In most countries economic prosperity is very unevenly distributed. Regional, urban and neighbourhood policies are often based on concerns about these kinds of disparities, and reducing such disparities is a key policy objective in many countries. High quality evaluation is central to understanding how to meet these objectives. However, impact evaluation - which seeks to identify the causal effects of policies - is often in short supply for spatial policies. In this viewpoint we highlight three barriers that hamper more rigorous impact evaluation. First, data availability constrains research. Second, identifying the causal impact of polices is difficult. Third, there are several practical barriers. We briefly consider each of these in turn, and make practical recommendations for change. Better policy design, more use of open data, and capacity-building for government analysts are three important and achievable steps in improving the extent and quality of future impact evaluations.