If you don't find an answer, please click here to post your question.

Discrepancies Between Ooyala Pulse and 3rd Party Data

0 Kudos
Started ‎03-24-2015 by
Modified ‎01-29-2018 by

Discrepancies Between Ooyala Pulse and 3rd Party Data

by Community Manager on ‎03-24-2015 11:21 AM - edited on ‎01-29-2018 01:45 AM by Ooyalan (601 Views)
  • Google Analytics in click/landing page tracking

When you try to use Google Analytics (GA) for this, you run into problems almost every time. The reason is that GA uses javascript to call a property named referrer, which normally contains the URL of the page that the user came from. The problem with this is that referrer is quite buggy, especially when the last click came from within a browser plugin (like Flash), and many times doesn't contain anything at all. So instead of referrer = "http://www.ooyala.com", for example, referrer = null would be the case. Because of this, sites that use GA to track where users came from, have a high discrepancy.

 

The solution is to use a redirect tracker instead, meaning use a link like http://tracker.com?site=one&redirect=http://www.one.com that first tracks the click-thru, and then forwards the user to the real URL.

 

  • No cache buster/random token

If you use URLs that always look exactly the same, most browsers will cache the response, and the next time the link is called, they won't bother to call it as they have it stored in their memory already. This leads to discrepancies especially with 3rd party impression trackers.

 

The solution for this is to always add cache buster (random tokens) to the impression URLs. Example: http://tracker.com?track=campaign1 becomes http://tracker.com?track=campaign1&rnd=[random]

Due to the nature of standard internet protocols, discrepancies can also occur when multiple servers are used to count statistics in a sequential order.

 

  • Sequential Latency

The initial request usually comes first to the third party server and only then the Adform pixel is fired. Very likely, there can be a difference of 0.2-0.8 seconds, depending on user’s connectivity and bandwidth to both Adform and the third party system. During this period, discrepancies appear as some connections may be cut off either due to user’s actions (such as closed browser’s window, interrupted page load by loading other webpages, or clicking “back”/“forward”, etc.) or lost internet connection.

 

  • Slower than usual page loading

It can depend on internet related problems or any connection problems with the item on the page (e.g. page elements from external sources, such as images, scripts, etc.).

 

  • User behavior on page

Depending on website’s specifics, a user might be able to navigate from the initial page to the page they are interested in without waiting for the whole content to be loaded. Closed browser tabs or windows also fall under this reason of discrepancies.

 

  • Differences in counting methodologies

Adform counts an impression as soon as a request to Adform servers is made. Third party ad server usually counts a tag load before providing the content and before a request to Adform is made. Other than that, some requests depending on user-agent string or IP might be classified as robotic and withdrawn from statistics.

 

  •  Ad blockers

Various ad blockers may block different ad servers. As a result, third party server may not be blocked while Adform is blocked. In such cases, Adform will never have a chance to count an impression.