Fighting Ad Fraud
A series of reports published on the topic of ad fraud last year estimated that fraudulent traffic would cost the industry anything from $7.2 billion to $16.4 billion by the end of 2017. After a number of high-profile legal cases with the likes of Uber and ESPN, Adland was shaken. But in the immediate months after the scandals, we saw most players in the industry taking measures to deal with the ad fraud issue head on. Although the panic has died down somewhat, a number of initiatives like Appsflyer‘s ‘Fools No More’ campaign and Adjust’s CAAF have emerged and most companies have begun to get their anti-fraud ducks in a row. To learn more about how ad fraud works on mobile and how companies like FraugLogix are working to detect and prevent it, we caught up with Head of BI & Yield, Shrey Tewari and Hagai Shechter, CEO of FraudLogix to chat about our new partnership.
Why do you think we saw more media headlines on the topic of ad fraud last year?
This conversation around trust and transparency has been building in our industry for a while and as more data surrounding the monetary results of ad fraud on businesses became known, I think all the players in the adtech industry really started to take the issue seriously and start looking for solutions to prevent this. Oddly, it also became a buzzword with every second article and opinion piece focusing on the issue with no real solution offered. But I think now, after the media frenzy has died down, we can really see companies taking proactive and long-term steps to tackle the issue as we all know it negatively impacts every player in the industry from advertiser to user.
What has PubNative’s approach been to fraud prevention thus far?
We developed internal tools based on different trends and identifiers we saw. We also have active quality controls on both sides to make sure suspicious Publishers don’t enter our ecosystem.
Why did PubNative decide to partner with FraudLogix to provide ad fraud detection and prevention?
We had been using our own internal fraud detection tool for some time but we wanted a more equipped suite of services to help provide the highest traffic quality for our customers. We spent a lot of time doing research and talking to a number of fraud detection partners but we ultimately settled on FraudLogix as our fraud detection partner because of their long standing commitment to working with networks, exchanges and SSPs to help fight ad fraud in all its forms.
How bad is the ad fraud problem on mobile and does it differ from desktop?
More ad dollars have been shifted to the mobile space recently, and because fraud tends to follow the money, sketchy activity has also increased in mobile environments. Platforms like PubNative are smart to take preliminary action against it – before it impacts clients’ campaigns. Ad fraud on mobile is conducted a bit differently than desktop so different logic and algorithms have been developed to detect it. However, it’s still similar to desktop in that fraudsters are faking actions that marketers are willing to pay for, i.e., mobile app installs and ad clicks. In desktop they may launch fake websites to monetize, while in mobile it may be fraudulent apps used to create fake ad impressions. Other tactics involve setting up schemes to fake app installs and actions within apps. Mobile fraudsters tend to be a bit more sophisticated because the process takes longer and there’s more scrutiny in the space. There’s much less volume with app installs and payouts are higher, for example, than desktop impressions, but mobile fraud can be more lucrative. Like desktop, much of the fraud is conducted by malware that infects devices with bots and scripts, creating an army of zombie devices whose goal is to create fake impressions, clicks, and/or installs – any actions that they can get paid for. The fraudsters get paid for the work their bots are doing and advertisers lose out because chunks of their campaigns have been wasted.
How does FraudLogix typically work with companies like PubNative?
For companies like PubNative, who are helping mobile publishers to monetize their inventory, Fraudlogix provides solutions that are tailored specifically to their unique needs: the ability to monitor high volumes of traffic, identify where fraud is coming from within thousands of partners, block fraud pre-bid, and satisfy the demand side’s expectations of quality. By using Fraudlogix’s IP Block List, for example, a platform can automatically block traffic and actions that are coming from known fraudulent sources, i.e., bots, malware, data centers, etc. What’s unique is that the data that fuels the blocking lives in the client’s environment and is updated hourly. So there’s no latency and platforms can deploy it on all of their traffic instead of picking and choosing which campaigns or buying partners to protect. The block list filters out a majority of fraud, pre-bid, and has proven to improve the effectiveness of advertisers’ campaigns.
How does the technology work?
Fraudlogix’s core technology is pixel-based and is collecting data from traffic directed to millions of different websites, which results in direct contact with millions of unique users and billions of unique devices a month. The pixel picks up informational and behavioral data. Using machine learning, baselines for human behavior is established and anomalies outside of those baselines are identified. A simple example of an anomaly is a large amount of traffic coming from one IP address. Fraudlogix has over 40 different anomalies that it detects, all of which are weighted based on type and severity. These are ultimately signatures bots, automated scripts, and compromised devices. Fraudlogix tracks where fraudulent activity is coming from (i.e., the IP address), and if the majority of an IP’s traffic is fake, it’s included on the block list. Because fraudsters move quickly, the list is updated hourly. Fraudlogix clients access the list and block ad requests coming from IPs that are identified as being high risk for producing fake traffic.
If you’d like to learn more about how we’re tackling fraud, reach out to your PubNative account manager or shoot an email over to firstname.lastname@example.org. To learn more about FraudLogix and their advanced fraud detection capabilities, fraudlogix.com or email email@example.com.