within 30 days.

What I'm considering is something similar to what google provides, but housed on our local servers and only tracking what we tell it to track.

Here is the basic outline of what I plan on developing:

-All of our Adwords marketing campaigns direct users to a landing page. When a user clicks a link on the landing page, a cookie will be stored on the user's computer with a unique ID associated with the campaign.

-Certain pages (depending on what we want to track) on the client's website will contain a block of code which will execute a javascript file located on one of our servers. Each client will have a unique javascript file based on what we are tracking.

-This javascript file will check certain conditions (Does the cookie I'm looking for exist? Is the user logged in? What page is the user on?), and if satisfied, will send certain information to our local server for storage.

-An example of what we might send would be the datetime of the transaction, the userID of the person logged in, the total of a purchase, etc. We would only want very basic information.

-Using the basic information saved to our own database, we can build a query for the website's database (which is sometimes housed by us, other times by the client) to generate more detailed and long-term reports related to the campaign.

-For example: We could determine the number of users which came from Campaign A who are still involved with the website one year after Campaign A ended. We could simply collect all UserIDs associated with the campaign from our database and query the website's production database for their last activity.We could also determine how much revenue the campaign had generated by collecting all users from Campaign A from our database and query the website's production database to total all sales from these users between the start of Campaign A and now.