In the policy arena we have tended to focus much of our attention on causal questions. For example, is something working or not or if I do X how will it affect Y. Predictive questions, what will happen in the future, tend to not be discussed. However prediction, using the methods of data science, can actually be very useful in policy making. Particularly when it comes to the allocation of resources. In this talk, Andrew Means, cofounder of The Impact Lab, will discuss the role of prediction in the targetting of policy resources. He will draw from examples health, politics, and more to illustrate how prediction can be used to improve the outcomes of policy interventions.