Not only do we have nation-wide polls on Clinton vs. Trump already, we are starting to get the usual slew of state-by-state polls. For right now, ignore the latter. Opinions and vote intentions don't tend to crystalize until fall, when the campaign crystalizes them. That doesn't mean we learn nothing by looking at polls, it just means we can't read too much into them yet. And state-by-state polls are harder to do because of potential turnout differences that wash out in the nationwide averages. So, as problematic as it is to do a nationwide poll right now, state-level polls are harder.
And remember, nearly every time, the electoral college victory will go to the winner of the popular vote. So, by watching the nationwide polls, we get a sense of where things stand now. As November approaches, we need to shift to the state-by-state polls, where the fine-grained distinctions can be made, and where they become important. Popular vote/electoral college splits are rare, but obviously, they do happen. When? In close elections! And this one will be close.
This matters now because if you look at the RealClearPolitics nationwide polling average, you see what is essentially a tie between Clinton and Trump, but the state-by-state polls have Clinton ahead in more battleground states right now, which makes her lead look more solid than it actually is. Don't trust those. For now, recognize that the polls themselves are of less value than they will be in a few months. If you must look at polls right now (and of course, I must), focus on the nationwide polls. Shift your attention to the state-by-state analysis around September, or August at the earliest.
Or, hell, obsess over every poll you see. Who am I kidding? That's what I'll really do, regardless of the advice I give.
I'm not well. I need help.