Graphics are important...but not in the way most people imagine them to be. The important thing about graphics isn't how preety they are or how many effects they have. It is whether or not they are
functional for that specific game. By that, I mean if they are potraying all the visual information the player needs in a clear manner and are not detrimental to the gameplay or the player's enjoyment in general.
Of course, the catch is that graphics do not need to be top-of-the-line in order to fulfill such a task...heck, even games on very old systems like the
Famicom/NES or
Commodore 64 were quite capable of having sufficiently functional graphics. Yet, very often, developers become so infatuated with technology that they end up focusing too much on that aspect, which results in the gamer being overserved in that aspect and underserved in others.
This isn't to say that it hurts if developers dedicate at least a bit of time to making the graphics pleasant to look at. However, nowadays, developers invest way more time, money and manpower into that aspect than it is truly necessary.
If I was a head of a game company, I would do what Iwata did to the developers of
Brain Age and give them at least one project which would have to be ready for demonstration in 90 days (well, more or less; it would depend on what kind of game). Not only would they not have the
time to argue with me, it would force them to strip down graphics to the most essential components and focus on other elements in order to have it ready.
...That example makes me sound really mean, doesn't it?

But then again, being able to work under limitations is what seperates a good developer from a great developer. And afterall,
Brain Age would go on to sell millions.