My number-one recommendation is to invest in people. Humans that are well trained are the inputs into this discovery process. And there's big opportunities still, I think, to do a better job of investing in people.