Liftoff Introduces New Cortex Innovations to Boost Mobile Advertising Performance, unveiling a trio of AI‑driven enhancements that promise faster, more precise bidding decisions for programmatic mobile campaigns. The Redwood City‑based firm says the updates—unattributed‑sample learning, multicast ROAS modeling, and sequential behavioral inputs—are already powering a 41 % lift in its Core Advertising platform revenue.
What’s new in Cortex
Cortex, Liftoff’s proprietary neural‑network engine, now ingests three additional data streams. First, “unattributed samples” let the model learn from conversions it did not directly drive, expanding the training set and tightening prediction accuracy. Second, a multicast model shifts ROAS optimization from aggregate estimates to user‑level valuations, enabling advertisers to earmark high‑value prospects with finer granularity. Third, sequential modeling replaces summarized metrics with raw event sequences—timing, location, and app context—giving the engine a richer, real‑time view of intent.
Why the upgrades matter
The three upgrades translate into measurable performance gains. Liftoff reports a 4× increase in experiment velocity and the ability to launch optimized campaigns in under 24 hours, compared with the typical two‑week ramp‑up without Cortex. For marketers, that speed means less budget waste and a tighter feedback loop between creative, targeting, and spend.
Industry implications
Mobile remains the dominant screen, accounting for more than 60 % of global digital ad spend, according to eMarketer. As advertisers chase higher ROI, AI‑powered bid optimization has moved from a nice‑to‑have to a competitive necessity. Liftoff’s new features address two persistent pain points: data sparsity (through unattributed learning) and the lack of granularity in ROAS forecasts. By delivering user‑level predictions, Cortex narrows the gap between demand‑side platforms (DSPs) and emerging identity‑resolution solutions from firms like The Trade Desk and Adobe.
How Liftoff stacks up against rivals
Competing AI engines—Google’s Bidding Engine, Amazon Advertising’s DSP, and Meta’s Automated Ads—rely heavily on first‑party data and large‑scale inventory. Liftoff differentiates itself by focusing on SDK‑derived signals from over 167 k apps, reaching roughly 1.4 billion daily active users in Q1 2026. This breadth of mobile‑only data gives Cortex a signal‑to‑noise advantage that many cross‑device platforms lack. Moreover, the multicast ROAS model mirrors capabilities that have only recently appeared in enterprise‑grade CDPs, suggesting Liftoff is compressing two technology stacks into one platform.
What enterprise marketers should expect
For large brands, the practical impact will be threefold. First, campaigns can be launched and iterated in a single business day, aligning ad spend with rapid product cycles. Second, the user‑level ROAS forecasts enable more aggressive bid scaling on high‑value segments without overshooting budgets. Third, the sequential behavior inputs open the door to creative optimization that reacts to in‑app actions—think dynamic ad creatives that adapt to a user’s last interaction within a game.
Market Landscape
The adtech market is at a crossroads of privacy regulation and AI acceleration. Gartner predicts that by 2027, 70 % of marketers will rely on AI to allocate budgets across channels, up from 35 % in 2023. At the same time, iOS 16.5 and Android’s privacy sandbox have curtailed third‑party cookie efficacy, pushing firms toward first‑party signal enrichment. Liftoff’s SDK‑centric approach sidesteps many of these constraints, positioning Cortex as a privacy‑compliant, high‑volume data source.
Competitors are responding. The Trade Desk recently announced “Unified ID 2.0” enhancements, while Google’s “Performance Max” campaigns lean on its own AI. Yet none combine the depth of mobile‑only event sequencing with user‑level ROAS modeling at Liftoff’s scale. For enterprises that already operate on Liftoff’s Accelerate tier, the upgrades represent an immediate lift; for newcomers, the promise of sub‑daily optimization may justify a migration.
Top Insights
- Speed to market: Cortex now supports sub‑24‑hour campaign optimization, cutting traditional testing cycles from weeks to hours.
- Data completeness: Unattributed sample learning expands the training set by up to 30 %, improving prediction confidence for low‑frequency events.
- Granular ROI: Multicast ROAS modeling delivers user‑level profit forecasts, enabling precise bid scaling on high‑value audiences.
- Behavioral depth: Sequential modeling ingests raw event streams, offering richer context than aggregated metrics used by most DSPs.
- Industry shift: With privacy‑first mobile data at scale, Liftoff’s Cortex positions AI‑driven bidding as a core capability for enterprise marketers.
