Barack Obama: Against the Odds
2008:
鈥淭here was a new actor in the campaign drama: Jim Messina. Obama convinced Messina to leave his political father, Sen. Max Baucus, by calling him the day after Hillary Clinton dropped out of the Democratic primary contest. The sales pitch was neither about hope nor change. 鈥淵ou鈥檙e really going to get to run a business,鈥� Obama told Messina.鈥� [1]
2012:
鈥淭he power of this operation stunned Mr. Romney鈥檚 aides on election night, as they saw voters they never even knew existed turn out in places like Osceola County, Fla. 鈥淚t鈥檚 one thing to say you are going to do it; it鈥檚 another thing to actually get out there and do it,鈥� said Brian Jones, a senior adviser.鈥� 聽[2]
Introduction:
The 2008 and 2012 US general elections were two of the most closely watched political races in modern history. Mr. Obama鈥檚 2012 campaign broke fund-raising records- attracting more than $1.1B in donations, most of which was raised online. The campaign also designed and executed the most sophisticated electorate mobilization operations ever seen in American politics to defeat Mitt Romney- a well-financed establishment candidate. Obama鈥檚 2008 candidacy was based on the electorate鈥檚 general dissatisfaction with the direction of the country. The 2012 race was staged as a referendum on Mr. Obama鈥檚 controversial health care and government-funded bailout programmes. While most of the campaign鈥檚 鈥渃ustomer鈥� promise could only be delivered whilst governing in office, the first, most critical task/promise was getting the underdog candidate re-elected 鈥� and the campaign operations were up to the task.
Customer Identification
Who was the Obama voter and how did the campaign use its operations to identify them? The campaign analyzed opinion polls as well as historic voter information from Mr. Obama鈥檚 previous statewide/national elections to identify the 鈥渃haracteristics鈥� of potential voters. The campaign then recruited a cadre of behavioral scientists to extrapolate and build out a database populated with millions of voters that fit the target profile. The 鈥渃ustomer鈥� journey from initial contact and voter registration to message delivery, persuasion and balloting, was mapped out in to distinct waypoints. A file was created for each voter that could then be tracked throughout the entire journey. [3]
Delivering the Message/Media Operations
The results of the Obama campaign鈥檚 target voter profile suggested that potential voters were of diverse backgrounds, and that they skewed young and non-white. The diverse nature of Mr. Obama鈥檚 voter base presented significant challenges- many of these demographics were motivated by sharply different causes. Millennials were concerned with expanding civil rights as well as what they perceived as America鈥檚 perpetual war-footing, baby boomers voted their views on income inequality and the direction of the economy, independent voters favored more measured federal policies. Obama鈥檚 message/media operations had to deliver a message that had broad appeal to these different demographics. To tailor these messages, the campaign鈥檚 media operations teams conducted extensive A/B testing. [4]
For example, the Obama team would first test fundraising emails on a small representative subset of the donor list with varied subject lines and content. The sample鈥檚 response and click-through rate data were aggregated and used to project fundraising revenue for the broader donor pool. Emails with the highest projected fundraising values were then sent out to the entire donor list. [5]
The campaign also created My.BarackObama.com- a social media platform that allowed enlisted volunteers to create customized pages, generate fundraising links and view likely-voter information for people within their neighborhoods. The platform included a 鈥淒ashboard鈥� that when integrated with a volunteers鈥� facebook account, could search for potential voter-friends who had not yet been registered. Volunteers could then contact these new voters with a customized message from the Obama campaign. [6]
In total the campaign facilitated 125M personalized phone calls over the course of the election. [7]
Dashboard also decentralized the flow of voter information within the campaign. Campaign headquarters, field organizers and volunteers all had access to the same information updated in real time. [8] Volunteer-updated voter rolls would immediately be visible to campaign headquarters, which could then modify where specific voters were in their respective journeys. [9]
Election Day – Get Out The Vote (GOTV)
The Obama campaign鈥檚 statistical team assigned two voter rating scales to every registered voter. The scales ranged from 1 to 100 and were predictive of both the probability that the voter would support Obama as well as the likelihood that they would show up to the polls. [10] Voters with high support scores but lower likeliness to turn out received targeted election-day attention from the Obama Campaign鈥檚 GOTV operation. To facilitate this attention the campaign set up 5,100 GOTV stations in battleground states and enlisted 700,000 volunteers to get voters to the polls. [11]
Results
Mr. Obama won re-election in 2012, at a time when 52% of the electorate was of the opinion that America was 鈥淪eriously off on the wrong track鈥� and that 鈥渢he Federal government was doing too much鈥�. Unemployment was stubbornly high at 7.8%, just a single basis point lower than when Mr. Obama first took office in 2008. The fact that Obama was able to win in such unfavorable electoral conditions is amongst other things, a testament to the seamless integration between his campaign鈥檚 charter, and its boots on-the-ground operations.
[1] http://www.thedailybeast.com/articles/2013/09/12/no-drama-obama-s-dramatic-2012-reelection-campaign.html
[2] [http://www.nytimes.com/2012/11/08/us/politics/obama-campaign-clawed-back-after-a-dismal-debate.html?pagewanted=1&_r=0&hp]
[3] http://www.nytimes.com/2012/11/13/health/dream-team-of-behavioral-scientists-advised-obama-campaign.html
[4]
[5]
[6]
[7] .
[8]
[9]
[10]
[11] http://www.huffingtonpost.com/2012/11/03/obama-voter-contact_n_2069289.htm
Very interesting post. I enjoyed looking at the campaign from a business and analytical perspective. It would be interesting to how analytics are used throughout elections overall, especially in regard to detecting voter fraud. In addition, I would be interested to know if the platforms and programs created were solely for the campaign. For example, is having better analytics and programming an advantage of just his campaign, or do all candidates use the same platform and it is just a matter of who uses the data the best?
I also enjoyed your post very much. Do you think the advantages that benefited Obama was a once-in-a-lifetime phenomenon because technological presence across candidates and parties will be closer to parity. Obama had the incredible advantage of a compelling message but he was, as you highlight, the first to successfully use many of these technologies. Will future candidates ever have the “tech edge?” Also, is the Obama model the blueprint for future campaigns, especially the 2016 election, or does technology change so quickly that the Obama Campaign model is now antiquated?
[Comments to RGO’s post below]
A well-researched post! I would be curious to take RGO’s comment one step further: Online tools have been pivotal to GOTV and awareness-raising efforts since at least the 2000 election, if not before. That said, given the long four-year span between elections and the rate of technological advancement during that period, voter/donor lists seem to be the only long-lasting asset spanning multiple campaigns. It is difficult to envision a campaign using the same tech strategy as an election four years prior. This trait lends support to RGO’s argument that a temporarily strong tech presence might be without value to later campaigns (through regression to the mean / “catching up”). However, if a tech strategy is what drives atypically good voter/donor registration, then the effects of a strong tech presence would be longer lasting, even if the actual platforms and strategies are wholly different.
Thank you for this post! Adding to the conversation above, I’d be curious to learn how the Obama campaign tech strategy evolved from 2008 to 2012, and to what extent Romney 2012 was able to co-opt Obama’s 2008 tech infrastructure with a very different target audience. It seems as if Obama’s campaign operations were uniquely suited to his target demographics (specifically, skewing young); older voters were probably less accessible through Facebook-based programs like the Dashboard.
I was also wondering about how this tech strategy originated. Was Obama’s campaign more open to innovative ways to get out the vote because he was an unexpected contender for the presidency? Maybe other campaigns can learn from the organizational structure that permitted a brand new idea to become the centerpiece- and driving force- behind a winning campaign.
Thanks for this interesting post! One thing that I’d be further interested in learning about is how this operating model lent itself to mobilizing so many volunteers for these elections. I don’t follow politics very much at all, but an unprecedented amount of people in my immediate circle actually got involved in Obama’s campaigns, beyond simply voting. From myself personally going out & knocking on doors to my best friend working as a Field Organizer for Obama full-time to another college friend actually taking a semester off to work on the campaign, it seems that a huge part of the success of his campaigns hinged on its ability to mobilize deep grassroots support. How did they think about strategically involving the most enthusiastic “Obama evangelists” to further spread their message?
Fascinating. What strikes me about this post is that it reflects certain business marketing themes that I would not have thought appropriate in a political context, but clearly they worked. This seems to be a data analytics model. The campaign was generating and collecting all types of data about voters and using it to hone its message to be increasingly effective. This raises something of a moral issue for me – if the Obama campaign can use data to run A/B tests on users (aka voters), what would stop any politician from doing the same? Are we voting on the candidate that best represents our views or does the campaign with the most money to spend on data analysis and technology end up being victorious every time? If we can tailor our messaging based on tests, what stops us from tailoring our message to individual voters so that there is ultimately no such thing as a single campaign promise, but rather a collection of promises that are most resonant to various voter subgroups, and what is the end result of such promises in action after an election?
It’s interesting to hear what happens when such techniques are adopted en masse. Obama’s campaign strategy has become the gold standard for campaigns across the country. These tactics are now being employed in campaigns across the country. Does this mean that they have now lost their competitive advantage. If so, what can be done to continue the success of such a model?