Audience discovery firm Comprendi has launched an open beta program for its automated audience discovery platform for social media, which has seen increases of up to 2500 per cent in app installs during initially tests.

The platform uses proprietary natural language processing and deep learning algorithms to hyper-target audiences on social media. A private beta with clients including Zynga, Glu and Kabam showed huge growth in app install volumes, as well as 30 to 60 per cent reduction in cost-per-install and cost-per-click, and 30 to 50 per cent increases in day seven ROI, one of the key benchmarks for measuring quality of users.

“Comprendis unique offering has been instrumental to our success on Twitter,” said Mark Beck, vice president of user acquisition at Product Madness. “They took us from zero to top 20 per cent of installed gaming apps in just a few months.”

The public beta aims to address the unique needs of large-scale mobile app, performance-based and brand advertisers with a solution that reaches contextually relevant audiences at scale, as well as raise awareness of Comprendis other tools for audience targeting.

“Our platform gives advertisers a completely new and effective way to optimise their user acquisition efforts across social media channels,” said Itai Ben-Zaken, co-founder and CEO of Comprendi. “The results were achieving for clients are truly unmatched in reducing user acquisition costs while increasing ROI and scale.

“Were uniquely positioned to offer this kind of solution in the market because our technology enables us to analyse high volumes of data across multiple sources, and interpret those results in real-time.”