Fans of online games with a large, diverse cast of characters often construct an elaborate tier system for competitive play. This makes sense, as not every character will have the same strengths and weaknesses, and some will be more suited to the metagame than others. Some players also build tier systems for games that do not have online or competitive play for the same reasons. Some characters, skills, or abilities start very useful but are quickly outclassed by others as the game progresses, while others require a lot of effort and often a change of play style to become useful, eventually outclassing most other options (think Magikarp from the Pokemon franchise - TVTropes named this concept after it), and still others are rarely useful for the amount of effort it requires to get them. Depending on how difficult it is to reach that point, the amount of time invested for relatively little gain can discourage most players from using it at all.
Many skills in RuneScape follow this pattern, and more will for as long as Jagex continues to only update skills at the highest levels. Herblore is a notable example, as there is very little reason to actually train the skill until the mid 80s, when untradeable extreme potions are unlocked, at which point training becomes a necessity if the player in question wants to compete. On the other end of the spectrum, Smithing is very useful early in the game, when players can get a massive experience boost from The Knight's Sword and make level-appropriate equipment, but by the time that player will be able to use steel or mithril armor and weapons, it becomes just another money sink with only a handful of niche uses.
An unusual case is Magic. Training is very easy early in the game and it's fairly useful in combat, especially with the many sources of free runes available to a low level player. Later, it can't keep up with melee, but is still useful outside of combat for teleports and high alchemy. Then, it becomes useful in combat again when the player gets ancient spells, and ultimately becomes stronger than melee when the player gets access to Storm of Armadyl. Unfortunately, the cost of training for very little reward may discourage the average player from training it, and the fact that it only surpasses melee in combat after the player completes Ritual of the Mahjarrat - a grandmaster quest with rigorous quest and skill requirements - will prevent all but the most dedicated players from ever using the skill at its best.
In any other game, this would most likely be a skill that players would avoid in normal play. It does end up being one of the best combat options, but only very late in the game and with a lot of time invested into it. To briefly return to Pokemon, the reason the concept works in that franchise is because characters affected by it are not affected to the point where they are entirely useless until that point, and if they are, only for a short time.
The only reason this works in RuneScape is because players are encouraged to raise every skill to as high a level as they can. That is the goal of the game and the metagame is centered around making the process easier for the players. Furthermore, there are a handful of enemies for high level players that require magic. This gives the skill a niche, but that does not make it balanced - merely inconsistent. It encourages players to grind solely for the reward, rather than to use it in normal play. The same can be said for other skills: players grind for certain levels because that's where the rewards are, and often that reward takes players into the 90s.
Some skills do this well. Prayer offers rewards at regular intervals, and regardless of what your level is, there is something useful that you have unlocked and something to work for. At lower levels, players have combat boosting prayers, then in the forties, they get protect prayers. At sixty and seventy, players can unlock chivalry and piety, and they'll probably be able to unlock ancient curses for more rewards into the mid nineties. Combat works the same way, as does every skill that is involved in Dungeoneering.
Magic, unfortunately, is limited by the fact that it is a combination of non-combat and combat skills. With the exception of the ancient spellbook, not every level will help you in a fight. Similarly, its non-combat options are relatively shallow, outside of teleports, alchemy, and the Livid Farm lunar spells, which are only available through grinding. It's a skill that can do anything, but not well.
The Polypore Dungeon is a step in the right direction when it comes to improving combat magic by adding better equipment, since the skill's accuracy is comprised almost entirely of bonuses from magic armor and there were only a handful of sets that were more than cosmetic, while the Polypore Staff effect is a cheap alternative to regular spells, though this staff requires level 80 magic and (as of this article) attack. Similarly, Dungeoneering incorporates the skill very well, as a number of enemies are weak to it and items such as the magical blastbox and celestial surgebox allow players to carry over sets of runes to each new floor, in addition to combining the best aspects of both the standard and lunar spellbooks into one.
I understand that skills should become more rewarding as the player reaches higher levels. The problem with skills like Magic and Herblore is that they progress very slowly compared to most other skills, jumping from "useless" to "incredible" very quickly at levels that most players will never get. Quite frankly, Jagex's recent trend of adding only content requiring levels between 70 (quests) to 90 (skills) is a problem: If they want to know why people grind their levels as quickly as possible, they just have to look at where the content is, or more appropriately, where it isn't.