Understanding Tracking, Viewability And Discrepancy

Article author
Knorex Support
  • Updated

On Tracking

Counting Impressions

Knorex XPO counts an impression when a request is sent to XPO Ad server to serve the creative and Ad Server return the response with Assets. Note that the impression is counted before Ad begins to load on the user browser or fully downloaded or viewed by the end-user. This is following IAB guidelines   on how to count the impression.

 

Downloaded Impression

We have another metric called downloaded impression which counts an impression each time a creative starts downloading on the user's browser or device. The downloaded impression is counted before the creative is fully downloaded and viewed by the user.

 

Delayed Impressions

Delayed Impression: Knorex XPO will count the impression only after creative is fully downloaded on the user device or browser and rendered it fully. However, this still doesn't guarantee that Impression is viewable or not. The total impression count for the delayed impression is always less than the downloaded impression which is less than the Impression count.

 

Counting Clicks

When a user clicks on a click-through component of a creative, a click request is sent to our Ad Server. AdServer will record this as a click if the request meets the below conditions and then redirect the user to the defined click-through or landing URL. Please note that the Knorex XPO will count the click if it meets the system-defined conditions and not when the user is redirected to the click-through or landing URL.

 

System-defined Conditions

  1. De-duplication Window: If the user has set the deduplication window in Knorex XPO system for clicks, then the system will follow the setting and will discard any click happening in the deduplication window. For example, if a user has set a deduplication window of 1 sec, this means multiple clicks happening from the same user( unique cookie or unique device id) within 1 sec will be discarded by the system. Even if a user clicks ten times during one sec, the system will record it as one click.
  2. Fraudulent Clicks: if the system detects the click request is coming from a fraudulent source based on AI-based proprietary algorithms or detected by our IVT vendors   and Forensiq  , the system will automatically discard those clicks.

 

On Viewability

Viewability is one of the metrics that matters when it comes to Display advertising. In XPO, we support viewability.

  • you can provide your viewability tag (e.g. from MOAT) and it can be embedded into the ad creative
  • if you are unsure if the partner you are using is supported in XPO, connect us with your partner and we can integrate with their system to embed their tag.

 

On Discrepancy

In a campaign execution, typically there might be multiple parties involved and where each uses its ad servers or trackers to track and deploy their respective campaigns. Due to a range of factors (see the factors below), there is always a discrepancy in the resulting reports. A difference of 5-20% is considered standard and normal in the industry, though again the difference might still differ significantly for various reasons (see article on this ref). However, whenever discrepancy exceeds 20%, it is necessary to look into what caused the difference.

 

Contributing Factors to Discrepancy

Each ad impression, click etc. can be tracked in myriad possible ways by different parties and hence are affected by a range of different factors. These are some of the factors that we are aware of that could affect the tracking and reporting as a result.

  • Ad blockers
  • Http and Https (or SSL vs Non-SSL)
  • Fraudulent traffic
  • Differences in measuring and implementation between different systems
  • Layers of systems or "daisy chain" of multiple systems (ref  )
  • Multiple layers of trackers
  • Network latency
  • Websites technologies and
    • AJAX/dynamic sites
    • One-pagers
  • Traffic
    • Non-human traffic - bots, crawlers, and spiders
    • Fraudulent traffic
      • pageview inflation
    • IP restriction
  • Reporting timeframe/timezone

 

FAQs

  1. Why are there differences in measurement when using a Site Analytic software compared with when using an Ad Server in measuring impressions?
    1. This article provides a good explanation https://support.google.com/dfp_premium/answer/177507?hl=en  .

  2. Why are the numbers between XPO reporting different or not matching with the numbers from other reports/systems? Why is there such a discrepancy?
    1. There are multiple reasons for this. Such discrepancies could be due to XPO and your other tracking partner are measuring different things e.g. when XPO delivers click numbers, we apply our own proprietary methods of de-duplication and fraud screening. We follow all industry best practices in the design, operation, and both internal and external monitoring of the click redirect handlers. This discrepancy would also occur between systems e.g. Google DFP vs Google DCM   and Google Analytics vs DFP  . Also, the placements of the trackers would also make a difference. See the above Contributing Factors to Discrepancy.

  3. Why is there a discrepancy between Google Analytics and XPO? 
    1. Google Analytics (GA) is a tool meant for measuring website traffic. It is not meant for counting ads impressions. In the case of measuring "clicks" and "visitors", GA counts “visitors” whereas XPO counts “clicks”. Google counts “visits” using a proprietary algorithm with a 30-minute sliding window. The GA numbers are de-duplicated by IP address over a 30-minute sliding window (repeat clicks in the window are counted once). XPO clicks are counted as they occur, with a 1-second de-duplication window.
      See the same comments from many other vendors
      1. https://www.simpli.fi/blog/adtech-mythbuster-series-discrepancies-google-analytics/  
      2. https://blog.kissmetrics.com/tackling-click-discrepancies/  

Was this article helpful?

0 out of 0 found this helpful

Have more questions? Submit a request

Comments

0 comments

Please sign in to leave a comment.

Didn't find what you were looking for?

Submit a request