Zeynep pointed to her New York Times op-ed, “Beware the Smart Campaign,” about political campaigns collecting and exploiting detailed information about individual voters. Given the emerging conventional wisdom that the Obama campaign’s technological superiority played an important role in the President’s re-election, we should expect more aggressive attempts to micro-target voters by both parties in future election cycles. Let’s talk about how voters might respond.
First, we need to think about the technical or policy changes that will likely make micro-tracking of voters’ activities more difficult. More and more users of web technologies are turning to anti-tracking technologies, such as ad blockers, browser plug-ins, and cookie controls. Browsers are moving toward tighter controls on cookies and other tracking technologies, so there will be an ever-growing group of users whose online activities will be more difficult to track. This is just one aspect of a move toward user choice that also includes initiatives like Do Not Track. In short, it will be harder for campaigns to collect information directly about browsing activities–so they may turn more aggressively to buying other forms of data.
Second, we will likely see greater pushback from voters against micro-targeted messaging. Targeted attempts to increase participation, such as get-out-the-vote and other aspects of the election-day ground game, are seen as mostly pro-democracy. But micro-targeted pitches designed to persuade voters to support a candidate are another story.
Voters have long disdained politicians’ tendency to tell every crowd what it wants to hear. In recent years the spread of cheap video cameras has helped to punish inconsistency. “Politician debating against himself” has becoming a recurring theme on the Daily Show and YouTube. The difference now is that a politician can tell every individual voter what she wants to hear.
The obvious response to micro-targeted messages is for voters to collaborate to find and highlight inconsistencies and differences. Given the right technology, this process can be automated. The easiest case to handle is email–voters might forward political emails to an aggregator that extracts and analyzes their content. (ProPublica used this method to collect a few thousand political emails this year.) Online ads can also be collected, for example using a browser extension, though that is more difficult technologically and raises greater privacy issues. (Anything like this would require robust, informed consent from the user, coupled with technical safeguards by the aggregator.) Even paper mail could be aggregated by asking voters to upload scans or phone-cam photos for analysis.
Analysis of targeted political messaging could have several goals. First, you could look for inconsistencies, where messages sent to different voters are in direct conflict. Second, you could examine differences in language and emphasis used to court different voter populations. Third, you could look for A/B testing, where a campaign seems to be trying out different variations on a message to see which is most effective.
Finally, to the extent that you have demographic information about the voters cooperating in the project, you could try to figure out how campaigns are targeting specific groups. The available demographic information might be coarse-grained–you might only know the voter’s sex, age group, and state–or it might be quite detailed. (Appropriate privacy safeguards would be necessary in this case.) Even without any demographic data, you might be able to infer something about the campaigns’ classification of voters by looking at correlations across different voters in your sample.
There is a rich opportunity for data collection and analysis here. The same technology that allows messages to be targeted and delivered can also help voters work together to better understand what politicians are up to. Voters deserve to know how the campaigns are selling themselves.