A probing Q&A with Dwight Ringdahl, CTO of RhythmOne.
A: Put simply, they are all retroactive. Most viewability and verification technology is forensic, meaning that it can only confirm when an ad was viewed after it has rendered. Don’t get me wrong – that’s a huge improvement over previous years and has inaugurated a new paradigm of how ads are billed. But non-viewable ads still render all the time. That’s the limitation of today’s prevailing methodology, and IAB and MRC standards reflect those constraints. But the real, structural, industry-wide transition to a viewable standard will only come when we can preempt non-viewable inventory, stopping it before it enters the marketplace and skip the make-good process altogether.
Q: You are an engineer who has worked on programmatic since its infancy. What problems are you concerned about as an engineer that other programmatic stakeholders might not be aware of?
A: Well, ad fraud keeps me up at night just like everyone else. Among the many types, bot fraud and domain fraud contribute most to the loss of sleep. Bot fraud is well understood, and widely feared – but I am spending more and more time dealing with domain fraud, which is on the rise and more prevalent than is generally perceived. Domain fraud, variously known as URL masking or impression laundering, is when fake “front sites” charge impressions by disguising themselves as legitimate publisher URLs. As a result, they get premium prices for junk inventory running on illegitimate sites. This strategy also means that domain fraudsters are incentivized to target digital video buys where CPMs are that much higher, and my teams are definitely seeing this play out. Domain fraud is already difficult to fight because the illegitimate site can be hard-coded into the ad tag. Blacklists work but can only do so much. What’s worse is these schemes are conducted in the interstitial grey area between demand and supply sides, and the responsibility for mitigating against it does not fall squarely with either the DSP or the SSP. That makes them even more intractable.
Q: How do you see our industry’s approach to fraud prevention evolving and what might hold it back?
Truth is, fraud prevention will always be an arms race, a never-ending a game of whack-a-mole. The most we can do in the long run is remove the incentive to easy money by making it the process of ad fraud more difficult and costly to perform. It’s important that we view this as an ongoing development and not rely on the prospect of a single solution.
Q: Coming from the tech side, do you have any recommendations for advertisers on how they approach programmatic?
A: It comes down to understanding the risks and letting that understanding inform your selection of partners. Programmatic is moving fast, but it is still in its infancy. Brands face a true range of options when it comes to preventing fraud and ensuring a safe supply. Every demand-side solution will claim to be the safest and the furthest ahead of the fraudsters. What’s most important is that they remain nimble and transparent with their brand clients. The space is simply moving too fast for anyone to have the one solution.
Dwight Ringdahl, CTO, RhythmOne, is an engineer on the forefront of programmatic and the quest for true quality control in the space. He and his team are leading the charge on pre-emptive vs. forensic methods for assessing and ensuring quality and viewability before the ads serve.