AI Agents & the Future of Marketing (MMM, Brand Lift & Benchmarks)

In this episode of Marketing Measurement Matters, Dr. Luca Fiaschi (Partner, PyMC Labs) shares how the world’s biggest spenders moved from fragile attribution to Bayesian MMM, why agentic AI will become the “universal translator” between models and marketers, and what their open-source MMM benchmarks reveal—including a head-to-head with Google’s Meridian.

What you’ll learn
Luca’s path from academic neural nets to Rocket Internet & HelloFresh (managing hundreds of millions in media)
• Why privacy shocks forced the shift to MMM—and how PyMC Marketing made advanced effects (adstock, saturation, seasonality, hierarchies, interactions) practical
• The next frontier: turning complex MMM outputs into business-ready decisions with AI agents
• Brand measurement beyond MMM: using Bayesian VAR to model long-term, upper-funnel effects and connect them to sales
• Synthetic panels: using LLM agents for consumer research (how to validate against human panels)
• Open-source validation: building transparent benchmarks with synthetic data at multiple scales; where PyMC Marketing outperforms and where Meridian struggles (speed, scaling, seasonality)

Chapters
00:00 Intro
01:00 Luca’s background & early marketing analytics
03:30 Rocket/HelloFresh: scaling budgets & measurement
05:00 Privacy changes → MMM adoption
07:00 Making MMM explainable for business stakeholders
15:20 Brand measurement with Bayesian VAR (long-term effects)
19:30 Agentic AI for coding assistance & “universal translator” insights
21:45 Synthetic panels for consumer research
27:00 Open-source MMM benchmarks: PyMC Marketing vs. Meridian
33:00 Wrap & resources

Subscribe for more deep dives on MMM, incrementality, and AI in marketing.
Drop your questions below—who should we bring on next?


#MarketingAnalytics #MMM #AIinMarketing #MarketingMeasurement #OpenSource