Just listened to this whilst eating dinner. I like his early point about how many data science projects fail because they fail to establish a clear value link. His points about pricing are similar to what we do but his are in a different context (insurance) and we're hoping to become much more of a turnkey solution.