At the time of writing, it seems that Joe Biden is on track to clinch the presidency. But, the blue wave did definitely not happen and the polling error seems almost bigger than in 2016. The popular vote currently stands at 50% – 48%. This will likely change, but we are far away from the 7%-lead that the average of nationwide polls showed. In addition to the close presidential race, Democrats are poised to lose seats in the House and will fall short of recapturing the majority in the Senate. Look at South Carolina or Maine where Democrats did not even come close to beating Republican incumbents. And all this happened while Democrats were literally burning money. My thoughts:

– If I compare the average of the polls with the actual results, it seems to me that in many states, it is particularly the Trump vote share that is different. In Florida and Ohio, for example, Biden's vote is fairly close to the polling average. It's the share of the Trump vote that is different. Either people lied and there indeed exist the "shy Trump voters." As one Republican pollster argued all along: People lie to their doctor, their spouse and tax authorities. Why would they honestly tell a stranger over phone how they vote? While that guy seems really smart now and has my full attention, it is also possible that undecideds went overwhelmingly for Trump. In the past, undecideds more often went for the challenger. But listening to Luntz's focus groups, it always did seem to me that voters who self-identified themselves as undecideds were actually leaning Trump.

– An election with an incumbent is foremost a referendum on the incumbent. While I think that that is still true, Democrats obviously underestimated the appeal Donald Trump has on half of the electorate. Democrats then built their campaign on fundamentally wrong assumptions, yet again. Good public opinion research and sound political marketing has a lot to do with listening, and careful focus groups should have revealed that appeal as well as the weaknesses of Joe Biden. I tweeted some weeks ago: Turn off cable news, stop looking at nationwide polls, and pay a lot of attention to people who don't pay attention. Or put differently: An opposition party that merely banks on the vulnerability of the incumbent party, no matter the screw up of the incumbent party, is often overestimating its chances to win.

– The Republican party invested a lot of resources into their Get-out-the-vote operation (3000 paid staffers, 2.5 million volunteers). Democrats did not do that until late in the game because of health concerns. Nothing beats door knocking. A party needs a ground game.

– "It's the economy stupid," James Carvilles famously said some 28 years ago. Joe Biden and his party have done little to build up economic competence during the campaign (do you remember anything in that respect from the convention?). I will never forget a focus group respondent after the first debate say that while Trump's behavior was despicable, "it does not affect my bottom line." The economy does. It's the economy, first, last and forever.

– Candidates matter. Let's be honest: 77 is not the ideal age for an American president. By saying this, I don't want to sound agist. The problem is not the age itself, but that Biden sounds, acts, and behaves like a 77-year old. A 25-years younger Joe Biden would have won this thing more convincingly.

– The minority vote is all but monolithic (see also the Michigan Senate race, no matter the final outcome). If Republicans put up a challenge for the minority vote (as did George Bush in 2004), Democrats must fight back. In that respect, it is also noteworthy that in many cases where Republicans were able to flip a House seat, it was done by a female or minority candidate.

– There was a record number of voters who voted early and/or absentee. Biden won those overwhelmingly in almost all states across the USA. Votes cast on election day, on the other hand, went overwhelmingly for Trump in almost all states. Pollsters and campaigns must learn the lessons and adapt to the new reality.

– If there is herding among pollsters, meaning that they copy each other and base their raw data on the same wrong assumptions, taking the average of polls doesn't make the data better.

Making the long story short, I'm always surprised just how stupid I was two days ago…euh, meaning to say, I'm impressed as to how much I can learn in 48 hours.