A/B Testing & Causal Inference — 64,000-Customer Email Campa...

A/B Testing & Causal Inference — 64,000-Customer Email Campaign Analysis (Hillstrom Dataset) A full randomized-experiment analysis, built to show the difference between running a statistical test and running one correctly — including catching and fixing my own methodology bug mid-project. Result: A clear, statistically defensible recommendation (send the Men's email campaign to the full customer base) with honestly quantified conditions under which that recommendation would change.