Any mention of the words “big data”, “analytics”, or “micro-targeting” in political conversation today runs the risk of conjuring up frightful images of an Orwellian Big Brother spying on citizens’ every move, infringing upon the electorate’s right to privacy by collecting information against their wishes, and feeding illicitly-obtained spools of data into a digital black box to produce nefarious propaganda and mind control. This science fiction couldn’t be any further from the truth of how big data analytics is used in modern campaigning.
Let’s begin pulling back the curtain on campaign data and analytics by eliminating what it isn’t.
Firstly, big data usage by campaigns is not creepy. The most plentiful, useful data to any campaign is publicly known and freely available to all political parties—Elections Canada’s Lists of Electors and Statistics Canada’s Census data.
Secondly, even private party data and lists are not secretly obtained. Since no party will ever know exactly who will vote for or against its candidates, they can only guess at future likely voting outcomes based on information freely given by individual electors. Of course, elector responses or actions can be used for future political interaction if a person browses a party’s website to donate, visits a local office to volunteer, or responds to a canvasser’s door knock or phone call asking which party they support. However, anyone can easily abstain from self-identifying simply by not participating.
Thirdly, microtargeting does not rely on consumer lists. Much fuss has been made about the use of consumer data in political analysis, for example cat vs. dog magazine subscriptions or bourbon vs. cognac drinkers. The truth is, most consumer data is not useful because either the difference between groups’ preferences in large lists is too small to be predictive (e.g., The Beatles listeners tend to divide 50/50 between Democrats and Republicans), or small lists don’t have enough coverage to make expensive data purchases worthwhile (e.g., paying top dollar for a list that covers 0.01% of the electorate is a waste of scarce campaign money).
Lastly, analytics should not be confounded with manipulating statistics in an effort to sell a misleading narrative, à la Karl Rove’s U.S. election night meltdown on Fox News.
Analytics doesn’t magically change who people are or what they believe—instead, it helps bring these realities into greater clarity for a campaign to act upon.
So then, what exactly does big data analytics do for campaigns? It allows a campaign to:
- Closely observe the behaviour of common political interactions between campaigns and the electorate (e.g., door knocks, phone calls);
- Remember the specific conditions and outcomes of each of these observations (e.g., if someone answered and responded on a particular day);
- Consolidate all of these conditions and outcomes along with a multitude of other data points (e.g., census, historic data);
- Explore behavioural patterns beyond the traditional geographic or polling crosstab groupings (e.g., social circles, affinity groups);
- Evaluate the significance of how all these disparate data points interact with each other;
- Recommend optimal next action steps to maximize a campaign’s effectiveness and efficiency (e.g., who to contact, when, how, and with what message or call-to-action);
- Empower quick, precise resource deployment customized down to the individual level (e.g., treating each and every elector as a unique person).
So why is all of this big data good for Jill and Joe Public? It’s very easy for disengaged citizens to feel that national, provincial, and even municipal elections are too large and too impersonal to care about what any single person thinks. Too often the electorate has been told they are looked upon solely as stereotypes, such as belonging to a large voting bloc of “hockey moms”, “rural voters”, “urban voters”, or “youths”. Samara’s website eloquently states that “[d]espite Canadians’ strong support for democracy and concern for the health of their communities, more people are opting out of the political process.” The personalized, one-on-one attention that big data analytics empowers a campaign to engage in with each and every elector holds the promise of helping to reverse the tide of citizen disengagement by demonstrating that each individual matters, and that their unique concerns, suggestions, and aspirations are important and heard and shared by others in their local communities.
Big data analytics is, at its very essence, listening to and interacting with electors as unique individuals, in a way that resonates best with each person, in the hopes that they become fully engaged and, ultimately, vote.
When campaigns use big data analytics properly and completely, they ensure that every dollar donated, every hour volunteered, and every minute spent listening to and speaking one-on-one with individual electors is respected, remembered, responded to, and acted upon in the most personalized manner possible.
In a nutshell, big data analytics gives a clear, amplified voice to each and every ordinary citizen, which can be good for democratic engagement if campaigns choose to listen closely and act accordingly.
Inspired by Ethan Roeder’s Op-Ed published December 5, 2012 by The New York Times.