I don't think it's necessarily a matter of exploration vs. combat or leveling vs. level-cap content.
Heck, I might argue that all content needs just four things:
- Minimally divergent behaviors among players doing the specific content.
Such comes from both (A) internal factors such as in dungeon design and its different opportunities and (B) external factors such as player culture and the content's reward system and its adjacencies.
The more ways of play and reasons to play the content differently, the more conflicts you're likely to have in that content. This is not to be confused with having internally consistent behaviors that are distinct from those common to other content types; such is often precisely what can make a piece of content feel particularly novel (a strong factor in enjoyment to many), so long as depth and breadth of attractors aren't thereby too reduced.- Relative sustainability of form.
If something is meant to be centered on exploration, and yet is meant to be played through some dozen, it must still be centered on exploration even on that dozenth time, even if it trades some aspects thereof for others (such as optimization).
For instance, if Haukke Manor's keys were meant to be an element in and of itself, rather than a sliver of extra flavor in an otherwise standard dungeon, they wouldn't likely drop in the same locations each time and, accordingly, the dungeon layout wouldn't be effectively linear outside of its single obligatory Return and they keys likely wouldn't be visible as directly named interactable. In its current design, that "exploratory" element isn't much even in its first run, and nonexistent thereafter. The same, of course, can be said for every side-room in any existing dungeon.- An enjoyable and applicable concept.
This is the most straightforward. If there's nothing to the content's form people want to do, no particular itches it scratches, etc., the content will be useless unless forced (and arguably worse, even, if it is forced).- Commensurate rewards.
Again, this much is obvious, but keep in mind the first criteria. However, it doesn't seem particularly obvious to XIV when we consider how roulettes rewards—and to a lesser extent, content rewards in general—have been handled. System-level changes (especially to the currently flat roulette rewards, which will tend to seemingly favor the shortest possible content available to a roulette since they skew rewards-per-minute in their favor, especially when the queues themselves are shorter) would help a lot, but the content, too, needs to consider the uniqueness and appeal of its rewards and the efficiency by which they are acquired relative to competing contents (especially those that do not have nearly the same, or far less total, intrinsic appeal).
But, we've seen quite a few ideas like this over the years. Sandpark's Frontiers, Shougun's Airships, alternate Exploratory Missions, dungeons with non-linear routes that can't nonetheless just be memorized and thereby optimized, some—frankly, far better—concepts of Rogue-like content, etc., all come to mind. ...I just wish I still had links to some.