US President Barack Obama's victory Tweetpic

US President Barack Obama: how he got re-elected

The issues and demographic shifts that gave President Barack Obama four more years are being widely debated. Less attention is going to how his campaign harnessed them to secure victory. Was this the first major election where, “It was tech wot won it”?

“A long campaign is now over. And whether I earned your vote or not, I have listened to you. I have learned from you. And you've made me a better president.”

President Barack Obama, victory speech,  7/11/12


The issues and demographic shifts that gave President Barack Obama four more years are being widely debated. Less attention is going to how his campaign harnessed them to secure victory. Was this the first major election where, “It was tech wot won it”?

Team Obama, led by David Plouffe and David Axelrod, made formidable use of the Internet in 2008 to promote the candidate, his themes of ‘Change’ and ‘Hope’, raise millions and turn out the vote.

In 2012, it changed its game.

There was plenty of social media and online advertising, but greater emphasis went behind the curtain on techniques such as data mining, profiling and targeted personal messaging.

That backroom work has been kept under wraps. The campaign believed (rightly, it turned out) it had data and tools that gave it a clear advantage over Mitt Romney’s rival e-team. There was also a fear that openly discussing how thoroughly OfA was profiling supporters could provoke a backlash from Democrats themselves.

Finally, the ‘top secret’ lesson from the pre-campaign data and analysis was that Obama needed a very different strategy from 2008. But that remains where the story begins.

The long game

Elections end and campaign teams usually disband. In 2008, Axelrod and Plouffe retained their core Obama for America web staff though some left to friendly e-politics start-ups.

The most important was Bully Pulpit Interactive (, founded by Andrew Bleeker, director of Internet Advertising at OfA. This year, BPI was the main agency for Obama’s online advertising spend of, according to the Federal Election Commission, more than $50m (Romney spent $26m).

During the president’s first term, OfA was described publicly as a vehicle for continued communication with supporters. Importantly though, its role extended into ‘two-way traffic’, turning registered users into a permanent focus group. That web team today comprises more than 200 programmers, analysts and other staff.

Before the election, OfA tracked supporters’ reactions to its output, use of its tools and content, and behaviour on associated social media. It used the data to model voting patterns and research what might influence opinions of individual issues and Obama himself. Then, once the campaign was under way, it really got to work.

‘Retail’ politics

“The comparison is with retailers,” one Obama staffer told E&T. “If you have one of their cards, they can see what you’re buying and when. And they use that profile data to send you coupons.”

It’s clear how this can be transplanted to politics. It is also about things you are or are not ‘buying’. The Obama campaign decided there were two particular features of retail profiling it needed to adapt.

The first was to build a big, stable sample and infrastructure.

“Your supermarket harvests from you every time you shop,” the aide continues. “It builds a dataset of trends and behaviour over time. To be able to do the same thing [by keeping the OfA activity alive], gave us a serious advantage over Romney.

“We had four years’ data, all off of a consistent methodology, all run through software that had gotten more and more mature. They had a year roughly of data of their own and then packets of GOP [Republican Party] data from various sources. It’s not easy to put all that together. And it’s not as deep.”

The second feature was achieving granularity.

“Real-world focus groups are organizer-driven,” the aide says. “You want to know as much as possible about the sample. But in the time you have with them, they [the groups] tend to come back to the hot buttons. And they’re snapshots.

“Online profiling is bigger and continuous. You look at way more people. And it’s getting better at throwing up little things and new things a traditional group might not but which could make a difference. You want ways to influence an undecided voter or - and this was important for us - make sure someone who voted for the president in ’08 came back for him this time.”

Using this template, Team Obama built out OfA 2.0, a version that places as much weight on analysing supporters as its original objective of motivating and organising them. A recent analysis by Evidon found 87 different tracking technologies on Obama’s website just before the election (Romney’s had 48). So what did all this tell the campaign?

Reinventing the world’s most famous man

Earlier this year, that growing databank (and some traditional research) led Obama’s campaign managers to conclude that the candidate must present a different persona to secure re-election. The new-look Democratic base built around and through OfA felt he had fallen short on ‘Change’ and ‘Hope’.

The rhetoric had to go. A more subdued, reflective and even contrite tone was required. During the campaign, the classic Barack Obama would not reappear until his victory speech.

But when the reelection persona made its national campaign debut at the first debate with Mitt Romney - disaster. ‘Subdued, reflective and contrite’ fell flat given the adversarial format. An energetic and combative Romney recovered a large chunk of ground he had lost over a miserable Summer.

However, while several commentators suggested that Obama’s performance would lose him the election, the campaign held firm to what the data was saying. The debate was a serious mistake, but turned out to be a case of losing a battle rather than the war itself.

The base, stupid!

A new presidential persona was not itself a magic bullet for the potential voter shortfall anyway. First, the candidate. Then, the issues.

Obama won in 2008 by broadening the core Democratic vote. He attracted younger people traditionally seen as disinterested in politics. He expanded the party’s appeal to women voters with progressive positions on abortion and equality in the workplace.

And he reached out to ethnic groups, particularly African-Americans (13 per cent of the US population) and Hispanics (10% and the fastest growing group in the country). Both felt Washington ignored them and historically had not voted in significant numbers.

For 2012, it was clear that trackable supporters in these blocks were wavering. They would not defect to the Republicans, but if they did not turn up at the polls at all, Obama would still lose.

Bases are important everywhere but particularly important in the US because it does not count the popular vote to elect a president but uses an electoral college allocating 538 votes proportionally across all 50 states. Each state’s representatives cast the official votes.

Obama’s route to a 270-or-greater majority required victories in a handful of ‘competitive’ states where, women aside, his core represented more of the population than the national average.

“It wasn’t all of them [the 2008 voters] and it wasn’t the same across every group,” the Obama aide says. “But it was totally obvious that the same enthusiasm wasn’t there. Our hold on them wasn’t as strong.”

These voters wanted a more sober president, but they also wanted their concerns to be noted. It was here that profiling-based campaigning was put into action.

Public informs private

The 2012 campaign was extremely negative. Through BPI, the Obama campaign used targeted online ads as warnings.

Romney was Wall Street red in tooth-and-claw and exported American jobs. Romney wanted to deport illegal immigrants and wouldn’t implement reform. Romney’s running mate Paul Ryan had authored a budget plan that attacked African-Americans. Romney had a binder of things he wouldn’t do on women’s rights.

As blunt as the messages from both sides were, there was some smart technology behind them.

Because BPI was integrated with OfA, feedback loops selected and monitored which campaigns worked, which didn’t and where ads were best placed. The objective mixed convincing ‘swing’ voters (this was one of the campaigns public facets) with addressing wavering 2008 voters.

But OfA went deeper. The ads fed the broader analytical infrastructure by capturing, among other things, email addresses, Facebook ‘likes’ and other data. New connections were analysed, including public profiles. Reactions from existing database members were treated the same. Everything was sifted in the data bank.

The results drove a separate, private, personalised campaign strand. Here are some techniques that we understand were used.

  1. Detailed personal information about prospective voters was given to phone and door-to-door canvassers.
  2. Personal emails were sent addressing the election theme or themes in which the recipient had shown greatest interest.
  3. ‘Active’ campaign volunteers were alerted to friendships and connections they had with potential Obama voters in the same area. A ‘multiplier effect’ was sought: messages are better delivered by someone you know than someone you don’t.
  4. Email, phone calls and visits were made to encourage voter registration and provide advice on where to vote.

None of this is especially new. Technology is evolving campaigning techniques. But the precision is increasing fast. A campaign manager could guess one technique might ‘chime’ with a certain community; today, though, they want techniques that sing an aria. To you personally.

Increased accuracy also provided a way to focus messaging that the Obama base especially required.

“Republicans are interested in a closely grouped set of issues,” the Obama aide says. “Democrats are more diffuse in the themes that will get them to engage. Or possibly disengage.”

Take immigration reform. Hispanics alone may have given Obama the votes that swung the final result, opting 71 per cent-27 per cent in his favour. Their naturalised community ardently wants reforms that will formalise the presence of members without papers, particularly Obama’s proposed Dream Act that offers a path to citizenship for people brought to the country illegally as children.

“But the Hispanic community is predominantly Catholic and has a significant evangelical sub-group. Many trend toward Republican positions on abortion or gay marriage. Many African-Americans do too,” the Obama aide says. “You want to make immigration a more important issue for them [Hispanics], when you present the president’s platform.

“Then, I did mention African-Americans. Immigration reform does not play well in that community. It sees cheap, illegal labour taking away blue collar jobs from black workers because the employers can force a lower rate and don’t have the same payroll obligations. For them, you need to highlight some of the Republican proposals on cuts to entitlements.”

Sending the right (and equally not the wrong) message to each voter was vital to mobilising Obama’s ‘rainbow’ coalition. Mass marketing here was tricky. Personalised marketing could get base voters to concentrate on their individual priorities.

Metric-driven politics

Election night at Mitt Romney’s HQ saw widespread (but actually not universal) shock as networks called Ohio for Barack Obama and disbelief that Virginia stubbornly remained too-close-to-call only to also plump for the president. Then, a tied Florida count was stopped and stretched into a second day.

Romney campaigners thought their man had edged victory in all three states. They did not believe that the youth vote would arrive or that Hispanics (significant in Virginia as well as Florida) would turn out in volume.

Strangely, that was largely because they had come to the same conclusion that Obama’s team had: these groups were no longer as enthusiastic about the president. They were right. But it wasn’t so much about the president now, and they underestimated that shift as well as its effectiveness in turning out the Democratic base.

OfA had been in the game longer. It had more time to gather data and refine analysis tools and delivery mechanisms. It used a platform that had evolved rather than being bootstrapped together. Its knowledge allowed for greater innovation.

Many traditional elements of a successful IT implementation were there, through to integration with third-party suppliers.

Obama’s technological advantages may not have won him a second term in their own right. Also, importantly, they augmented and informed but did not replace the campaign’s very strong grass-roots organisation. But they still may well have put Obama ‘over-the-top’ in a very tight race that required exact regional and demographic campaigning. Every little helps.

So, the president did indeed listen and learn, in more ways than most Americans realise. Will it make him a better president? The next four years will tell.

Recent articles

Info Message

Our sites use cookies to support some functionality, and to collect anonymous user data.

Learn more about IET cookies and how to control them