Anirudh Suri

The election of President Obama in 2008 was by no means always the expected outcome. As analysts look back and analyze what gave him the edge over strong candidates such as Hillary Clinton in the primaries and then John McCain, many point to the Obama’s team of “New Organizers”. Four years later, the re-election of Obama was again by no means always the expected outcome. Except this time around, as analysts looked back and analyzed what gave him the edge over Mitt Romney, many pointed to not the “New Organizers” but the “Cave of Data Scientists.”

Zack Exley explains the secret behind Obama’s grassroots organizing success: the team, he argues, “undogmatically mixed timeless traditions and discipline of good organizing with new technologies of decentralization and self-organization.” This was in stark contrast to earlier efforts where the organizers were  either so top-down that the volunteer army felt “choked” or so bottoms up that they sidestepped the need for planning, management and coordination.

The Obama neighborhood teams were run on the motto of “Respect. Empower. Include.” This meant that the organizers really empowered other organizers to develop their own teams of organizers, each of whom would then further build out their own teams. There was no end to it. This hub and spoke model (except there were multiple connected hubs) that the Obama team built is very reminiscent of the map of the Internet or even social networks. The Obama team was simply linking people (bridging capital) and then tapping their personal or professional networks (bonding capital).

While empowerment and larger field teams might have been the explicit objective of this organizational structure, the most important outcome in my mind (and maybe the most important driving factor of this strategy) was “efficiency”. Volunteer teams as well as field organizers were all able to increase their productivity greatly, as the local volunteers were empowered (through technology and training) to carry out voter-facing activities on their own without much assistance from the top. This meant that a less well-funded (at least in the initial stages) Obama team could compete more effectively against the Goliath that Clinton’s fundraising machine was.

Sasha Issenberg weaves the story of how President Obama’s campaign used data analytics to target and mobilize  individual voters – but really this story also boils down to Obama’s effort to achieve greater ROI on its various organizing and mobilizing efforts. The strength of the work that Dan Wagner was leading lay neither in the fact that “Wagner successfully predicted the final margin within 150 votes—well before Election Day” nor (as Sasha argues) in that Wagner represented a “new way of thinking” in which voters were not stuck in old political or demographic or consumer marketing driven categories.

Rather, the strength really lay in the ability of the Obama team to use Wagner’s predictive scoring models and randomized controlled experiments to help with all resource allocation decisions (whether it was allocation of field volunteers or advertising dollars). The analytics team ultimately drove decisions of not only field strategy (i.e., which voters were most persuadable (based on the estimated persuasion score) and hence which voters should the field teams be focusing on) but also media buying (finding the persuadable voters in the cable companies’ billing files and hence buying media that these persuadable voters would be watching). This was evident as the media buying software platform, Optimizer, helped the campaign conclude what would be “the more efficient way[s] of reaching persuadable targets.”

In both the field and the media buying teams, the analytics team then really played the role of helping each team achieve what I call “efficiencies at scale”! The pre-requisite in the case of the New Organizers was the organizational structure and recruitment of teams. In the case of the Data Scientists, really the fundamental element was the existence (and to some extent) collection of relevant data. Once both the teams had the basics down, then there was no stopping them from scaling. So in a way both efforts were really well suited to achieving efficiency at scale very quickly after the initial foundational effort.

In a lot of ways, this is not unlike what entrepreneurs (at least in the tech space) do – when it comes to consumer marketing. The idea is, with limited budgets, to reach niche consumer segments (defined less by the categories that would traditionally describe them than by the behaviors that really define them at a deeper level) that are more likely (than not) to use the startup’s product. The challenge of course is to find the pre-requisite data that can really help achieve these “efficiencies at scale” that in turn can make startups into efficient mobilization and persuasion machines.

Image credit: LAtimes.com