Google first launched the MMM in March 2024 and has since tested it with ‘hundreds’ of brands globally.

Meridian uses Bayesian causal inference, with the model combining prior knowledge with signals learned from data to estimate media effects and quantify uncertainty.

The model analyses campaign performance based on key performance indicators (KPIs) including sales, website visits, profit and conversions. Users can run multiple optimisation scenarios and include other non-media variables to assess the impact of pricing and promotions on KPIs.

Writing in a blog post, Harikesh Nair, senior director, data science, said: “Traditional MMMs, built for offline media and branding, have historically been unable to fully measure performance media, like Search ads, and AI-powered campaigns.

"They lack a modern approach, which may lead to inaccurate budget decisions. They may not update assumptions based on incrementality experiments or provide granular insights, such as optimal frequency or regional nuances.”

Google said it will continue to invest in the MMM in the coming months, with new features and updates to its methodology.

Alongside making the model available to everyone, Google has launched a partner programme with a group of agencies and measurement vendors, with the aim of helping marketers to use the model.