Causal inference requires contrasting counterfactual states of the world under pre-specified interventions. Obtaining counterfactual contrasts from data relies on explicit assumptions and careful, multi-step workflows. Causal diagrams are powerful tools for clarifying whether and how the counterfactual contrasts we seek can be identified from data. Here, I explain how to use causal directed acyclic graphs (causal DAGs) to determine whether and how causal effects can be identified from ‘real-world’ non-experimental observational data. I offer practical tips for reporting and suggest ways to avoid common pitfalls.