Liftoff Upgrades Cortex

Why It Matters

By boosting predictive accuracy and real‑time decisioning, Liftoff helps advertisers cut wasted spend and capture high‑value users, a critical edge in the increasingly competitive mobile app market.

Key Takeaways

• Cortex now learns from unattributed conversions, expanding training data
• Multicast modeling enables user‑level ROAS optimization for high‑value users
• Sequential modeling ingests full behavioral sequences for richer intent insights
• Real‑time adaptive decisions improve targeting, optimization speed, and ROI

Pulse Analysis

Liftoff has become a cornerstone for mobile app advertisers, offering a suite of performance‑marketing tools that rely on data‑driven bidding. In a market where user acquisition costs are rising, the ability to predict which impressions will convert is a decisive advantage. The company’s neural‑network engine, Cortex, has powered these decisions since its launch, but the rapid evolution of app ecosystems demands ever‑more sophisticated modeling. Upgrading the core prediction layer signals Liftoff’s commitment to staying ahead of algorithmic competition and delivering measurable ROI for its clients.

The May 19 upgrade introduces three technical breakthroughs. First, ‘unattributed samples’ allow Cortex to incorporate conversions that occurred outside Liftoff’s direct attribution window, enlarging the training set and reducing bias. Second, multicast modeling shifts from aggregate forecasts to user‑level return‑on‑ad‑spend calculations, enabling advertisers to pinpoint high‑value individuals and allocate spend with surgical precision. Third, sequential modeling feeds raw event streams—timing, location, app context—into the neural net, creating richer user profiles and allowing the engine to adapt in real time as behavior unfolds. Together, these features boost predictive fidelity and speed.

From a business perspective, the enhancements translate into faster campaign optimization and higher return on ad spend, especially for advertisers targeting high‑value users in competitive app categories such as gaming and fintech. Real‑time adaptive decisioning reduces wasted impressions, a critical metric as mobile CPMs hover around $2‑$3 in the U.S. market. Competitors like Adjust and Appsflyer are also investing in AI‑driven attribution, but Liftoff’s focus on user‑level ROAS gives it a differentiated value proposition that could attract larger brand budgets and drive industry‑wide adoption of granular, sequential modeling.