Why start using uplift models for more efficient marketing campaigns
summary
Uplift modeling replaces propensity scoring to optimize incremental revenue by estimating conditional average treatment effect per user using Uber’s open-source CausalML (meta‑learners, uplift trees, neural nets, instrumental variables). A production pipeline uses holdout/explore/exploit splits (explore ≈10%), scores millions (20M+ users across 5 mega regions, 11 countries) monthly, and in one campaign delivers over 100% of original incremental value while targeting ~50% of the population, cutting spend and identifying negative “sleeping dog” segments.
AI-written summary · about
chapters (11)
- 0:00Introduction and speaker background
- 1:26Traditional marketing models and inefficiencies
- 3:32Problem with propensity models in marketing
- 7:24What is uplift modeling
- 9:35Use cases and benefits of uplift modeling
- 10:40Uplift modeling algorithms in CausalML
- 13:34How to use uplift modeling results
- 15:59Evaluating uplift models
- 17:26Real-world use case at Uber
- 21:39Q&A session begins
- 30:13Closing remarks and thanks
related talks
38mHow to Enhance Customer Targeting in Marketing
13mOptimizing Subscription Services with Uplift Modeling
43mBeyond Just Prediction: Causal Thinking in Machine Learning
01h 00mAI Powered Uplift Modeling and Stream Forecasting in Music Industry
44mIntroduction to Uplift Modeling
24mImprove your marketing reach using ...
31mScaling up Lyft’s Growth via Causal ML and Bandits
24mUnlocking revenue: newcomer classification for boosted profits