Failure is part and parcel of research, but discussing it sometimes seems to be taboo in science. It doesn’t need to be.