The goal is to create a marketing optimization tool that Blue Field Agency can integrate into its routine budget allocation processes for RESET. The tool aims to enhance RESET’s key performance indicator, brand awareness. By applying econometric models and optimization algorithms, the research seeks to address the following question:

The researchers implemented a Marketing Mix Modeling (MMM) approach, using various machine learning algorithms including Elastic Net regression, linear regression, logistic regression, and random forest to account for factors like adstock and saturation effects. They also utilized the Nevergrad optimization algorithm to fine-tune model hyperparameters.

The findings indicate that while certain trends and patterns in campaign performance can be identified, some models, especially the Nevergrad optimization algorithm, faced issues with convergence and reliability. Different algorithms showed varying effectiveness depending on the KPI being optimized, revealing limitations in directly applying model predictions as future indicators. The study provides insights into seasonal impacts and campaign efficacy, concluding with recommendations for optimal budget allocation based on the best-performing models for different KPIs.