Solution:
We implemented a solution based on audiences with artificial intelligence that prioritizes the preponderance of user conversion. We developed a Machine Learning model and pipeline automation to export data from BQ to GA4 using Measurement Protocol, this process is intended to enrich the categorization of users. The goal was to automate the whole process so that the likely audience was sent to Google Analytics and then to Google Ads with the shortest possible delay.
Results:
Effective Targeting Campaigns: Thanks to the forecasting model, we were able to target specific campaigns to user groups with a high probability of conversion. The implementation of an automated microservice to update the models with recent data supported this approach.
Improved Insurance Purchase Campaigns: We saw considerable improvement in all campaigns focused on insurance purchase conversions, significantly outperforming more tactical remarketing audiences.
Cost Reduction: We experienced a 5% reduction in Cost Per Acquisition (CPA) compared to remarketing audiences, which contributed to effective budget optimization.
Increase in Video Conversions: Recorded an impressive 33% increase in total video conversions, outperforming remarketing audiences.
New audiences that do not lose efficiency with platform optimizations.
