Cydersoft Uses Machine Learning to Battle Fraud-Bots

September 19, 2016

12:56 pm

Machine learning is, as simple search for news can prove, a hot topic in the tech industry. Its varied applications make machine learning truly unique: from the existing ones, such as disease prediction or loan assessment, to future applications, like generating videos from photos or identifying pixelated faces, the list is virtually endless.

Another important application of machine learning has to do with Internet security. More specifically, an issue that affects the mobile advertising industry: fraud. The rise in fraud is one of the greatest challenges currently that mobile video advertising has to face these days, as bots and other sources can have a negative impact on the advertisement’s effectiveness.

Along with its rise, fraud is also becoming increasingly more sophisticated. All this causes advertisers to lose important parts of their investments, with all the consequential poor-quality traffic. In order to avoid this, Cydersoft is launching a real-time fraud prevention module for VideMob, a comprehensive set of mobile video advertising technologies and monetization services.

VideMob’s Anti-Fraud solution differentiates between legitimate and fraudulent traffic, such as bots, harnessing the power of machine learning. In addition, it uses that information to block frauds in real-time, or to alert on invalid traffic. This allows for both publishers and advertisers to create a safe and effective advertising environment.

With VideMob, all traffic is monitored through every step of the ad workflow. Most existing alternative systems can only detect fraud at the request level, with the use predefined black/white lists instead of real-time monitoring. This Anti-Fraud system, however, uses dynamic events to detect abnormal user behavior, greatly increasing its effectiveness and reducing the impact of fraud. Here is a breakdown of this system’s unique features:

  • Real-time Analysis and blocking: based on step-by-step analysis of the full workflow, from request to impression. Enables real-time fraud blocking and provides users with in-depth monitoring and reporting.
  • Machine Learning Algorithms: constantly adapts to evolving threats using machine learning technology.
  • 24/7 Monitoring: constantly monitors all streams, so that users can be alerted to potential fraud at any time.
  • Detection of Non Human Traffic (NHT): detects NHT and abnormal user behavior by using various tools.
  • Actionable Insights: provides clear and detailed graphs and tables to ensure actionable insights.

Semion Rotshtein, CEO of Cydersoft, briefly explained why VideMob’s Anti-Fraud system is valuable:

“One of the toughest struggles in video advertising is fraud, which affects both publishers and advertisers. With our real-time built-in anti-fraud solution, analyzing more than 5 billion daily requests with up to 1.5M QPS, we are cleaning up the industry, and providing both parties with a solution to maximize their investments.”

That value can also be explained with the use of statistics. Over the past 6 months, VideMob made a pre-launch monitoring of its customers traffic, discovering a 46% fraud and invalid traffic rate – nearly half of all mobile video advertising traffic. From that rate, 73.9% was blocked in real time at the request level (mainly comprised of non-human traffic), and 26.1% was detected from the full workflow including the impression level.

So, if you are present in this field and want to improve the effectiveness and security of your mobile advertisement, the VideMob platform is a serious option to consider.

Did you like this article?

Get more delivered to your inbox just like it!

Sorry about that. Try these articles instead!

25 y/o, born and living in Portugal. Majored in Biology, but tech and computers were always a passion. Wrote for sites like Windows.Appstorm and MakeTechEasier.

Leave a Reply

  • (will not be published)
Startup_Mixology_300x250