Managing and Minimizing Campaign Tracking Discrepancies


girlglass.jpgA common trend has emerged in recent years among campaign tracking platforms where the number of successful actions reported by the Network is typically lower than Advertiser data. We all know this difference as the “discrepancy,” and for many Networks it has become a recurring annoyance. And while it seems as if this is something we are stuck with, there are ways to manage and minimize the problem.

Large discrepancies result in unnecessary administrative complications and other disadvantages for a Network. I see it all the time – one offer is being run using distinctly different tracking platforms, with split testing resulting in one of the Networks reporting more successful actions than the other, resulting in larger revenues for the Publisher. When this happens the Publisher has an easy decision as to which Network through which they’ll continue to run the offer. Additionally, Networks that choose to update statistics to show unreported leads may spend several hours and waste valuable resources doing this data importation.

So how do we eliminate these tracking discrepancies?

I should first point out that I don’t think there is any way to successfully track campaigns with 100% accuracy, as there are too many variables that can skew data and thus disrupt the tracking process. However, I do think there are methods available to close the gap and report data with improved accuracy.

Don’t Rely On Browser Cookies

Most tracking platforms use persistent browser cookies and pixels to track conversions. However, the majority of platforms aren’t prepared for customers who have cookies disabled on their computer, which results in the customer not being tracked through the conversion process. So what do we do if a customer has cookies disabled? In Commission Junction’s Publisher FAQ they address cookie tracking accordingly:

Q: What about customers who have their cookies turned off?

A: Over 99% of all Internet users have cookies enabled in their browsers (less than 1% of all users have their cookies disabled). The fact is that cookie-based technology is a reality that will remain a part of consumer tracking for individual Internet businesses. In the highly unlikely situation that a customer elects to have cookies disabled, the customer is out of the range of publisher tracking and sales cannot be tracked.

I fully agree with CJ that cookie-based tracking is the best overall method for tracking customers. Cookies allow easy identification of unique customers, and they can help prevent fraudulent activity. However, I strongly disagree with CJ’s position that “the customer is out of the range of tracking” when they have cookies disabled.

I should point out that although less than 1% of users have cookies disabled, an unknown percentage of cookies are being removed or deleted in other ways:

  • Customers manually delete browser cookies;
  • Cookies often expire based on time constraints;
  • Cookies are regularly removed by anti-virus and anti-spy software.

We can only speculate as to the adjusted percentage of cookies that are actually unavailable by the time a conversion occurs. It seems more than feasible that the 1% of customers with cookies disabled, combined with those removed through other means, could inflate the percentage closer to the three to ten percent level that Networks are currently experiencing.

By logging specific user characteristics that do not require browser cookies, we may be able to more accurately identify customers and better track conversions. Although cookies may not be available, we still have access to other information that can be evaluated at the time of conversion. These include:

  • IP address
  • User agent (browser)
  • Operating system
  • Screen resolution
  • Language

If we log all of these characteristics at the moment a click occurs, we may be able to identify customers at the time of conversion (when the pixel fires) and eliminate a portion of the conversion discrepancy. I have yet to hear of anyone using this type of tactical measurement approach, and it may open up the possibility of reporting false-positives if not properly executed. But it merits closer examination and testing.


Please enter your comment!
Please enter your name here