Rethinking Data-Driven Campaigns
I became a political pollster because of my love of data and my passion for politics, including Democratic and progressive causes. Both of those things took a pretty big hit on last Tuesday and I, like many, have done some soul-searching since I learned Donald Trump was going to be our next President.
I am confident in the analytic ability of the polling and political data community to understand what happened Tuesday night, why it happened, how to improve upon it, and to build better tools in the future. I know I plan to spend a lot of time on these topics because we should learn from our mistakes and improve upon our methods.
There will be a temptation for some to say that all polls are garbage and we should ignore them forever. I believe that would be a mistake. Yes, the polls missed. Yes, we need to figure out how to improve our predictive tools. And while I suspect one of the conclusions may be that elections are uncertain and there is no perfect tool to predict the future, public opinion research will continue to play an important role in understanding the world we live in.
We are going to revamp polling, but that is not what this column is about. I have been thinking more about a bigger question on the role of data and research in modern political campaigns and specifically within the Democratic Party. I fear that we have become too focused with data-driven tactics and lost sight of the original point of research — which is to understand the electorate. What motivates voters? Why do they vote in certain ways? What resonates with them?
Take the issue of turning out Democratic voters. I believe that, too often, Democrats treat turnout as an endgame tactic — something that we address at the end of the campaign in Get-Out-The-Vote operations. We forget that we actually need to give a reason for voters to show up. A great ground game or social pressure mail may be proven effective at squeezing out a point or so on the margin, but neither provide voters with a substantive rationale to go to the polls.
Donald Trump had a message and it resonated with his voters. If Democrats want to be successful in the future we need to spend more time figuring out the right message to motivate our voters. The best minds and leaders of the party are going to have to provide vision, ideas, policy, messaging, and more.
And yes, to figure that all out, we are also going to need research. We will need to conduct polls and qualitative research, we will need to explore new social media data and techniques, and we will have to employ more advanced techniques like predictive modeling and experimental design. I still believe in all of these tools and think they are fundamental to our future success. But we need to do more to consider why we are doing them, what purpose they serve, and how we can use them to inform our overall political strategy.
I’m not suggesting that every successful message must be poll-tested (Trump may have disproved that on Tuesday), but we need to spend more time studying the American electorate, especially listening to our own voters, and understanding what makes them tick. We need to understand how people are feeling about the economy and how they are responding to the Trump Presidency. That will be essential to rebuilding our party and regaining what we lost on Tuesday night.
What I do know is that the Democratic Party needs to avoid the classic Moneyball fight. In the classic baseball story, the quants fought with the scouts for primacy and control. In the end, the best franchises employ both quantitative experts and subjective evaluators and successfully weave in the expertise of both sides. In our party, we will need to do the same thing — we’re going to need the data minds to work with those who have the ideas and the makings of a successful message. But the ideas will need to come first.
This post can also be viewed on Medium here.