How Kepler helped easyJet discover the true impact of their ads on Meta technologies

An uncertain measurement model

For easyJet, the Marketing Mix Model (MMM) is an important part of its strategy. easyJet combines MMM with geo-testing and publisher lift tests, with geo-testing serving as the core principle for calibrating the MMM. This approach helps guide multi-million pound budget allocations to drive profitable growth.

But the business faced a significant measurement challenge within this approach. While measurement was in place, there was a high degree of uncertainty around its social channel, with the MMM generating a Margin ROI with wide and unreliable confidence intervals.

This was compounded by an integrated media framework, where social activity was highly correlated with other channels, like display. Lacking a robust, holistic measurement strategy, easyJet’s model struggled to isolate and confirm the distinct impact of its ads on Meta technologies, and the business couldn’t confidently allocate their budget to the right channels.

Taking steps towards incrementality

easyJet wanted to fundamentally improve the accuracy and reliability of the MMM to enable more confident, data-driven investment decisions.

The goal was to resolve uncertainty around their measurement approach by creating a stable, causal signal for Meta technologies. The airline worked closely with Kepler to develop a rigorous, experimental program that would help them to measure the true impact of their social ads.

The approach

Working together, Kepler and route1.io, easyJet’s marketing analytics partner, designed a multi-faceted program to achieve this, combining a platform-based Meta Conversion Lift study and several geo-lift tests.

The Meta lift study, which ran from December 2024 to January 2025, used a multi-cell design to test a high-value audience strategy against a business-as-usual approach. route1.io then synthesised the insights from these experiments using their advanced Bayesian modeling methodology, where this causal data was integrated as a prior within the MMM to inform and stabilise the overall channel measurement.

Regaining measurement confidence

The approach delivered transformative results for both campaign execution and measurement confidence. The Meta lift study identified a winning targeting strategy that drove an 11.48% incremental lift in purchases and a higher Margin ROI.

After cross-validating these findings with geo-lift tests, the causal data was integrated into the MMM, leading to a confident, evidence-based decision to increase the Meta budget by 35% in the following quarter, as well as: