In some ways you’ve got to think
of your work as a portfolio of
experiments. Some of the high
risk ones will not work, but you’ll
learn from it.’