Phase transitions can influence mantle convection, inhibiting or promoting vertical flow. We are motivated by two examples: the post-spinel reaction proceeding via akimotoite at cool temperatures and a curving post-garnet boundary. Some have suggested these could change mantle dynamics. We find this is unlikely for both reactions: the first due to the uniqueness of thermodynamic state and the second due to the low magnitude of the boundary’s  slope in pressure–temperature space and density change.