Member Only Access
Summary
As privacy regulations and the deprecation of third-party cookies limit access to individual-level consumer data, advertisers are increasingly forced to rely on aggregate metrics to evaluate marketing effectiveness. This MSI working paper introduces a novel, two-stage, marketing mix modeling framework designed specifically for cookie-free environments. By combining machine-learning–based directional prediction with classical econometric calibration, the approach demonstrates how firms can extract reliable signals about campaign effectiveness—even from short, noisy, aggregate time series—while maintaining interpretability and practical relevance for marketing decision-making.
