a) Is it working? If not, why? b) Do you think you know how to do it better? If yes to both then do it. If no to a) then you'd want to be fairly confident that

a) Is it working? If not, why? b) Do you think you know how to do it better? If yes to both then do it. If no to a) then you'd want to be fairly confident that they got it wrong in some way and didn't just prove it to be a bad idea.