Meridian 2.0: Introducing Meridian GeoX & Automated Calibration, Full-Funnel MMM, Agentic Skills, and More

[kmm022]

New update v2.0 - We're excited to announce the largest update to Meridian and the general availability launch of Meridian GeoX, available to everyone! Meridian GeoX is Google's new open source geo experiment solution that enables you to measure cross-platform causality and seamlessly calibrate your MMM. With this v2.0, Meridian now offers automated experiment calibration, a new full-funnel MMM framework to capture the longer-term effects of Brand media, an agentic skills module, and faster optimization.

Register and join our Discord livestream on Sept 9th at 9:00 AM PST to learn more about these Meridian updates and Meridian GeoX and speak directly with our Product team.

NEW FEATURES

1. Easily ground models in real-world proof– We’ve introduced a new geo experiment solution, Meridian GeoX, under the Meridian umbrella. With this, you can run Meridian GeoX independently and use the new automated calibration module in the Meridian library to seamlessly integrate causal lift results directly into your MMM. We’ve also expanded the model health module to help you quickly see how calibration affected your model and identify channels recommended for experimentation.
   Next Step: Check out Meridian GeoX and review the Meridian Prior Calibration Demo for implementation details.

2. Connect Brand to outcomes with our Full-funnel MMM Framework - With Meridian’s new full-funnel MMM framework, you can capture the full effect of brand media by acknowledging how upper-funnel awareness creates later lower-funnel demand. By incorporating Brand Equity signals like branded GQV directly into your models, you can accurately track the downstream, indirect impact of upper-funnel media. This seamlessly connects brand to outcomes, ensuring brand investments receive their proper credit across the entire funnel for a complete understanding of true ROI. Next Step: Review the developer documentation and follow the instructions in the Colab notebook to implement the new methodology in their Meridian model.

3. Get enhanced guidance with Meridian Agentic Skills – Meridian’s agentic skills repository is a new MCP-based toolkit that empowers you to employ AI agents, such as Antigravity CLI, for interactive, step-by-step modeling guidance. Leveraging Meridian’s EDA and model health modules to proactively recommend solutions for detected issues, it provides an automated and fast path to resolution. This AI-first solution removes the steep learning curve of traditional developer documentation, ensuring Meridian users have seamless access to high-quality modeling. Next Step: Follow the developer documentation to complete the simple install and start using Antigravity CLI to build, run, and get accurate guidance for their next Meridian model.

4. Optimize spend faster with more flexibility - Meridian now runs 2x faster and 4x more efficiently out-of-the-box with its new JAX backend. Also, Meridian's enhanced optimization framework makes it even easier to run budget scenarios with more flexible time frames, enabling you to plan weekly vs quarterly or yearly. This gives you the flexibility to quickly optimize your marketing spend for the exact time frames that matter most to your business.

Please be aware that v2.0 is a major release with breaking changes. You should review the CHANGELOG to understand the potential impact before upgrading models to v2.0.

[liftlabanalytics]

Great update on Meridian 2.0, especially the integration of GeoX experiments with automated MMM calibration. Connecting causal lift results with MMM can make measurement and model calibration much more actionable. For teams exploring Incrementality Testing & MMM Calibration, this is a valuable development to follow: https://liftlab.com/solutions/incrementality-testing/