Taking Political Control With Digital Analytics

How analytics have come to dominate the political landscape


‘Political organization is basically a matter of list-keeping. You canvas a state by foot and by phone to find out who is for you, who is against you, and who is uncommitted. Once you have the list, you cross off the ones against you, barrage the uncommitted with pleas and information, and make sure your supporters get to the polls.’

Hunter S Thompson wrote the above while following candidates in the 1972 Democratic primary. The principles of political organization remain very much the same. The reliance on lists is certainly still there. These lists, however, now contain substantially more information than when George McGovern ran against Nixon in ’72.

One of the main elements of Barack Obama’s victory in 2012 was a massive analytics drive, led by campaign manager Jim Messina. Analytics was applied to every aspect of the campaign, and data-driven decision-making was used to choose advertising buys, raise money, and model voter movements as the race came down to the wire. Messina later said: ‘We were going to demand data on everything, we were going to measure everything... we were going to put an analytics team inside of us to study us the entire time to make sure we were being smart about things.’

To achieve this, Messina employed a 100-strong analytics team, 50 of whom worked in a dedicated analytics department. A further 20 more were embedded throughout the headquarters, while 30 were sent out into the field. They worked their way through dozens of terabytes of data using a combination of the HP Vertica MPP analytic database and predictive models with R and Stata. Every aspect of the campaign was geared towards getting as much data as possible. All staff across the entire organization were regularly evaluated on their input levels, to ensure that they were putting in every last bit of data so it could all be collected.

While profiling potential voters is nothing new, it has never before been done on the scale seen during the Obama campaign. Messina’s army of data crunchers managed to create a so-called ‘megafile’ on potential voters. This didn't just show how to find voters and get their attention, but allowed them to run tests which predicted the specific kinds of appeals that would persuade different types of people. Using this information, they were also able to rank call lists in order of ‘persuadability’.

Analytics were also used in the field. Using Airwolf, volunteers out canvassing would record a voter's particular interests and feed them into HP’s Vertica database. The digital team would then run email blasts from the local organizer to voters that corresponded to a voter's favorite campaign issues, helping to target their messages.

Consumer data was also examined to determine who was giving money to the campaign. One of the most important aspects of Obama’s victory was fundraising. Before Barack Obama, the Republicans historically raised substantially more than the Democrats. Since 2008, however, the pendulum has swung firmly in favour of the Democrats, despite a number of heavy hitters such as the Koch Brothers backing the Republicans with substantial sums.

This was thanks in large part, to the substantial number of small-dollar donations raised by a digital grassroots campaign. Anyone who ever donated money to the Barack Obama campaign will be well acquainted with the deluge of e-mails it sparked, and they may - at least they would have been before the e-mails started - be glad to know that it worked.

The Republican campaign, meanwhile, saw no need to build the same sort of in-house data analytics departments until late-on, preferring to outsource it instead. Mitt Romney’s Digital Director, Zac Moffatt, said in July: ‘I don’t think we thought, relative to the marketplace, we could be the best at data in-house all the time.’ When they eventually did set about building an in-house team, it was a tenth the size of Obama’s. A massive gulf has developed between the Republicans and Democrats in terms of how many people they employ for their campaigns with experience in technology and data. Indeed, between 2004 and 2012, Democratic campaigns hired 503 staffers in digital, data, and analytics, while the Republicans hired just 123.

The Democrat’s focus on analytics is set to continue into the next election under likely candidate Hilary Clinton, who it is believed will hire 1,000 ‘data geeks, techies and digital gurus’ if she’s successful. And the Democrats already have a well-oiled data analytics machine and wealth of expertise in place to build upon. It is likely that it will be some time before the Republicans are able to catch up, even if Donald Trump has thrown his toupee and $8 billion chequebook into the ring.


Read next:

Working At The Boundaries Of Aesthetics And Inference