Admetrics, a technology provider for digital media tracking and user tracking, now also offers a solution to counter digital ad fraud.
In this interview, Admetrics founder Markus Repetschnig answers the most frequently asked questions on Fraud Prevention.
How did Fraud Prevention come about? Can you tell us a little bit about the background?
Sure can do. In our day-to-day consultations, a number of clients showed interest in a solution that can measure and eliminate fraudulent inventory, possibly because digital ad fraud has been such a prominent topic recently.
We thought that the ability to programmatically detect fraudulent traffic patterns would be a logical addition to our range of ad tracking, verification and analytics solutions. So we put a dedicated team on the job, reverse-engineered bots, bought and analyzed bot-net traffic and invested a lot of time in training our probablistic and deterministic models.
Eventually, we began rolling out in Q2 2016 and we're very happy with the results of Fraud Prevention's pattern detection, especially when combined with Device ID, Admetrics' cookieless device tracking solution.
What did you see in terms of results?
Fraud Prevention's results were evaluated against the results of two other leading technology providers in a real-world test run on a number of live campaigns. We basically saw similar results across the board. With values in the low single digits, the amount of fraudulent impressions were not as alarming as the worst cases you hear about in industry discussions. However, that wasn't exactly a surprise since the tested campaigns ran on high quality inventory. We find it hard to talk about averages or benchmarks in general, since so much depends on the way a specific advertiser runs its campaigns, where the traffic is bought from, and so on.
That being said, the amount of fraud that you see in general is without a doubt significant enough to optimize against. We definitely would advise every advertiser to do so.
What would be common use cases for Fraud Prevention?
Programmatic. The most common request is to leverage Fraud Prevention data in an RTB stack.
There's basically two ways to do this: Since we can qualify impressions in real-time, one option would be to use the data in pre-bid environments. The other option is making post-impression data available to bidding algorithms, for example to blacklist domains with a high probability for fraudulent traffic.
In terms of other use cases, I would suggest to get in touch and evaluate the specific situation and environment to see how Fraud Prevention can contribute.
I guess that's it. Thank you very much!
My pleasure. Be well.
Learn more about Admetrics Fraud Prevention on its dedicated product page.