An Adotas Q&A with Mike Andrews, Chief Scientist and CTO at Forensiq, explores the impact of real-time technology on ad fraud detection…and more.
A: Real-time detection allows us to catch new bots much faster, before they can do more damage. And that means that there’s increased accuracy overall, because cleansed devices can be served ads sooner.
Looking at the question from another angle, 26.9% of ad-fraud-committing-devices are newly infected, day-to-day. The massive rate of newly infected devices underlines the importance of identifying non-human traffic in real-time.
Q: How will mobile ad fraud be detected with mobile in-app and mobile web?
A: On mobile, fraudsters still have to hit the exchanges and interact with advertisers, which is where we can catch them via traffic patterns. Also, app testing and monitoring can help. App fraud is often perpetrated in all apps from a given publisher, therefore publisher reputation should be monitored. Another thing to look for is device spoofing: automated scripts that hit ad exchanges pretending to be mobile devices. Finally, mobile device emulation does occur, so we must monitor for this scenario as well.
Q: While many fraud companies are focusing on ad fraud, what other related areas should we be aware of?
A: There are advertisers defrauding publishers, too. For instance, they may hijack the publisher’s entire page (site) and redirect elsewhere, like clickless app install redirects. Also, targeting is diluted by bots. That’s why we should be looking for definitive human browsing patterns associated with devices. This, in combination with interest targeting, could be a winning combination.
Q: How do you see the ad ecosystem evolving?
A: We expect to see an increase in ad blocking, along with more pure brand advertising that is not personally targeted.
Q: How does ad blocking affect ad fraud?
A: Ad blocking increases the observed level of fraud, because fraudsters never block. Strategies such as publishers forcing consumers to choose their advertising experience could weed out some fraud.
Q: Do you forecast an increase or decrease in fraud?
A: It all depends on the actions of the players in the ecosystem. But with the Internet of Things, any connected device can be a target. With that in mind, we expect not only an increase, but the emergence of new threats like we’ve seen with mobile device hijacking.
Q: What do you see as the challenges or issues as programmatic advertising becomes more commonly used?
A: The prevalence of programmatic advertising today has already led to increased opportunity for fraud. As programmatic becomes even more utilized, everyone must become ever more vigilant.
About Mike Andrews
Mike has a remarkable background in applying scientific principles and methods to business. He has worked with great success both on Wall Street and most recently in ad tech. Mike’s experience has included directing research and modeling at DoubleVerify where he built their first botnet detection product and at CPXi where he launched botwatch.com. Earlier in Mike’s career he ran an acclaimed research program at Bell Labs and served as a visiting scholar at NYU. The author of dozens of highly cited publications, Mike received his PhD in Experimental Physics from MIT.