Very interesting discussion. I think the most important thing to recognize is that there isn't a single correct % for all situations. Even if you aren't varying your size based on the strength of your hand, which I think you generally should not, there are a number of factors that should affect your size:

Effective stacks – with shallower stacks, the concern you note about not building a pot with your nut hands is less of a concern. Consequently, there is less of a cost to using a small size with your whole range.

Shape of your range – what ginger says about betting bigger with a more bluff-heavy range is most applicable in polarized range vs bluff-catcher situations, which occur most frequently on the river but can happen on earlier streets as well, especially on static boards where you have a clear range advantage (222 or AA2 flop when you are UTG raiser, for instance). This is really only applicable with quite deep stacks, though, because in most tournament situations with shallow stacks you don't gain from overbetting nut hands, which won't have trouble getting all in anyway.

Board texture – static boards usually call for smaller sizing, because there are fewer protection considerations and position matters less.

Position – all other things equal, you should probably use slightly larger sizing (and bet less often) from OOP. This is because your opponent needs better odds to call with the same hands when you have position than when you don't.

Some other factors mentioned here, like number of opponents, have (I think) a bigger effect on whether you should bet a given hand than on how much you should bet when you bet.

C bets usually don’t “mean” much of anything, as many players will do it with the majority of their ranges. The more valuable thing to consider, then, is what it means when someone does NOT c bet, as this is usually the much more defined and exploitable range.