Trump or not, 'big data' could be huge in 2016 vote

"Big data" could play a huge role in the 2016 US election, even if Donald Trump doesn't think so.

Trump, who sailed through the Republican primaries using unconventional campaign rallies and Twitter messages, has indicated that he sees little use for popular data analytics tools to help target specific voters.

But analysts say he may be at a disadvantage in the general election if he stays on that course.

"In the primaries, he was only looking for Republican voters, and in the general he needs all voters," said Alan Rosenblatt, a digital political strategist.

"Unless he starts to be more sophisticated in how he targets his message, he's not going to have a good sense of where he's going to win and not win and where he should put his resources."

Trump's rise appeared to take the wind out of the sails of data crunchers and the techniques that helped the 2012 re-election campaign of President Barack Obama.

Trump is defying the traditional wisdom in both political parties, which have been ramping up the way they use data to help candidates and their backers personalize their pitch to raise funds and get out the vote on election day.

Jordan Cohen at the digital marketing firm Fluent, which works with candidates, said the Trump campaign "has exhibited the least amount of digital sophistication or interest" and that this could hurt his efforts.

Cohen said email fundraising, which has become a staple for candidates, requires considerable work to build a supporter base.

"Now that the Trump campaign wants to raise funds from the public they are certainly at a disadvantage by not having spent the past year building up a big database," Cohen said.

- Better data crunching -
The 2016 campaign could see more sophisticated data efforts thanks to advances in computing technology.

Campaigns have been speeding up how they collect and analyze data to form the so-called predictive models that can help determine tactics and strategy.

Consulting firms boast of having as much as 1,500 "data points" or bits of information about voters, and dozens of "voter models" that categorize individuals for targeted messages.

The data ranges from voter registration databases to public records on hunting or fishing licenses or gun ownership, to online "cookies" or small programs which track Web browsing and purchase activity.

"This is 'Moneyball' for politics -- it's about running smarter, more efficient campaigns, and the more data you have, the more effective you can be," said Tom Bonier, chief executive of TargetSmart, a political consultancy, referring to the book and subsequent hit movie using data to build a winning baseball team.

Because the campaigns can't reach everyone, they want to find the most "persuadable" people, according to specially designed formulas or algorithms.

"In 2012 for the first time we were able to successfully model persuadability," said Dan Porter, a member of the Obama data team who co-founded BlueLabs, a consulting firm now working with the Hillary Clinton campaign.

"We've had four years now and we think we are able to develop a persuasion model with more precision."

Contacting the most persuadable voters is critical for campaigns. Reaching out to those opposed may backfire by making them even more committed to an opponent.

- Beyond 'cookies' -
argetSmart, which works with the Democratic party and political action groups, said it has begun working with Facebook data that eliminates the need for tracking cookies.

The consultancy has matched 191 million voter file records to Facebook users to be able to deliver targeted messages to people on whatever device they are using, said Bonier.

Bonier said this is more accurate than using tracking cookies because it is based on user log-ins regardless of the device.

He noted that Facebook's platform provides important clues because it uses people's "likes," which can be important for political campaigns.

"We have a lot of our own data, but what Facebook brings to the table is the 'likes,' the attitudinal information." 

Some Republican candidates have used predictive models from a firm called Cambridge Analytica, which offers profiles of voters for so-called behavioral microtargeting.

Chris Wilson, a consultant for Republican Ted Cruz's campaign which used the Cambridge scores, argued that using this data is positive for the democratic process.

"You have so many people who have become disenfranchised because they do not believe the political system speaks to them," Wilson told a recent Washington forum.

"Once we are able to understand what an individual voter cares about and we are able to talk to them about that issue, it creates a level of connection between the voter and the candidate, and that is what politics should be about."

Wilson said Cruz -- who dropped out of the White House race -- effectively used these tools, but that Trump prevailed because he was able to reach more people with media coverage, which he equated to "$2 billion in free advertising."

But while better data can help campaigns operate more efficiently, the growing use of these tools also raises privacy issues, says Jules Polonetsky, who heads the Future of Privacy Forum think tank in Washington.

"There is a lot of data sloshing around the system that, if breached, could put the system at risk," Polonetsky said.

"Every piece of information tied to your name is available to these groups."

Source: AFP