So really the choice comes down to the problem you want to solve: if it's easier to sell to reviewers do it in academia, if it's easier to sell to customers do it in industry.
(* if not the detail - non academic impact is worth quite a lot in uk academia nowadays, and you may at times measure that through e.g. sales figures, but these in turn must ultimately be sold to reviewers)
Speaking as a former tenured professor, I can say this is no longer very accurate. What would be more accurate is that your customers are grant review panels, and your income is indirect funds on grants (at least in the US, and increasingly elsewhere).
