Automating SEA Retargeting for Smarter Audience Engagement and Higher Conversions
summary
An automated pipeline identifies high‑propensity site visitors via batch‑processed user journeys stored in AWS S3, engineered features (pre‑conversion events + reservation params), and models trained with Dask dataframes and CatBoost with Bayesian optimization; audiences upload daily to Google Ads (seed + quantile selection) and target non‑converters from the previous day via retargeting pixels. Results: 9% audience uplift, 113% conversion‑rate uplift, 22% lower cost per conversion, 30% spend reduction.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:08Meeting agenda
- 2:11SEA and retargeting context
- 7:12Data blueprint
- 11:48Why ML is essential
- 16:48ML problem statement
- 19:17Feature engineering
- 21:25Handling data imbalance
- 22:48Model training pipeline
- 25:58Audience performance results
- 27:44Q&A
related talks
19mSession: Re-Engagement In 2026: How To Design & Scale Winning Retargeting Campaigns
22mSession: Meta Ads: The Next Chapter
14mMarketing Automation Solutions To Power Up Your WordPress Website
48mGoogle Shopping & Performance Max: How To 2x Your ROAS in 2024
41mFuture Of Retargeting - Hopes & Expectations
18mAdvanced ad strategies that drive revenue and boost engagement
27mGenerating targeted AI revenue at mass scales
39mUnderstand and Engage Your Top Users with Dynamic Audiences