Cyber security company PerimeterX is introducing a solution to ad fraud and other automated attacks that uses technology that identifies what legitimate users look like -- and blocks anything that’s out of the ordinary, in real time.
The "behavioral fingerprinting" tech recognizes anomalies in requests and behavior patterns in the user, device and network that bots can't hide. For instance, a bot, unlike a human, will not move a mouse in a random way and click on something.
If a bot is detected, PerimeterX blocks it, protecting against attacks like ad fraud, brute force (trying out a bunch of password and user-name combinations until one works); price scraping (competitors sending bots to see what your prices are); content scraping (extracting data or content from a Web site); layer-seven distributed denial of service (bringing down a site with too many requests) and scalping (buying goods, then selling them at a higher price).
Started in 2014, PerimeterX's goal is to change the way Web sites are protected, said its CEO Omri Iluz. The company, which closed $12 million in seed and Series A funding, has released the solution to 25 large sites since Bot Defender's September pre-release.
When the company first started, it set out to identify bots for the commerce and banking industries. But soon after, people started using it for advertising. Six months into the life of the company, Iluz said, and advertising is one of its strongest areas of growth. One reason? Iluz said other ad-bot solutions use weak signals like IP addresses or user agents.
It's easy to see why marketers want insurance that traffic is legitimate. A study from from the Association of National Advertisers (ANA) and Forrester found almost 70% of marketers cited higher bot fraud in programmatic buys as a concern. Advertisers are expected to lose an estimated $7.2 billion globally in 2016 as a result of bots, according to an ANA and White Ops study.
PerimeterX has offices in San Mateo, Calif. and Tel Aviv, and is currently building its sales and marketing teams.