By Shari Balouchi
Sewanee graduate Rayid Ghani (C’99) gave this year’s Sherwood Ebey Lecture, an endowed lectureship dedicated to making mathematical ideas accessible to the general public. His talk on “The Role of Data, Technology and Analytics in the Presidential Election” focused on Ghani’s experience as Chief Scientist for the 2012 Obama campaign.
Ghani was born in Pakistan and relocated to the Unites States where he attended The University of the South. In college he realized his life ambitions through studying Computer Science and Mathematics. “I was motivated by two goals: one was to study and understand how we [humans] learn; and two: I wanted to solve large practical problems by making computers smarter though the use of data,” said Ghani. After graduation in 1999, Ghani attended graduate school to study Machine Learning and Data Mining before joining the Obama For America 2012 campaign as Chief Scientist.
Though Ghani previously worked for a technological consulting firm, politics was not an unnatural route for him. “Living in the US for seventeen years, you tend to follow the politics of the country because it does affect every person. You read about it, [and] discuss it with co-workers and friends,” he explained.
Ghani’s role as Chief Scientist was to improve fundraising efforts, volunteer involvement, and voter participation using analytics, social media, and machine learning. Data is useful in a political campaign because it informs users about patterns in voter behavior. Knowledge of voter behavior patterns can be used to make decisions about the most effective group to target in order to minimize costs. “The core of the work I was doing was looking at a large amount of data and making sense of it to help other people make better decisions,” he said.
There are many methods to consider from the cold calling approach to television advertisements to social media networking. Since resources are fixed, maximum efficiency is vital. “You could pick up the phonebook and just start calling everyone, but you’ll either waste calls on people who are already going to vote, or on people who can’t be persuaded to vote your way. But with a data-driven approach, you can target those voters who are much more likely to be affected by that call and pick up voters you didn’t have.”
Through a series of experiments, the Obama For America 2012 Campaign team found that sending many e-mails a day was more effective in raising money than sending out a few e-mails. This result seems counterintuitive because one might think that sending many e-mails would lead individuals to block or unsubscribe from the mailing list. However, because the statisticians were mindful of their audience and targeted individuals who already supported Obama, the multiple e-mails served as a reminder rather than a nuisance. In the end, $500 million was raised from e-mail fundraising alone.
Though the team was able to implement some new marketing strategies during the campaign, Ghani emphasized that the data they worked with was very limited. Contrary to popular perception, campaigners don’t have access to privileged information about individuals’ purchasing habits or recreational interests.
“Most of the data we had was from data that we collected either from interacting with people, which might mean either we called someone, someone donated money to us, or if they volunteered, or from voter registration records,” he said.
The two most important pieces of data were the likeliness of an individual to vote in favor of the Obama campaign and the likeliness of an individual to vote in general. The 2012 campaign focused primarily on individuals who were likely to support Obama but who were unlikely to vote.
Campaign volunteers were trained for door-to-door visiting. They would ask individuals whether he or she intended to vote, if a person knew where his or her voting location was, and when he or she planned to go to the site. “What’s happening in this interaction is their likeliness to vote has increased,” explained Ghani, “because you are causing them to think seriously about the details of the voting process.”
In addition to gaining votes by increasing voter participation, other powerful strategies were implemented. Those who were already supporters of Obama and registered voters served as integral players in the campaign, not just as monetary supporters, but also as persuaders. Electronic ads can be overwhelming and unmoving, while a neighbor’s heartfelt opinion may hold more value.
In an age of technological detachment, finding a link between the cyber and personal worlds is critical. Ghani explains how the two worlds were interwoven throughout the campaign in an effort to combine efficiency with connectedness. For example, a door-to-door visit would result in a follow-up e-mail with a relevant video attachment. On the other hand, an e-mail would be followed-up with a personal phone call from a campaign member. Ghani expressed his desire to use technology to increase public awareness of politics through a “grassroots” approach to political campaigning.
“There is a lack of data, so political parties need to start collecting…then they need to use it to understand the voters and allocate resources more efficiently. Parties that focus more on grassroots organizing are the ones most likely to collect and make more effective use of this data and as this process gets more mature and democratic, I hope it leads to a better educated public making informed voting decisions that are good for the country and its people.”
Though the campaign is won, Ghani looks forward to continuing his work in the non-profit world and to integrating data efficiently so that people can become more engaged in policy-making. “Data isn’t a secret weapon but an enabler of better democracy and more public participation. I see the future use of data as enabling more personalized and relevant interactions with voters, to get them more education about issues, more involved in political discussions, and have them even participate in creating public policies.”