With programmatic advertising spend expected to reach $21.2 billion next year, the industry becomes more complex and advanced. So does fraud. In fact, according to the Association of National Advertisers, ad fraud will cost the industry $7.2 billion this year.
Monica Vazquez heads the Fraud Prevention team at Fyber, with a goal of maintaining the industry’s most robust, brand-safe programmatic video marketplace — one where DSPs and media buyers are assured of the quality of ad placements, and publishers and app developers get the most value from their inventory.
Fraud has been one of the top concerns for advertisers, how serious of a threat is this?
Fraud is part of the discussion about quality — and yes, I think there are areas where both buyers and sellers are rightfully concerned. But first, we have to take a step back and remember that fraud is an issue with all digital advertising, not just programmatic.
We’ve learned quite a bit at Fyber in recent years. For example, the realization that while you do need third party providers to combat fraud, you can’t only rely on them. Ad tech companies like us have to be diligent — that’s why we have a dedicated team on board to check each and every impression, to analyze strange patterns, and prevent bad inventory from getting into the system in the first place.
Anti-fraud tools are getting better, but the bad guys are getting smarter, too, meaning you have to constantly invest in beating them. And we firmly believe that only tech vendors and suppliers who are willing to make that investment and stay ahead of the curve will survive.
What are some of the programs and partnerships that Fyber has in place?
With pre-bid programmatic traffic monitoring through Forensiq, all of Fyber’s inventory is graded before any ad request converts into an impression. On the other hand, Pixalate helps Fyber track strange patterns or behaviors post-bid, in order to ensure that ads only run on premium sites and apps. In addition to these third-party verifications, Fyber’s Fraud Prevention team also provides 24/7 traffic quality monitoring through a combination of human and machine-based analysis in order to preserve inventory quality within the entire Fyber stack.
What is the difference between pre-bid and post-bid verification?
Pre-bid verification entails sophisticated machine-learning techniques that grade the inventory before an ad request is sent to the DSPs for bidding.
Post-bid verification is the combination of human and machine-based analysis that helps track strange patterns or behaviors to ensure that ads only run on premium sites and apps, and that interactions are real.
How does Fyber hold other members of the ecosystem to similar high standards?
It really comes down to education. At Fyber, we are regularly in discussion with our networks and publisher partners on what standards we have, what we expect from our partners, and we make sure to share best practices.
We also set the bar high for our app publisher partners. Our selection process for partners is strict and it is something that we see as necessary in order to create an honest marketplace. For app publishers, we ask for proof of SDK integration, test the quality of the traffic, and we test every app that goes live with us.
On the programmatic side, we review every single site and application before it goes live. We require our partners to work with third party verification tools so that they feel responsible for the quality filtering as well. We measure our partners using such indicators as quality of their traffic or fill rate numbers. Therefore, the more they can filter on their end, the more beneficial it will be for their numbers as well.
How long have you worked in ad tech and fraud prevention? What changed in the ecosystem in that time?
I started in AdOps four years ago, moving into Support Operations and taking care of integration projects such as Salesforce Cloud later on. Then my career path took me into Ad Quality and Malware Detection, which is how I got interested in the topic of publisher fraud detection and prevention.
The primary change I’ve witnessed is that fraud tactics have become more advanced. The industry is definitely taking it more seriously, but there is still a lot more for the industry to do to address fraud. For example, viewability has become an important issue for advertisers.
The product and ecosystem would be cleaner if all the companies agreed to certain standards. I know it sounds utopian, but the money that each ad tech company loses to fraud could be used instead for innovation.
Do you think things can improve and how?
It is going to take more knowledge sharing across the industry to stop fraud globally. For example, a single shared database of fraudulent players would significantly help reduce fraud. Unfortunately, as of now the members of the ecosystem don’t share this information readily. Let’s see if that’s a step that companies in the ecosystem are willing to take, I am hopeful for a change.