Google Upgrades Meridian With Agentic AI and Upper-Funnel Measurement

From ‘reactive report card’ to dynamic measurement

Google is updating Meridian, its open-source marketing mix model, with agentic AI capabilities that help marketers audit data quality, resolve errors more easily and get real-time guidance while building models. Back-end updates also improve speed and efficiency. Google says the changes are part of turning measurement from a “reactive report card” into something more dynamic. Here’s what’s new in Google Meridian MMM.

What’s changing

• Agentic AI: data quality audits, error resolution and real-time model-building guidance.
• Upper-funnel measurement: marketers can feed more brand-building signals directly into Meridian, improving visibility into harder-to-measure channels like TV and out-of-home.
• Meridian GeoX goes global: a publisher-agnostic incrementality tool for geography-based experiments is now available worldwide after beta. Google says incorporating GeoX insights improves MMM accuracy.
• Last week, Meridian also partnered with Circana’s Liquid Mix, a self-serve AI MMM platform.

Broader measurement updates

Google is integrating Data Manager directly into Google Analytics and Display & Video 360, adding enhanced conversions to both to improve ad relevance while securing customer data matching. Google Ads is also getting a Data Strength Uplift Metric that calculates conversions recovered through an advertiser’s first-party data setup.

Context and concerns

Google launched Meridian two years ago as a privacy-minded, durable alternative during third-party cookie deprecation. Google later dropped its plan to kill cookies, but marketers kept investing in alternative measurement, and MMM interest has revived. Brands such as Alaska Airlines have credited Meridian with uncovering new revenue opportunities. Still, the familiar concern remains: Google grading its own homework.

Practical takeaways

MMM is for cross-channel budget decisions

Platform attribution tends to overstate each platform’s role. MMM compares TV, out-of-home and digital in one model to guide allocation. As three ways to reclaim credit that search took from paid social show, capturing upper-funnel contribution that last click misses requires MMM plus incrementality tests.

Cross-validate

Because Google’s tool evaluates Google’s ads, design your own geo experiments and cross-check with other measurement sources. With platforms like Snapchat opening unified attribution to all advertisers, setting one neutral standard is where good measurement starts.