Randomized trials with primary data collection are the preferred way to estimate causal effects, but they are not always feasible, ethical, or timely. The number of clinically relevant questions exceeds the number of possible randomized trials. Therefore, researchers use data from observational studies to answer causal questions. Causal inference from these data mimics a pragmatic randomized target trial. The protocol specification of the target study helps to define the causal question and the target quantity of interest. Protocol components include eligibility criteria, treatment strategies, treatment assignment, outcomes of interest, start and end of follow-up, and causal contrast. Precise specification of the target study allows for its emulation, but it may fail. The article highlights the reconciliation of eligibility and treatment assignment at time zero when replicating the target study.