I knew I was going to love this job here,” says Anthony Lake, Unicef's executive director, pointing at his heart beneath his suit jacket. “But,” he points to his head, “it is fascinating here too. There are no boring bits.”

Aid has always been heart-led. Charities tried to move donors by showing pictures of starving children. Donors sent money. Whether the money helped, nobody quite knew. The only thing in aid that was habitually measured was the input: the amounts donated. The output - lives saved or improved - almost never was.

Yet since Lake took over the United Nations' children's charity in 2010, the amiable American has watched the so-called “evidence-based revolution” in aid. Finally, academics, donors and even some aid agencies have begun measuring what works. Very slowly, development is becoming a science.