I am of the belief that the "higher" the concept, the shorter the campaign will be. Therefore, if one wants to run a long campaign, a "low" concept is necessary. A "high" or "tight" concept is one that contains a lot of information about what the campaign will be like, while a "low" or "loose" concept does not. There is not a stiff border between the two, but here are some criteria that I use to judge whether a concept is "high" or not:
1) Does the pitch contain a reference to a specific television show or movie which the DM is trying to imitate?
2) Does it look like we will have to spend a long time between character creation and play sorting out what each person is and how they relate to one another so that the premise won't collapse in play?
3) Do I expect that this is the kind of game in which someone will say at some point "But that's not the kind of thing that happens in [the source that inspired the game]"?
4) Is the pitch longer than one short paragraph?
5) How tightly is the premise going to determine what happens in game?
The basic problem with a high concept is that it contains too much. I have seen pitches or explanations of games that contain within in them the characters' origins, their relationships, and the specific kinds of challenges they will overcome. I find these pitches become less interesting the more detailed they are, as they close off more and more possibilities to engage with the story or world with each word and sentence.
A narrative that is closed, that has a finite universe of possibilities and options which can be fully explored and then exhausted, is obviously one which should end, as it has been completed. The fewer options and possibilities PCs have, the faster the narrative is exhausted. This can be useful if one wants to run a one-shot, or a short game, but trying to drag a game out once this has happened is a recipe for boredom.
In contrast to this, a long campaign relies on underdetermination. Underdetermination is the use of ambiguity, polysemy, surprise, ambivalence and mystery to provide PCs with just enough information to make meaningful choices in their situation while avoiding or obscuring a total comprehension of everything that is going on. You need the occasional intrusion of the unexpected, unexplained, unknown, and unrelated to refresh and expand the narrative.
Underdetermination also pushes PC agency to the forefront. If they don't want to be Power Rangers defending the earth from alien invasion, then they don't have to be Power Rangers defending the earth from alien invasion. Underdetermination means that how the PCs engage with the world is not determined for them, but something that emerges organically in play, and that is subject to revision and change through their choices.
A low concept campaign works especially well when you have a detailed world with lots of possible origins for PCs, places for them to go, NPCs to interact with, and challenges to overcome. You have to avoid having a world so detailed that engagement with it requires the PCs to know a ton of information beforehand, which basically just replicates the same problems as a high concept game through a different means. Or a metaplot that crowds out the PCs and turns the setting into a bunch of high level NPCs jockeying for control.
While I have a lot of respect for Glorantha, I have very little interest in playing in it, since most of the modern adventure concepts I hear people pitch are just a mish-mash of references to canon that convey an attitude of "Let's enact the timeline!" This is something I try to avoid when I write for the Dawnlands, not always successfully, but I want to leave it open enough that individual referees feel that they can "own" it and alter it as they please while still feeling like they are running a "Dawnlands" game.