If you drove a Prius in 2012, data suggested you were 80 percent more likely to vote for Barack Obama—a nugget that neatly captures the appeal of big data in politics. Too often when we talk about big data, it’s suggested that someone out there has cracked the code on how to manipulate swirls of data to perfectly target and influence voters. That’s not the reality.
As David W. Nickerson and Todd Rogers argue in their paper from 2013, Political Campaigns and Big Data: “Despite sensational reports about the value of individual consumer data, the most valuable information campaigns acquire comes from the behaviors and direct responses provided by citizens themselves.”
Leaving aside all of the targeting advancements American political practitioners have made, this remains the reality for most European campaigns. That, far from cross-referencing demographic information with, say, purchasing history to triangulate the perfect voter to target, it is actually simple voter responses to simple questions that are most valuable. For a European campaign that wants want to run a smart, effective campaign based on data, it’s critical to start gathering data points in a rigorous way from day one.
The Regulatory Challenge
For many European countries, the use of big data to target voter segments with email marketing, canvassing or social media advertising is simply not acceptable. Germany, for example, has had quite restrictive data protection laws in place for many years that are stronger than the recent General Data Protection Regulation (GDPR) of the European Union.
As Thomas Jarzombek, the digital affairs spokesman for the CDU said in parliament, “There is no other country where it is more difficult to work with data.” He was referring to the business community, but the same rules apply to politics. Many countries use television debates, political journalism on radio, or large public rallies to learn about politicians and campaigns. The discourse is public, high-level and avoids the microtargeting techniques favoured in the US.
The Cambridge Analytica scandal would have you believe that Facebook was ransacked for consumer profiles that were then used to effectively influence large segments of voters. Certainly, Cambridge Analytica were happy to suggest (before the scandal broke) that they exerted this control. After all, they had a vested interest in promoting their abilities with data manipulation in order to sell their consultancy services. Having hard data on how effective those efforts actually were is another thing.
The other output of big data firms tends to be area profiles like Geographic Information System (GIS) maps. These profiles take, for example, electoral precincts and attempt to distill them into an actionable form for political campaigns. Is the area a stronghold for a political party? Does it have a certain income profile? Is it made up primarily of families or young professionals? Based on these area profiles, campaign tactics develop to target certain areas that have the highest potential return with specific campaign messaging.
For example, the French political consultancy, Liegey Muller Pons, did exactly this, working up 60,000 area profiles across France in the run-up to the last election there. Working with the Macron campaign they identified 20,000 of those areas that would be worth targeting with door-to-door outreach.
Communication Over Influence
So, how do you go about getting the benefit of big data insights while working with local data? Start with the basics: a voter file like an electoral register or property register. Then voter contact by street canvassing, door-to-door canvassing and focus groups that would pump direct voter feedback into the database.
If these insights are used correctly they can, in the words of Nickerson and Rogers, be used to, “produce individual-level predictions about citizens’ likelihoods of performing certain political behaviors, of supporting candidates and issues, and of changing their support conditional on being targeted with specific campaign interventions.”
Going back to the original example of Obama’s ‘08 and ‘12 campaigns, the reality was that he followed the best campaign template out there: the local campaign. Local campaigns tend to have excellent local knowledge of issues that concern the electorate (because they live there) and they have excellent knowledge of individual voters and their histories and propensities (because they know them personally). The best of those campaigns apply that knowledge to speak to them in a way they understand and that reflects their concerns. This is essentially what Obama did, but he scaled it up to the national level.
For European campaigns, understanding this model may be the key to unlocking smart campaigns at scale without resorting to big data practices that influence rather than communicate with the electorate.
Brendan Tobin is the Head of Growth at Ecanvasser.