This is the second of four parts of a larger article I wrote for the Campaigns & Elections special edition on the CampaignTech conference, which is shipping with the current issue of the magazine. This piece appeared earlier in the C&E blog, and see also Part One, Retail vs. Wholesale Online Politics
Data-Driven Politics: Should Pollsters Be Nervous?
Let’s think more deeply about where data-driven politics is really catching on: Targeting and field. Not coincidentally, both areas are already havens for data nerds.
Direct mail mavens have been using voter files and consumer databases to target messaging and fundraising appeals for decades and data-targeting online ads is a natural extension of that mindset and skillset. Likewise with grassroots, since field organizers are consumed by numbers related to canvassing and its results. Walk-lists, after all, come from a database, and voter contacts yield plenty of information ripe for tabulation and analysis.
Polling is a third area of data-heavy politics. Online polling has its problems, chief among them the eternal issue of getting a representative sample of the population whose opinions you’re trying to measure. That’s where digital advertising comes in. Google and Facebook ads can help figure out what voters care about at a broad level, and campaigns can also use them to test which messages resonate with which voters in which contexts.
Three years ago I was on a panel with Eric Frenchman, John McCain’s digital ad expert in 2008. He talked about being able to see trends in the political environment days ahead of the rest of the campaign by looking at which ads were being clicked on. In 2010, my friend Soren Dayton was already combining Facebook and Twitter ads to test messages on particular slices of the electorate for the campaigns he worked on.