BPR Interviews: Harrison Kreisberg


Harrison Kreisberg is the lead strategist for Blue Labs, a technology and analytics company formed in 2013 by senior members of the Obama campaign. Kreisberg’s team at Blue Labs works with political campaigns to optimize campaign strategies using data analytics. Kreisberg worked on Obama’s 2008 and 2012 campaigns, Terry McAuliffe’s 2013 Virginia gubernatorial campaign, and Hillary Clinton’s 2016 presidential campaign. Kreisberg graduated from Brown University in 2010.

BPR: Data-driven political campaigning is a fairly new development that many people don’t understand. How would you define data-driven campaigning?

HK: If you use the term data, it often scares people off. But fundamentally, what we are trying to do is take all the information coming into a campaign and make a decision. So the idea is to take the information we have and use that to help us decide whom [a voter] is going to support. Ultimately, we can’t talk to everybody, so we have to make guesses about people before we spend money on campaigning. You can’t answer every question [using data], but you can come to some basic decisions about whom to talk to and how to talk to them.

BPR: Were Hillary Clinton’s losses in key swing states caused by a failure to use data correctly? How could the Democratic Party use data more effectively in the future?

HK: There is no doubt that a lot of things were surprising to people on election day. Most polling aggregators got it wrong, and there was a fundamental shifting in the electorate with big changes in October and new players in the election like Russia and fake news. Data is as susceptible to these types of changes as anything else. Just as analytics was not [the] magic formula to victory that many claimed following Obama’s 2012 election, it is not fundamentally broken, as some are claiming in the wake of last November’s elections. Hillary Clinton [just] needed to win either Pennsylvania or Florida and one other swing state. The Clinton campaign should get a lot of credit for spending a lot in those two states, and that was something that data was able to bring to the fore. That said, there was much to learn from the 2016 outcome.

BPR: How has social media changed campaign strategy, and how can data analytics be used to target social media users?

HK: Social media is the next frontier in a lot of this, if we want to understand how folks are communicating. Not everyone is on social media, and if you only listen to social media you will get a biased vision of the world. Just like we need to understand which doors to knock on and which mailboxes to send letters to, we also need to understand which social media users we need to talk to. If the social media population is different, we need to recognize that and talk to them in a different way. It is tricky, because social media is so balkanizeddifferent groups of folks are talking to themselves and not to each other. When people are so segmented, it makes it harder to break into it.

BPR: Do you think using data instead of a more traditional political calculus to drive political decisions poses a risk?

HK: You can use any tool for good or for bad. Data is certainly a very powerful tool, and with more power, you have the ability to have a larger impact on both sides of the coin. Maybe there are some politicians who say, “tell me what the most popular belief is, and then I will say that.” But I don’t think that is common. You need something that is going to be authentic and believed, and the only way to do that is to truly believe it. If I told Hillary Clinton the best thing to do was to play to nativist fears, I don’t think it would have been authentic coming from her. [Some people] treat the data practitioner as a mercenary, but people in my shoes on both sides of the aisle care about politics. You can make a lot more money working in data and analytics in finance [than in politics]. You do this because you believe in it.