Abstract: Understanding how customers would react to various offers is of significant importance in a large retail organisation such as M&S. However this task is non-trivial and requires us to identify the causal impact of one set of action on another. In M&S we use machine learning methods in conjunction with causal inference to understand the "why" and "what-if" scenarios. In this talk I will give an introduction to how we use causal inference and meta-learners to provide the most optimal offers to customers who are part of the M&S loyalty program "Sparks".
