What is the ideal approach to integrating marketing mix and attribution with causal levers and experimentation, including field experimentation?
What advances are possible to generate efficient experimental designs, especially in online contexts?
• How should one balance causality and prediction?
