The tech behind campaigns: “It shouldn’t only be a bunch of nerds in the basement that can do this”

Raffi Krikorian is the Chief Technology Officer of the Democratic National Committee, where he is focusing on building foundational technology at the DNC, empowering the technology ecosystem around the Democrats, and equipping candidates up and down the ballot. Prior to joining the DNC, Raffi served as the Director of Uber’s Advanced Technology Center.

Mikey Dickerson is the Executive Director of the New Data Project, which builds better technological tools for civic engagement and strong democracies. He formerly served as the founding administrator of the U.S. Digital Service.

Given their combined experiences on the 2008 and 2012 Obama campaigns and now with a number of Democratic House races, we sat down for a three-way conversation on how technologists can make an impact on campaigns and how the state of campaign technology has evolved over the past ten years.


How do campaigns use technology and their technologists on staff?

Raffi Krikorian (RK): In campaigns, I’ve seen a conflation of the terms technology, digital, and IT. A lot of our time is spent explaining that technology is not IT. Most of the time we talk about technology, we’re asking how it can boost the efficiency of the work that we’re already doing, but people rarely ask how we can use it strategically to transform the way we run campaigns. We try and bring people to a new mindset for looking at how to make data driven decisions or rapidly experiment on the ground. We’re trying to normalize technology, yes, but also the mindset for trying new things quickly, failing fast, and then trying the next thing.


That seems like a huge goal to shift that mindset. How do you do that? Are there specific tools or tactics you are using?

RK: If you’re looking to actually build impactful technology, it requires a lot of upfront work— talking to people, building prototypes, doing test appointments in the field, analyzing results. I just described a six month process. But too often, campaigns want to focus on the next thirty days. If you want to run a real technology strategy starting today, your target is the 2019 election at the earliest. Not the 2018 election. Technologists who want to go help right now should go knock on doors or work on a social media strategy. But if you want to help in the long term, we need to think ahead.

Mikey Dickerson (MD): It’s not going to change at the level of the local field office, in my opinion. The phone banking and door knocking very often have no measurable effect and sometimes have a backlash effect. But it’s what campaigns are going to do and they’re not going to stop doing it. If we want to have any success as a party across the board, those structures need to be built somewhere that doesn’t rely on campaigns because they only exist for a few months and then they’re gone again. We don’t have a good home for that kind of work right now.


Are phone banking and door knocking effective in the current moment?

MD: If you go knock on five doors that are supporters for your candidate and encourage them to show up and vote in the fall, that will have some very small, probably less than one percentage point effect on the likelihood of those five people voting. And now you’ll knock on the sixth door and you had bad data or you made a mistake and somebody with a Make America Great Again hat answers the door and sees you with your Hillary blue T-shirt and clipboard. You have now dramatically increased the probability of that person voting because they feel threatened. There are similar results with phone banking.

I think we’re in a hangover where people are realizing one by one that they need to make changes but we don’t know yet how to tell ourselves a different story as a group. The Instagram of each candidate right now will still be full of people renting vans and driving to neighborhood and knocking on doors—because that’s what we know campaigns to be. It’s what our parents did. Somebody’s got pictures of their mom campaigning for whoever in 1964. We haven’t redefined the cultural experience of campaigns.

RK: We know a lot about what works with campaigns, but we don’t know what the right mix of tactics is. Some campaigns want to stop running television ads and go fully digital, but is that swinging the pendulum too far? We’re all trying to answer the central question: if your campaign only had $1,000, how would you spend it to get the best return on your investment?

MD: One thing we do know is that friend to friend relational organizing has been a ray of hope showing strong results.


As technology and data-driven decisionmaking become a larger part of campaign strategy, how do we sustain and spread that knowledge across varying campaigns?

MD: The 2008 and 2012 Obama campaigns hired a lot of in-house tech people, so they amassed all of this knowledge and intellectual property that became property of Obama for America (OFA). If you try to go looking for that knowledge even two years later, there is no one to answer the phone. OFA has been reconstituted into a different corporation and nobody knows who has the rights to any of that stuff. Other places pick up the slack, like Civic Analytics, which tried to institutionalize some of what was learned in 2012, and capture some of the talent so it didn’t just disperse to the four winds immediately.

RK: I think 2016 was a little bit better. At the DNC now we have all of the technology and data that the Hillary campaign built out for 2016, and we strategically hired a few people from the Hillary campaign. Our goal has been to break this four-year boom/bust cycle and be able to take this presidential code and sprinkle it down ballot. We used some of it in the 2017 races, and we’re using a bunch more in the 2018 races. Paired with that, we ran a whole bunch of data boot camps for campaigns on how to use this technology. One of the other things that we are trying differently is collect code from state parties and all their campaigns, so the DNC tech team runs a Github repository to preserve information cycle after cycle. We are trying to be that place that houses the tech.


How are these smaller campaigns reacting to the push for improved technology and data?

RK: It mostly depends on their staff right now. Some of the house races do have data and analytics staffers— but it shouldn’t only be a bunch of nerds in the basement that can do this. It should be everyone in the campaign. Our goal is for everyone to run a data-driven campaign by default. We’re sadly nowhere near that. I’d say maybe 25% know how to do it. There’s a big gap in training, and we’re working on that as a way to raise that number.


How are campaigns evolving? What will be different in the races over the next few years compared to 2016?

MD: Here’s what I see as what has fundamentally changed over the last 12 years: we’re not having high quality conversation about policy ideas. Most candidates are not trying to make the case for why the individual mandate is a good feature of the affordable care. They are fighting much lower quality fights— did they actually serve in the army? Is that Facebook photo of them even real? There’s a false picture of Christine Blasey Ford from 1980 with 150,000 shares, and there are equivalent examples in every single race. And it will dominate weeks of the campaign.

RK: On the DNC side, we’re doing our best to not be surprised. 2016 was surprising in a lot of ways. So we’re planning for the worst, in a way. Instead of celebrating that registration numbers in California and the turnout of Democratic primary voters are through the roof, we’re thinking ahead and trying to anticipate all the challenges that could come our way.