Recent molecular divergence estimates for placental mammals recapitulate their explosive interordinal diversification inferred from the Paleocene fossil record, but place this event some 10-20 million years earlier than apparently observed in the fossil record. We show that current models of molecular evolution do not adequately account for parallel rate changes, and result in dramatic divergence underestimates for large, long-lived mammals such as whales and hominids. Calibrating among these taxa has shifted the rate model errors deeper in the tree, inflating interordinal divergence estimates. We employ simulations to understand how such rate errors interact with fossil calibration strategies. We find that focusing on calibrations for taxa that retain “primitive” life-history characteristics substantially improves molecular dating accuracy. Applying this strategy to the empirical data favours the traditional palaeontological interpretation – a few Cretaceous placental lineages gave rise to an explosive diversification following the 66 Ma Cretaceous-Paleogene boundary mass extinction. This scenario closely aligns with a growing consensus for a similar model for bird evolution. Looking further back at the origin of mammals, we combine morphological and molecular data to merge fossils into the modern mammal tree. As a remarkable forerunner to the later placental diversification, we find that monotremes and therian mammals originated during a rapid adaptive radiation associated with the Triassic-Jurassic mass extinction event. Ancestral state reconstruction and lineage survival patterns suggest strong selection for traits enhancing endothermy around this origin of crown mammals.