Liftoff Upgrades Cortex

Liftoff, a provider of performance marketing and monetization solutions for the mobile app economy, has upgraded Cortex, its neural network prediction engine that powers all bidding and performance decisions across its product suite, with enhanced predictive power and performance:

"Cortex fundamentally changes how advertisers compete by replacing static models with real-time, adaptive decisioning," said Andre Tutundjian, chief operating officer of Liftoff, in a statement. "It processes exponentially more data and runs continuous testing, so campaigns get smarter and more efficient with every impression. The result is better targeting, faster optimization, and stronger ROI at scale."

Liftoff is introducing the following innovations in Cortex:

• Unattributed samples so Cortex can now learn from conversions it did not directly drive. This expands the engine’s training dataset and provides a more complete view of user intent across campaigns.
• Multicast modeling for user-level ROAS optimization, which marks a shift from aggregate predictions to individualized, user-level valuation, allowing advertisers to identify high-value users and maximize return on ad spend.
• Sequential modeling for deeper behavioral insight. Cortex now leverages raw sequential inputs, moving beyond summarized metrics to ingest full behavioral sequences, including timing, location, and app context of each user action, resulting in richer user profiles and a more dynamic, real-time understanding of intent.