Mobile attribution and marketing analytics player, AppsFlyer, has unveiled a new data forecasting study. According to AppsFlyer, app marketers will lose $100 million in 2016 due to verified mobile app install and engagement advertising fraud -- as well as revealing the countries that have the highest rates of this type of ad fraud. The study's findings are based on the company's new anti-fraud technology, called DeviceRankTM, which was unveiled recently, as well. The technology provides protection for mobile app advertisers by automatically "cleaning" fraudulent devices from marketers' campaigns aimed at driving app installs and post-install activities.
While the study found the U.S. to be the most targeted region by fraudsters, the countries with the highest rates of app install and engagement ad fraud, when factoring for mobile population, are: Germany, Australia, China, Canada and the U.K., followed by the U.S., Russia and France. The research shows that fraudsters attempt to target specific countries depending on the potential payout they can get from falsifying their location to commit fraud. Countries with the highest cost-per-install and cost-per-action payouts have a higher fraud rate, while regions with relatively low payouts - including Indonesia, India, Brazil, Vietnam and Thailand have a lower fraud rate.
Further, while roughly 50% more fraud is perpetrated on Android, iOS is not immune and is often targeted heavily since, if successful, payouts tend to be far higher on iOS than Android for those perpetuating app install and engagement fraud.
With marketers increasingly spending higher budgets and putting more efforts behind mobile app install and engagement campaigns to attract loyal users, the threat of mobile app install and engagement fraud continues to grow. Most install fraud solutions rely on IP filtering and data modeling at the user or app-level.
AppsFlyer's DeviceRankTM technology works similarly to a Credit Score, identifying questionable behaviour and offering enhanced protection. It leverages a proprietary big data-powered algorithm to build an anonymized, multidimensional rating of every mobile device. Each device is rated on a scale from C (fraudulent), through B, A, AA and AAA. Devices with a "C" rating are automatically excluded from AppsFlyer's attributed installs and analytics. With more than 1.4 trillion mobile interactions catalogued in the internal database over the past five years, and 98% of all mobile devices across the globe already rated, DeviceRank aims to represent the most comprehensive fraud prevention technology in the world. Additionally, DeviceRank's unique architecture and machine learning allow the database and algorithms grow, learn and adapt as new mobile devices come online, new interactions are catalogued and user engagement patterns evolve.
"As we've seen from our global study, fraudsters and scammers are growing increasingly sophisticated, tricking advertisers into paying for both installs and in-app engagement," said Oren Kaniel, co-founder and CEO of AppsFlyer. "DeviceRank takes a radically different approach, cutting off fraud at the source and adding transparency to our industry in order to protect advertisers, our partners and the entire market. AppsFlyer is in a unique position to apply this game-changing technology and leverage our scale across the mobile industry, arming marketers worldwide with the best solution to combat fraud. We are excited to add DeviceRank to the AppsFlyer Active Fraud Suite, as we remain committed to fighting fraud and helping cleanse the app ecosystem of this problem."
According to the new study, the $100 million cost to marketers in 2016 stems from fraudulent click data, paid installs from fraudulent devices, fraudulent and simulated in-app events (CPA fraud) and the projected impact of these false installs and events on lookalike targeting and retargeting. Additional insights in the research include a breakdown of mobile app fraud by country, operating system and more.
The insights in the study highlight key issues for mobile app marketers. Ratings are built from over a dozen points of metadata, including device details, engagement history, historical fraud modeling scoring, IP address and geolocation, transactional data and verified transaction data and more. Each device's rating is based on the likelihood that device is an authentic user (rated AAA) or a scammer (rated C). AppsFlyer's proprietary machine-learning scoring system updates each DeviceRank score with every interaction, providing real-time insights and protection.
No comments:
Post a Comment