Google Meridian | Treatment Prior Types

Why It Matters

Proper prior selection in Meridian directly improves media attribution accuracy and drives more effective budget decisions across paid and organic channels.

Key Takeaways

• Meridian splits treatments into paid media and organic/non‑media categories.
• ROI, MROI, and contribution are the three primary prior types.
• Use contribution priors for organic treatments lacking spend data.
• Choose MROI for conservative budget optimization aligned with historical allocation.
• Align prior selection with the business metric you prioritize most.

Summary

The video explains how Google’s Meridian platform classifies marketing treatments and selects appropriate statistical priors for each.

Treatments fall into two buckets—paid media, which has direct spend data, and organic or non‑media actions that lack spend. Meridian offers three dedicated prior types—ROI, marginal ROI (MROI), and contribution—plus coefficient priors, allowing analysts to translate business knowledge into model‑ready inputs.

For organic and non‑media treatments, the contribution prior is recommended because it measures the share of total outcome without requiring spend. Paid media users can choose ROI when lift studies are available, MROI for conservative budget optimization, or contribution when percentages better reflect domain expertise. A decision tree illustrated in the tutorial guides this selection.

Choosing the prior that mirrors the metric a business cares about ensures consistent attribution, more accurate budget allocation, and better alignment between statistical outputs and strategic goals.

Original Description

Explore the various prior types available for different treatment variables in Meridian. This video breaks down when to use specific metrics—such as ROI, marginal ROI (mROI), Contribution, and Coefficient priors—for paid media, organic media, and non-media treatments so you can best align the model with your domain knowledge.

Learn more:

https://developers.google.com/meridian/docs/advanced-modeling/how-to-choose-treatment-prior-types

#GoogleMeridian #MarketingMixModeling #BayesianStatistics #ROI #MarketingAnalytics #causalinference