As for the appeal of the "new," I think that's the worst thing about contemporary pop culture. What's "new" this week, or this month, or this year, isn't likely to be truly fresh or innovative or even novel (in the short-term sense); it's just whatever is being released, promoted, advertised, and/or "viral."
Someone else goes on to mention in the comments that the canonization process for culture is slow for a good reason. Most of the Sight & Sound votes aren't from the last 20-30 years because it takes at least that long to determine if a movie (or anything else, really) has a long-standing contribution to make to our culture. It has to outlast its own novelty, and that's not something that can be predicted.
This is why I squirm at labels like "instant classic" (an oxymoron if there ever was one). Not just because they imply a rush to judgment, but because they imply a lack of understanding of why a canon of classics in any category is assembled in the first place, or what it takes.
One of the fun things about reading literary criticism from even a few decades back is seeing what books were considered a big deal in their day, and how so few of those books have left so much as a footnote for the present day. Who today talks about, let alone reads, Grace Metalious's Peyton Place (god help you if you do), or John P. Marquand's Late George Apley, or John O'Hara's Ten North Frederick? A good many of those books vanished for a justifiable reason: they had no staying power. They were so much of the moment — not just of the issues of the moment, but the point of view of the moment — that they had nothing to offer readers further down the line once the noise and heat blew over.
It's been said the biggest reasons we still read Shakespeare and Melville are because we keep having them shoved down our necks in college by one generation of teachers after another. The unspoken implication is that there's nothing really there apart from tradition's rusty gears turning, which is demonstrably false. Good stuff survives at whatever level it can continue to an audience, and if the audience of thirty years hence doesn't exist for it in any form — academic or popular — let's not leave out the possibility that the work's sell-by date has long passed.
We also have a tendency to conflate the creator with his work — that one brilliant book out of a stretch of also-rans or never-wases constitutes a successful career — and I think we do that in large part as a way of padding out the canon with creators instead of individual works. The creator is always worth studying, but not for the sake of putting all of his output under glass and behind velvet ropes.