Incubeta

The winning formula

Data Innovation: Incubeta

Triangulated measurement for a global fitness platform

The Challenge

A global on-demand fitness subscription brand needed to prove what its paid media actually added in the US market. Around 60% of demand was baseline organic. CPA spiked sharply in peak trading periods, and last-touch reporting could not separate genuine growth from demand the brand would have won anyway.

The Approach

Incubeta built one measurement foundation from three independent sources of evidence:

• Google Meridian MMM: A Bayesian model fed with granular data from the Google MMM Data Platform, GA4, and CRM subscription data. It isolated ROI and saturation for each lever: Performance Max, Demand Gen, YouTube, DV360, and Brand versus Generic Search.
• Geo-lift experiments: Controlled tests to confirm the model reflected causation, not correlation. The results calibrated the model’s priors.
• Scenario planning: Response curves and constrained optimisation (±30% per channel) to model how to move budget out of saturated channels and into formats with headroom.

The Results

• 16.7% incremental lift for Google media, confirmed through geo-lift testing.
• 83% model fit (R²) and 16% error (MAPE), well ahead of the 64% R² of the regression model it was tested against.
• Budget reallocated at flat spend: +4% Paid Search and +2% Google Video, with Meta and TikTok each down 3%. This grew conversion volume and improved incremental CPA.
• Scenario planning in leadership hands: US leadership now uses automated scenario planners to guide budget decisions.

Why It Matters

The client now allocates budget on proven incrementality, not last-click credit. Meridian is now its measurement standard in the US, with a proven framework ready to extend to other markets.

The Team

Led by Andrew Thomson, with Ben James, Sam Kerr, Jia Ming Low, and Wei Hao Low (Incubeta).