Okay, so here's what's going on wrt presidential elections.
The US electoral system for presidents means that only about a dozen states (those near 50-50 like FL and PA, or with recent strong swings like WI and MI) could even plausibly matter. And then within those states, only the voters who could plausibly either switch sides or decide whether or not to turn out actually matter. So a minority of a minority of all voters is what determines who actually wins. I certainly don't think this is a good setup but there is no chance of changing it by 2020.
These voters are disproportionately low-information (high-information ones have much more rigid political allegiances and high probabilities of voting). "Full swing voters" - those who could change their mind and vote for either party or stay home - in swing states are disproportionately white and working-class, which is inconvenient for the Left as it's currently construed but could be won back with a good candidate. Such a politician would be good at messaging to these voters. They would also, ideally simultaneously, be good at turning out minority and young populations in these states, many of whom are half-swing voters: they may turn out or not, depending on the level of engagement the candidate has with those populations, and vote overwhelmingly Democrat when they do turn out. Voter suppression is targeted at these people - politically active young and/or minority voters will still vote, but adding some further hurdles drives down turnout among the less engaged - those whose decision is between voting and not. The things that unite both full-swing and half-swing voters are disproportionately low socioeconomic class and low political engagement.
Obama appeared to have a "Blue Wall" because he was good at both of these tasks, easily winning the Rust Belt swing states on a combination of full-swing mostly-white voters and half-swing young and minority.voters. Hillary Clinton managed to lose those states because she was very bad at both of these tasks, suffering historic losses among full-swing voters in swing states and not turning out young and/or minority voters at the same time.
Most current Democratic Party politicians are much more Clinton-like than Obama-like in that they're also bad at winning plausible full-swing and half-swing voters, which is why they keep managing to pull defeat from the jaws of victory at all levels, even in the face of such a farcically bad candidate as Trump. It's true that some half-swing Republican voters were induced to vote by some combination of Trump's racism/sexism/etc and by his message of tearing down the current system. Those voters are numerically smaller but still enough to provide a margin of winning (edit to add: in an election as close as 2016).
I'm agnostic about whether such a candidate should be a full-throated economic populist - like a Sanders 2.0, now rebooted to be younger and and appeal better to minorities, or just a standard Democrat who is unusually charismatic and good at winning full-swing and half-swing voters - I call this Obama 2.0. It is plausible that either one could win.
What we definitely do not want is Clinton 3.0, or anyone who reminds people more of H. Clinton, or Schumer, or Pelosi, or Reid, than of either Obama or Sanders. There's some sort of delusional belief among these types of people (the "establishment Democrats", minus the few Obama-like ones with actual charisma) that they can swing enough upper-middle-class traditionally Republican voters to their side to actually matter on the basis of disgust with Trump and, even more delusionally, disgust with other Republicans. They also tend to overestimate demographic trends on short timescales. Yes, NC and GA and AZ, and behind them TX, are all trending Democrat ATM. But they're still more Republican than the nation on average, and the trend is slow enough that they will do maybe a couple percentage points better (compared to national vote swing) in these places 2020 than in 2016, and moreover the trend is flat in FL.
The only real ray of hope for these types of people is that the Sun Belt might eventually flip to them sometime in the 2024-2036 timespan, unless of course demography isn't really destiny and they manage to lose a compensating number of multi-generation Hispanic and/or Asian-descended voters, which actually did happen in at least some places in 2016. Also, they still have a whole lot of Senate seats to lose in Rep-trending states, as the 2018 map shows all too well. They are pretty close to rock-bottom from an EV perspective in the Midwest and Northeast though, with only 16 EVs left to lose (MN, NH, ME-statewide). But they're definitely not poised to claim any sort of advantage right now, and we'd best get the gerontocratic Third Way types to retire and replace them with younger, more charismatic, and probably more economically populist candidates.
