On the Struwe-Jeanjean-Toland monotonicity trick