> The author admits the difficulty in measuring it and that we may have to rely on "non-scientific" measurements.

But that's the whole reason why we reward outcomes in the first place. If it was possible to reward only "well-directed" effort regardless of outcomes, we'd be doing that already!