As I See It: How IT Decided The Election
December 3, 2012 Victor Rozek
In retrospect, analyzing wars to determine the exact tipping point for the victorious side is tricky business. There are multiple battles and often multiple fronts. Strategic decision making, tactical challenges, advantages in weaponry, motivation of the fighting force, and a hundred other factors all play a part. As, sometimes, does chance. Just deciding if divine favoritism played a part would be a subject of testy debate, although that advantage is usually claimed by the winner.
The aftermath of elections loosely resembles the aftermath of wars, at least in the extensive postmortem analysis. Where victory and defeat are concerned, both sides are obsessed with the question “why?” The winning side wishes to replicate its success, the losing side wishes to avoid a repetition. As with war, many people on both sides have to accomplish a great many tasks simply to engage viably in the struggle, but a strong argument can be made that this was the first presidential election to be decided by computing expertise. To put it plainly: Obama won because Democrats had better analysts and better coders.
To appreciate the part IT played, it is necessary to recognize (ballot headings notwithstanding) that none of us casts a vote for the President of the United States. The Electoral College (which is either viewed as outdated or, as the antidote to the tyranny of the majority) sees to it that we only vote for the president of the state in which we reside. Our individual votes are of little consequence on the national scale. As Al Gore discovered, winning the popular vote by a margin of more than 500,000 votes did not guarantee him a stint in the White House.
As we are reminded ad nauseam on election night, the critical calculation is how best to reach the 270 electoral vote plateau that triggers the playing of Hail to the Chief. And, given the relative parity of the red/blue divide, success depends on winning a handful of swing states. There, the ability of the campaign to get out the vote (GOTV) is tested. Which is precisely where technology and analysis provided the winning margin.
Time magazine succinctly summarized the new political reality in its commemorative election special: “In politics, the era of big data has arrived.” One of the first to recognize the magnitude and possibility embedded in that reality was Obama’s campaign manager, Jim Messina. “We are going to measure every single thing in this campaign,” he said, and promptly hired an analytical department “five times as large as that of the 2008 operation,” a group that had produced transformative results in its own right.
The metric-driven campaign began by combining dozens of stand-alone databases, merging information compiled by pollsters, field workers, fundraisers, and consumer trackers. They blended those data with Democratic voter files, social media, and mobile contacts. Exactly how the coders, quants, and data crunchers performed their alchemy is a closely held secret, but the results are a matter of record and disclosure.
For one thing, the Obama campaign was thought to be financially overmatched. No one seriously believed the Democrats could approach the $1 billion the Republicans were expected to raise. (An absurd amount to win a job that pays $400,000.) But careful testing and analysis created customized Internet appeals, with “different subject lines, senders, and messages.” The tailored appeals raised “10 times as much” as standardized requests. And, a Quick Donate program allowed giving on-line via text message “without having to re-enter credit-card information.” Voters were analyzed and targeted based on their propensity to give and their preferred method of giving. The super-PAC advantage of the Republicans was neutralized.
But getting out the vote is the name of the game. The team was able to create predictive profiles of voters in swing states (29,000 in Ohio alone), to increase the effectiveness of everything from telephone and direct communication, to ad buys, mailings, and social media contacts. For the first time Facebook was used to encourage voting, friend to friend. Leading up to the election, volunteers using smartphones were provided with a list of persuadable voters, and instructed where to go, who to contact, and what to say.
Polling and voter contact data were continually reprocessed to account for every possible scenario and shift in voter allegiance. Using computer simulations, they “ran the election 66,000 times every night.” Each morning they had detailed reports showing which areas provided their best chances of winning, and allocated resources accordingly.
The Republicans similarly hitched their elephant to the technology wagon. Theirs was an expensive and highly compartmentalized project, dubbed ORCA. It was designed to provide the campaign with real-time poll monitoring and drive the GOTV efforts on election day. Andrea Saul, a Romney campaign spokesperson, told Business Insider that ORCA was “the Republican Party’s newest, unprecedented, most technologically advanced plan to win the 2012 election.” It would, in her estimation, surpass in its effectiveness the software used by the Democrats.
For reasons that aren’t clear, the development team chose to withhold details of the project from everyone outside of Romney’s inner circle until the week of the election. What is clear in hindsight is that the software was never properly beta or stress tested. The rollout, according to volunteers, “was a complete and utter failure.” The night before the election, some poll watchers received a daunting 60 pages of PDF instructions while others received none. Regardless, they proved not only to be unwieldy but unusable. The campaign failed to provide volunteers with poll-watching certifications, and they were turned away. On election night, an army of 30,000 operatives and volunteers were standing by, waiting for the big download. But nothing happened. John Ekdahl, a Romney volunteer writing for Ace of Spades HQ, said that “the end result was that 30,000-plus of the most active and fired-up volunteers were wandering around confused and frustrated. . . . ” A communication strategist told Business Insider that “the digital strategy was so incomprehensible–they were playing with Super Nintendo while Obama’s people had PS3.”
Inadequate security and, in one visible instance, a lack of user sophistication, may also have contributed to the Republican’s digital woes. The hacker group Anonymous claims to have hacked ORCA to prevent voting machines from flipping votes, which infers that the project had much broader goals than the reported poll monitoring. But that claim is unsubstantiated. Meanwhile, a hacker emailed Gawker alleging to have hacked into Romney’s private email and DropBox accounts. He did so, according to the email, by guessing Romney’s favorite pet in response to a security question.
Ultimately, ORCA failed because of a lack of proper testing and training. When it was needed, the program crashed, and with it Republican hopes of securing the presidency. But that will in no way diminish the role of IT in future elections. One thing is certain. As more and more of our personal information is gathered and analyzed, and our profiles become nuanced, political parties will expand the range and proficiency of their digital influence.
And, in a close election, the party with better software is likely to win.