Interview question asked to Data Scientists interviewing at CRED, NVIDIA, Discord and other companies. Original question asked: How would you turn a 3% observed difference between campaigns into a rigorous causal analysis?.