Cognitiv

Advertising Goes Predictive with Deep Learning

Adweek's Adspeak podcast features Jeremy Fain, co-founder and CEO of Cognitiv, discussing how deep learning is transforming advertising by leveraging massive datasets, predictive algorithms and real-time optimization. The conversation emphasizes using first-party data, granular audience signals and log-level data frameworks to predict creative performance before impressions are bought, run continuous learning loops, personalize creative, and improve targeting. Fain frames AI as an efficiency multiplier that can unlock incremental gains at scale — citing the competitive value of relatively small percentage improvements when applied broadly — and stresses that deep learning in marketing is fundamentally a big-data problem rather than merely a creative tool.

Employee Spotlight Series: Q+A with Alex Nguyen

ML Research Scientist Alex Nguyen on model performance, a noisier open web, and why staying curious is the real strategy in adtech.

Fanatics Shifts to Value-Optimized CTV Campaigns

Fanatics Sportsbook changed its performance marketing approach after experimenting with adtech vendor Cognitiv, moving away from manually defined audience segments toward machine-learning models optimized for customer lifetime value (LTV). Fanatics reports a projected 19% increase in customer LTV from campaigns run with Cognitiv’s full-funnel CTV solution. The work relied on Cognitiv’s AudienceGPT product, which builds behavioral segments from natural-language prompts and trains custom ML models against advertiser-defined business outcomes. Fanatics’ paid-marketing director Andy Magnes and Cognitiv CEO Jeremy Fain described the shift as replacing audience hypotheses with algorithmic identification of high-value customers across channels including CTV and online video.