Isn't the perfect picture of balance in a PVP game equal distribution of usage?
No. Or at least not quite. To a degree, you do want to approach that, but there are other considerations.
For one, in a CCG like this people won't have an equal distribution of cards so there won't be an equal number of people running a given deck simply because some decks will inherently require more investment (either of time, money, or both). Handlock is fairly expensive, Warrior Control is even more expensive, and Zoo is really quite cheap (though some more expensive variants exist). This and the prices of other decks help to influence distribution into a more uneven state, but is not necessarily indicative of an actual balance issue.
Secondly, it depends on what format you're referring too. For example, the frequency of deck usage changes based on rank (and to an extent, the time you play as well). Watching Trump's stream the other day, he had been playing his Token Druid deck, and I think out of the 10 or so games I watched, only 2-3 of his opponents weren't playing Druid.
Now, obviously, that sample size is WAY too small to really be significant on the larger scale, but it does show how certain deck-types might be more popular among specific sub-metas. Zoo, Aggro Warlock, Aggro Paladin, and others tend to be far more popular among lower ranks, based on testimonies I've read and some other streams I've watched. Part of this is the aforementioned "expense" issue, since lower ranks tend to have more people that have invested less into the game due to how acquired skill and experience will generally assist someone's climb in the ranks even if they didn't also improve their deck. But part of it is also that that bracket has developed its own sub-meta. It's influenced by the overall meta (and influences the overall meta to some degree) but it has it's own mutations.
If we look at tournaments, the most popular tournament format is 3 different decks with 3 different classes, or team-based games where players on the same team will often have different decks (though there are some repeats). There are decks that are specialized to take out expected other decks, to work as counters to a popular strategy that that player might otherwise have trouble with personally. It's sort of like using a Sideboard in MTG in that regard. But those Counter decks won't necessarily be evenly distributed. Perhaps the others players don't have that issue with whatever deck the counter is there for. Perhaps they simply prefer to power through using their other options, etc. That doesn't mean the game is unbalanced, but rather that the game is flexible enough to have options and variations that account for individual play-styles and strategies. Even if some specific cards/decks see more use, it is simply the current "general purpose" card/deck for the current meta. As long as that doesn't become stagnant, it's not a problem. As long as there is still room for personal variations, and to apply different strategies and play-styles, the game is very likely well balanced even without even distribution.
Of course, tournaments are by far the minority of constructed play, but in theory a game's balance is best measured on that sort of playing field, as supposedly the people playing there understand the game well, will be playing competitive (rather than casually), and so on. On the other hand, it doesn't mean that casual play balance should be completely ignored, but because there are some strategies which are inherently more difficult to counter without as great an understanding of the game, or without having put in the proper investment to get the cards/decks that counter it, there can be a theoretical imbalance. For example in League of Legends "Master Yi" has on-and-off been considered a "pub stomp" champion. He's generally considered "balanced" or (more often) even below average at higher levels of play, but his play-style and mechanics can be difficult to deal with if you're playing as an uncoordinated team, or simply lack the experience/understanding to properly deal with it.
If Yi is "balanced" for low-end play, he'll be even more out of balance for higher-end play, and effectively removing the option completely if they aren't careful. Though there is no team-play and a smoother learning curve for Hearthstone than League of Legends, many of the principles still apply. What is "balanced" at higher level play might not be at lower level play and vice verse. Since Lower-level play is more common than higher level play (particularly for a free-to-play game) that means that total distribution (and even win-rate etc.) is different compared to higher level play where (supposedly) the balance of cards/decks can be measured more objectively with fewer issues of player error or investment capability impacting the data.
Of course, for Blizzard, the health of the game, and ability to retain players is important, so they shouldn't forsake lower-end play entirely or anything like that, but it should be the goal to have players learn how to deal with given strategies if they are only problematic to inexperienced players; and of course to provide more options to those players that don't require as large an investment (for example, better access to Common/Rare cards that could compete more with late-game Legs and Epics would open up that playstyle to newer players, and introducing more common tools for certain classes to deal with a given strategy in general can also help, etc.) But large or drastic changes should likely be avoided unless there is a proven substantial balance issue at all levels of play.Above all though, perfectly even distribution is not a sign of something being well balanced, as some play-styles are simply more popular than others.
There is not an even number of Timmy, Johny, and Spike style players that play the game.