This project benchmarks various casual inference/uplift modeling approaches for their ability to increase profit of online marketing campaigns. Treatment effects are modelled on four different datasets.
The models used are:
- Causal Honest Tree
- Causal Honest Forest
- Causal Boosting
- Causal Bayesian Additive Regression Trees
We found that using causal model for targeting in marketing campaigns would yield additional 50.000€.