In 2013 I helped run a progressive campaign for Mayor of my hometown, El Paso, TX. We were aggressively outspent by the two favorites campaigns, each raised about $250,000, and we raised $40,000. We couldn’t afford any web, advertising or really any consultants so we did everything “in house.” When I indicated my natural affinity for communications, the campaign collectively told me to manage the voter database and work on online advertising.
The Texas Democratic Party (TDP) maintains a database named the Voter Activation Network and commonly called the VAN in the trade. The VAN was created during the tenure of Howard Dean as the chairman of the Democratic National Committee, in which he pursued a “50 State Strategy” of making the Democrats viable even in red states. Part of this strategy was helping create voter databases in each state, including Texas. Most databases combined publically accessible data with some commercial data and user inputted data. We as a campaign decided to purchase the VAN data and data interface from the TDP. Since we were running a citywide campaign, we only provided the voter precincts we needed and paid roughly $500.
The data set and interface we received was great at providing the publically available data into easy to use spreadsheets. We would create “walk-lists” of voters in neighborhoods and make calls based on the criteria we believed would be likely voters. My own lack of sophistication with manipulating data at the time prevented me from communicating the best ROI (Return on Investment) voters. In addition, the data lacked city election level data, so we didn’t know if someone who voted in the 2012 Democratic primary voted in the 2011 city elections and would thus be likely to vote in 2013.
Some population numbers for perspective. The El Paso-Ciudad Juarez Border Metroplex region has over 3 million people and is the second largest border region in the world. Within the city limits of El Paso, there are about 650, 000 people. Of that population, about half are registered to vote. Of that registered population, half turn out to vote in presidential election years. In midterm elections, half turn out to vote compared to Presidential years. Finally in city elections, the turn out rate is below 10%.
We had a constant debate in our campaign, should we just target highly likely voters or try to bring out voters who maybe had never voted for Mayor or city council. We eventually settled on targeting likely voters in precincts that were not being targeted by the other campaigns – usually in less affluent areas. Yet, in the back of my mind I had concerns for democracy writ large when campaigns only target those likely to vote.
Our experience with online advertising was instructive as well. We used facebook advertising to target a population of voters that were female, interested in politics and some other small factors. The utility of facebook data is that the user herself supplies it when she fills out her profile.
My own experiments in data driven campaigning lacked the sophistication of professionals at the highest federal level of elections. While the practices are trickling down, I maintain some reservations for our democracy in the digital age.
First of all, national campaigns have to raise a lot of money. The diversity of platform and the fragmentation of audiences will require much more spending by campaigns in the future. The “money election” de factodisqualifies candidates from the actual election if they cannot get their message out. This of course has serious implications for the health of our democracy.
Second, presuming national candidates raise the money to advertise on many platforms – it may have deleterious effects on those campaigning for state and local level governing. Whereas Presidential advertisements may become omnipresent, it will become harder for local campaigns to cut through the noise. National campaigns can advertise in video games, online videos, apps and many others yet to be created platforms.
Finally, micro targeting of political advertising threatens turning the great public debate into smaller echo chambers. A democratic campaign armed with data that a household has voted in every Republican primary since 1980, subscribes to Guns and Ammo and shops at Bass Pro Shop will likely skip that household for the next. The sharp increase in polarization in American politics isn’t accidental.