Read: the Democratic path to electoral success in 2018.

After every election, debates over that election's meaning begin to consume both the winners and the losers.  Few people actually change their minds, regardless of the outcome or their previous predictions.  Few people ever walk away from an election and think, "Most Americans disagree with me."  Instead, they tell themselves that they lost because of deceit (the people were conned by lies and manipulated by powerful forces -- the Koch brothers or George Soros), messaging (we had the wrong spokespeople or the wrong campaign strategy, but the informed people actually agree with us on issues), or uncontrollable events (war, terrorist attacks, recession).  No movement or group is immune to this.  Indeed, the very premise of this blog is to reassure readers of what I perceive to be the positive realities that underlie a very negative state of affairs!

This tendency reveals a flaw in one of the essentials of democracy: political parties and movements tend to react to demographics rather than electoral feedback.  Elections are blunt instruments, passing judgment on people, policies, and parties all at once; in their aftermath, we see modern-day reenactments of ancient divination, as politicians and pundits grope around in the guts of a sacrifice to try to interpret the will of the holy American People.  Polling makes this easier, since people accept individual polls as factual representations of the will of the nation -- even though a poll's wording, methodology, or timing might completely change its outcomes.

Given this, there is no certain answer to the questions, "What do the people really want?" or "How do we appeal to voters?"  Different people have different arguments, instead.  For right now, they mostly revolve around the degree of obstruction that Democrats should present.  Should it be total?

In "Democrats’ Best Bet for 2018: Accentuate the Negative," Ed Kilgore argues in New York that the best chance for success for Democrats lies in obstructing Trump and the GOP as much as possible while still appearing reasonable and approachable:

It is true that midterm negativity can be taken to counterproductive lengths, as occurred in 1998 when the GOP effort to impeach Bill Clinton backfired pretty badly. Four years later, Democrats struggled even more in a midterm, partly because they alternated between attacking and agreeing with George W. Bush in that security-haunted era, and partly because 9/11 gave Bush a level of extraordinary popularity that had only partially worn off by 2002 (his approval rating on Election Day was 63 percent).
By and large, though, the presidential party loses ground in midterms, and the opposition party can enhance the natural backlash by piling on, not by offering alternative courses of action that most voters will simply ignore.

In "The New Party of No," Michelle Cottle, writing for The Atlantic, essentially agrees.

While nobody does obstructionism like McConnell, House Democratic leader Nancy Pelosi has a keen appreciation of the value of just saying no. In early 2005, a freshly reelected George W. Bush was hot to overhaul Social Security. Republicans held the White House and both chambers of Congress. They had the vision. They had the numbers. What could possibly go wrong?
Everything. Even thinking about messing with entitlements is politically fraught, and Pelosi decided to make Bush bleed for it. For months, she kept her troops focused on beating the bejesus out of Bush’s proposal, driving public support for it down, down, down. The negativity became so relentless that some Democratic members began to get twitchy. When would the caucus move beyond slamming Bush and put forth a reform plan of its own? Famously, Pelosi’s stock response became: “Never. Does never work for you?”
By spring, Bush’s plan was deader than disco—and stayed dead despite his efforts to revive it, which ran right through the 2006 midterms. That November, Democrats retook both chambers of Congress in a wave election that startled pretty much everyone.

In the New York Times, Will Marshall suggests in "Why Democrats Should Work with Trump" that now is the time for Democrats to try to strike a deal on infrastructure with a deal-hungry White House -- in exchange for major concessions.

Unlike depriving millions of Americans of health insurance, revamping America’s outdated tax code and modernizing our run-down infrastructure are progressive causes Democrats should be for. And unlike Republicans, whose ideological rigidity and strident partisanship often border on nihilism, Democrats still hew to the quaint notion that the people elected them to solve problems, not prevent them from being solved. McConnellism is not in the party’s DNA.
But if moderate Democrats are disposed to cooperate with the deal-maker in chief, they ought to exact a high price. On tax reform, for example, they should insist that Mr. Trump deliver tax relief to the middle class, not the wealthy, and that he jettison Mr. Ryan’s proposed border-adjustment tax, which would hit consumers and business with big price hikes. The administration needs to find better ways to pay for a sharp reduction in the corporate tax rate. Democrats don’t have to love big business to recognize that our antiquated tax system forces companies to pay much higher taxes than their overseas competitors. That makes American workers less competitive and gives our companies incentives to move investment abroad — and keep profits there — to avoid the higher rate.

Jonathan Chait of New York strongly disagrees in "Delusional Democrats Want to Prove They Can Work with Trump," as he consistently has, arguing that the electoral history is clear: voters reward presidents who get big accomplishments finished in a bipartisan way, and any proponents of a Democratic deal need to recognize that hidden cost: Democrats will not only be conceding whatever concessions they give up to meet in the middle with the White House, but they will be badly damaging their chances of victory in 2018 and 2020.

Trump also seems to grasp that infrastructure gives him the chance for a high-profile bipartisan deal. Describing his “beautiful” relationship with leaders of both parties yesterday, Trump boasted, “We’re about to make a big deal.” The infrastructure bill would provide the sweet bipartisan coating to the bitter pill of his otherwise orthodox right-wing agenda. Democrats think they’re calling Trump’s bluff, when in all likelihood, Trump is calling theirs.
...
The Democrats’ most deeply mistaken premise is their belief that the voters will make specific judgments about which congressional party is more effective. That is simply not how voters think. They make judgments about the state of the country, and hold the president and his party accountable. They do not separate their judgments about the congressional party from the president — indeed, they are so single-minded in their focus on the president that voters actually base their vote for state legislature on their assessment of the president.

For what it's worth, I agree with Chait.  Any bipartisan bill needs to be spectacular for Democrats to justify giving bipartisan cover and a major achievement to the Trump agenda.  It shouldn't be ruled out, but no bill that seems like a prima facie fair compromise is remotely worthwhile.  The enormous, electoral advantages are a hidden concession that needs to be considered.