Hitmetrix - User behavior analytics & recording

Obama and Romney: Big Data, bad data

If the 2008 Presidential election was about social media and digital calls-to-action, then 2012’s election was about data. It’s about what happens when one side has the ability to make decisions based on good, clean data and the other side simply does not. As the post-mortems for both the Romney and Obama campaigns begin in earnest, the demarcation is pretty clear.

Often, providers of data solutions extol their ability to present clients with “actionable” data—a term so overused it’s been practically drained of meaning. But the Obama and Romney camps demonstrate that that despite the word’s reliable presence in vendor marketing copy, the distinction between actionable data and data is something clients still need to consider.

The Obama campaign, as reported by Time, had disparate lists from its 2008 presidential effort comprised of various groups (donors, fundraisers, etc), which Obama’s camp consolidated for the 2012 reelection initiative. Every decision the campaign made was based on data. Why, for instance, was Sex and the City star Sarah Jessica Parker chosen to host a fundraiser dinner in her New York brownstone?

The Obama campaign’s data-centric policies extended across multiple channels, including email and SMS. It sent variations of emails with different subject lines, signed by different members of the campaign—from Obama to first lady Michelle Obama to White House deputy chief of staff Jim Messina—to test who got the best response from certain demographic segments. Makes sense: eDataSource reports that Obama emails outnumbered Romney emails 20-to-1.

On the day of the election, the Obama campaign reached out to its constituents via SMS, asking them if they would be willing to make a phone call for a final get-out-the-vote (GOTV) push. Those that agreed were pushed an SMS message with the first name and phone number of a voter in a swing state.

It’s the sort of intense targeting one can accomplish only after organizing data, making sense of it, and developing communication and outreach policies based on that data. Moreover, data collection is dynamic: each email blast, for instance, gathered even more consumer information that informed the next initiative.

On the flip side, there’s the Romney campaign, which by all accounts was a disaster. By now, it’s common knowledge that Romney relied on seriously inaccurate polls. It’s not that the Romney campaign rejected data-driven initiatives. It didn’t; the data it had, however, was crap.

Consider Project ORCA, which was supposed to use Romney’s Boston headquarters as a hub that processed real-time information from swing states, sending enthusiastic supporters to targeted locations for a GOTV drive.

It’s hard to assess whether the data coming into ORCA was accurate because the data going to Romney’s ground forces—those supporters who were supposed to lead Romney’s GOTV push—was totally wrong. A firsthand account from an ORCA volunteer provides the grit. Though ostensibly a mobile campaign, volunteers received 60-plus page PDF packets that they were expected to print out. The packets contained incorrect or omitted data—for instance, the fact that volunteers would need a certificate to participate. Usernames and passwords to access the ORCA site were incorrect. The password reset feature didn’t work.

And prior to the packets going out (and it seems volunteers didn’t know whether those packets would arrive via email or post), the ORCA “web app,” supposed to be the portal that linked mobile volunteers, wasn’t even an app. It was an HTTPS mobile-optimized website, meaning volunteers looking for ORCA via Android or iPhone app stores couldn’t find it, and volunteers that keyed in the URL—thinking it was a simple HTTP address instead of a secure protocol—weren’t routed to the site.

In the end, perhaps it didn’t matter. ORCA echoed a similar Obama ’08 initiative called Houdini.

Like Obama’s Houdini, Romney’s ORCA crashed. Unlike Houdini, there was no backup plan.

So when we finally assess the way Obama and Romney used data, Obama’s campaign leveraged what it knew to micro-target both supporters and voters. Romney’s campaign couldn’t even organize ground forces and made strategic decisions based off a series of false assumptions derived from data that has since been proven misleading.

When we think about the way marketers leverage Big Data, just one year ago companies were mostly concerned with collecting data. That’s not the chief concern anymore; companies are now trying to figure out exactly how to use all the data that’s available to them.

Assessing both the Romney and Obama campaigns is like looking at the Big Data timeline. One camp—the victorious camp—took swift action; the losing camp thought it was sprinting when in fact it was still just learning how to crawl.

We now return you to our regularly scheduled program of David Petraeus’s humiliation.

Related Posts