The campaigns (and other organizations) do scientific state-specific swing-state polling. The national stuff is important for sentiment, but its the handful of states that decide the election that are seeing numerous polls in the field continuously. That said, no one like to vote for a loser, and seeing your candidate behind significantly nationally can hurt turnout. What's interesting to me is that, (IMO) Trump won because people basically wanted to blow the whole thing up, and turned out to lob a grenade. With Covid-19, and Black Lives Matter, Biden could get the vote out as people of all races are very passionate right now about the pro-equality agenda in the cities. I saw a stat (I am paraphrasing) that if 5% of the African-American voters who stayed home in 2016 in any one of Milwaukee, Detroit, etc had turned up, the election would have flipped.
FYI: people who follow this stuff carefully for financial reasons (or other) know that pollsters have adjusted their sampling and weighting procedures to correct for some of the problems that occurred in 2016. Not saying they are ironclad, but they are supposedly much improved.
Let me put it this way... if Trump improves by 5 points (that's huge) in every single state, Biden will win handily. So contrary to all the supposed improvements made to the polling, Trump needs a big, big swing to get back on top. He's in real trouble in MN, PA and MI right now, of course NC, AZ, WI, OH and FL are in play, and nationally he is in double digit disapproval land.
You can't see the internals here, but this is good way to keep up on the public polling:
Latest Polls
www.realclearpolitics.com