Recent Ad Fraud Schemes and How to Fight Them
Each year, marketers set out to solve ad fraud, but the systemic struggle is a long-term fight with no instant fix. To that end, I checked in with Paul Roberts, CEO of audience marketplace Kubient, to assess the state of ad fraud and how providers can keep it at bay this year.
Kubient recently discovered two ad fraud schemes, SynthNet and Wease.IM. What did these schemes involve? How did you identify them?
SynthNet is a fraudulent synthetic network which was designed in such a way where web-based or computer-generated traffic was sold as legitimate mobile app traffic. Their target was to sell fraudulent traffic, which was perceived to be originating from premium apps with the highest demand and CPMs.
The Weasel fraud exposed the fraudulent actions of a company called Wease.IM, which was running secondary auctions within display ad units via PreBid.JS. They devised custom code to alter PreBid.js requests and obfuscate the location of the iframe, hiding the true identity of the app.
Kubient’s fraud prevention team identified these schemes using Kubient’s Artificial Intelligence tool (KAI). KAI is trained to identify anomalies in the traffic, raising red flags when there is anything suspicious that humans cannot identify at the first glance. Once our fraud prevention team noticed these red flags, we took a closer look and uncovered the end-to-end working of these fraud schemes, as well as the significant impact they were having on the ecosystem.
How much should publishers and advertisers be concerned about ad fraud? How can they combat it?
Brands and publishers should and are very concerned about the impact of ad fraud. Right now, we live in a reactive world regarding ad fraud, meaning advertisers and publishers will continue to lose money to these schemes because we aren’t getting ahead of them.
Both sides need to make ad fraud a priority internally and externally. This is more than implementing software to check for fraudulent traffic and block it. It means understanding who is involved in the transaction, where traffic comes from, where ads come from, as well as heavily vetting partners you choose to work with to ensure they meet the industry standards and your own.
How does ad fraud affect CTV? Is CTV just as susceptible as other channels?
CTV is heavily impacted by fraud and is more susceptible than the other channels because it is not standardized across all platforms and devices. On the web, there is a standard protocol for browsers that works for everyone; mobile apps on Android and iOS have standards that also work for everyone. These standards were set to allow for transparent sharing of vital information used in fraud prevention.
CTV has numerous platforms, and they don’t adhere to any single standard. ROKU, TIVO, Apple TV, Android TV, ChromeCast, Amazon Fire, and Smart TVs like LG, Samsung, and others make it such a fragmented space that the traditional fraud prevention companies software can’t keep up. The reluctance for access and data from these device makers also prevents analysts from understanding just how much fraud exists in CTV. KAI takes a different approach with its AI technology and is able to work across the CTV ecosystem, regardless of platform, thus removing a key barrier others face in trying to detect fraud in this space.
How will ad fraud and efforts to fight it evolve in 2021?
In 2021, all ad fraud detection needs to move towards pattern-based approaches instead of label-based approaches or our industry will continually be playing catch-up to fraudsters and bad actors in the space. Ad fraud schemes are evolving and becoming more sophisticated, and that won’t stop.
It’s up to our industry leaders to work smarter, become more creative, and ultimately continue to prioritize stopping ad fraud to create a safer environment for advertising. This will protect ad budgets and ensure that ad dollars aren’t going towards false traffic.