For a book, the local value of a page may not accurately represent the value added to the completed book. In fiction, a relatively boring page may be intentional as a kind of palate cleanser. In non-fiction, an early hard to understand section may make later sections easier to understand and avoid unlearning.

In programming, it may be challenging to distinguish between You Aren't Going To Need It (wasted work that adds complexity) and brillant foresight that avoids trying to retrofit an abstraction over independently developed code.

I am not sure that a manager needs to reliably judge the quality of the product to reliably deliver quality products. Judging skill and integrity (and promoting the development of such) does not seem to require having the same skill.