Replication is the cornerstone of science - but when and why? Not all studies need replication, especially when resources are limited. We propose that a decision-making framework based on Bayesian philosophy of science provides a basis for choosing which studies to replicate.