will target reduction in output & training review from 1% to 0.1% and claim not reviewing things means it's improving.