Discrepancies tend to be a headache for all publishers. Through this practical case we will show you how we've helped one of our customers solve a discrepancy issue and achieve a 225K€ yearly uplift in ad revenue.
One of our customer’s best bidders generates 250K€ annual ad revenues per year.
Thanks to Pubstack, this publisher implemented a discrepancy analysis workflow to compare his prebid performance against each bidder’s report.
The analysis shows that there is a 15% revenue delta between the bidder’s reporting and what Pubstack calculates.
This allows us to easily conclude that the bidder bids in gross amounts instead of bidding in net amounts just like all the other bidders. Which gives him a clear and unfair competitive advantage over the other bidders. He wins impressions that he shouldn’t be winning.
After checking in the wrapper setup and a discussion with the bidder’s account manager, the publisher becomes sure that the discrepancy comes from Net vs. Gross bids.
A bid adjustment module is set. In this case, the bid adjustment module will scale down the bid before it enters the auction to make sure the bidder’s revenue share has been removed.
15% of the impressions that were unfairly won by the bidder could have been won by another bidder with a better bid. Which represents more around 5% uplift on annual programmatic revenues.
⇒Publisher’s annual programmatic revenues: 4.500.000€
⇒Annual programmatic uplift= 4500000 x 5% = 225.000€
And this just one optimization out of several others Pubstack allows you to achieve.
Naturally, this resolution requires proper metric tracking and superior data granularity to allow for deep-dive troubleshooting to fix your ad stack. One of our most recent customers reported his use of our Analytics module, which you can find here : Use case : Highfivve.
Another one of our customers that found great value in our “slice and dice” capacity and highly granular data was The Moneytizer. Despite having a different situation to that of Highfivve, Pubstack’s solution has helped them reach optimal yield while reducing the guess work. You can read their use case here : Use case : The Moneytizer.
