By pairing them with business people.
Creating a study or degree paper, as the current satisfactory output from scientists, is not what yields innovation, for most of those studies do not lead to the crucial and well-defined “last mile” of such discovery turning practical innovation.
To describe the difference is perhaps best illustrated by an exchange I had with a leading figure and friend of a well-known research institute in Silicon Valley when I presented him with a company doing something special, his rather smug answer being: we invented a similar thing 20 years ago. To which my response was: what did the implementation yield?
Not much apparently, as the scientists promoting the discovery were unable to apply the science to an application area where the imperfections of relativity (inherent to science) did not negatively, but instead positively affect the real-world application. My company became a success; the scientists’ approach died on the vine.
Statistically, it is also demonstrated that the discoveries of scientists run by scientists lead to very poor commercial outputs (I recall a percentage of less than 5%), and I make it a point to ensure scientists drop down a few notches on the totem pole of a startup company’s leadership for precisely those reasons. I learned from experience I am very wary of supporting any company with a scientist as the CEO.
Science is an incredibly valuable practice to challenge and expand the fringe of human discovery to uncover how much more there is to know about evolution and nature’s rule. And scientists have a fantastic propensity to deliver new and higher normalizations of truth, but rarely (exceptions Bose, Dyson) are those same scientists able to match a discovery of long with the wherewithal, investment-return expectations, and the reality of short.
So, embed scientists often and early with an army of savvy product management and product marketing people to deliver a bed those scientists can lay in with milestones you can hang your commercial hat on.