Wednesday, February 6, 2013

Is Big Data Killing Democracy?

NOTE: The point of this entry is to raise the question of how analytics are being used in the political process.  The techniques described and concerns raised apply to all  politicians and all parties. This is not meant to question any specific candidate but rather the system as a whole and the way that it is using data and analytics.

The 2012 Presidential election will go down in history as a turning point in American politics.  While much attention is placed on the outstanding job that President Obama’s analytics team did, other candidates jumped headlong into the deep end on analytics as well.  In the future, we will see increasing investments of a campaign’s time, people, and money placed on crunching numbers.

Generally speaking this shouldn’t be a problem.  Good analytics give greater insight into customer needs and ideally allow organizations to improve their products and services to meet the needs of those customers.  That should bode well for the American people.

The problem is that in the election analytics weren’t used to create a better product or idea.  Instead, they were used (by both sides) to efficiently and effectively manipulate individual behavior.  In others words, the data weren't used to get the candidate to respond to better to the person.  They were used to get the person to respond better to the candidate.  So what's the big deal?  Isn't the the whole point of marketing?  Candidates have been trying to achieve this since people started being elected to political office.  However, when our understanding of voters and voting behavior was more crude, candidates had to think and focus broadly to win votes in order to hedge their bets.  However, now, we've reached a level of sophistication i understanding and predicting behavior that fundamentally changes the dynamic between candidate and voter.  This creates three problems:

1)      The electorate becomes resources to be optimized in support of the candidate's needs rather than the candiate being a resource to help his or her constiuents.
2)      Candidates can now win (and win much more efficiently) by reaching out to fewer rather than more of their constituents.  In fact, true resource optimization is about reaching the fewest people necessary to win.
3)      Candidates can  more effectively (they’ve always tried) manipulate voters into basing their decisions on things other than the candidate’s policies, abilities, and vision.

In this past election data weren't being used to understand people’s needs in order to formulate policy.  They were being used to manipulate people's attitudes and actions.  As Sasha Issenberg stated in her article, How President Obama’s campaign used big data to rally individual voters:

"The campaign didn’t just know who you were; it knew exactly how it could turn you into the type of person it wanted you to be."[1]

Isn’t that backward?  Our government and our leaders are supposed to be serving our interests.  We are not here to serve theirs.

Advances in accurately predicting voter behavior at the individual level allows candidates to pay attention to fewer and fewer people.  This played out clearly on the election with repeated campaign stops in certain locations and almost no attention played to others.

In politics, uncertainty necessitates being inclusive.  If a candidate doesn't know for sure where someone (or a group) stands or how they will behave, he or she has to try to reach more people.  On the other hand, certainty enables exclusivity.  Candidates now know who is more and less likely to respond to their message and will use their time accordingly.  As a result, candidates will increasingly reach out to people with whom they are already have an ideological connection and will miss the opportunity to hear and take into account opposing views.  In the name of expediency and votes, the emperors of tomorrow may never encounter the people who see that they have no clothes.

For example, suppose a candidate’s team had the following data for a group of 100 voters:

% Favorable
% Expected to vote
Expected votes
Candidate A
* Actually, 3.15 but only “whole” people can vote

Candidate A has four choices (in increasing order of difficulty)

·         Try to get four more of his or her supporters to get out and vote
·         Try to win over the three undecided voters who are likely to vote
·    Try to win over and convince enough of the undecided voters who aren't voting
·         Try to win over two of the opponent votes

Reports from the election indicate that the first two options seemed to be the preferred approach.  This approach allows the candidate to exclude over 80% of the other voters since they don't have to worry about those who are already voting for them and they don't need the votes of those who they know are not.

This type of strategy makes sense in business.  Businesses calculate the "lifetime value" of a customer.  This helps them determine how much, if anything, they should invest in acquiring and retaining that customer.  After all, why invest $50 to acquire a customer who is only "worth" $10 to you.  This helps them maximize the use of their resources.  It's good business.

The problem is that the same principle falls apart in politics. Do we really want our political leaders to determine whether to engage us based on how much they believe we are worth to THEM?   Don't they work for us?  The flap over Mitt Romney's statement about not focusing on the people who weren't going to vote for him drew rightful criticism.  However, a large part of President Obama's analytic's success was based on the same premise.  Whether doing it through simple, "old school" policy or a sophisticated analytical model it's wrong.  Unfortunately, the new analytical tools make it easier and more extreme than in the past.

Finally, for those who are targeted, it’s going to be easier to manipulate their behavior. The winners of future elections might not be the people with the best ideas for the nation but rather the ones with the best marketing departments and targeted and tested micro-messages.  Zeynep Tufekci, in his Wall Street Journal Op-ed, Beware the Smart Campaign[2]s states

“What I really worry about, though, is that these new methods are more effective in manipulating people. Social scientists increasingly understand that much of our decision making is irrational and emotional. For example, the Obama campaign used pictures of the president’s family at every opportunity. This was no accident. The campaign field-tested this as early as 2007 through a rigorous randomized experiment, the kind used in clinical trials for medical drugs, and settled on the winning combination of image, message and button placement. I agree that his family is wonderful and his daughters are cute. But an increasing role of “likability” factors, which we now understand better how to manipulate, is not good for democracy.”

This election was won by skillfully and expeditiously manipulating the margins  (and this is not meant to be an indictment against President Obama’s campaign.  Mitt Romney’s campaign tried as well as did all of the other candidates throughout the primaries.  The President’s team just did it better).  As analytics help candidates hone those margins to even greater levels of granularity, what will happen to those of us who are statistically determined to be “in” or “out” and therefore don’t warrant attention.  Will our voices be heard?

I know I’m being na├»ve but I hope that our politicians will see the power in analytics to isolate the core issues that unify us as a nation instead of continuing down the path of using data to separate us to get a vote.  Personalization is great for businesses and consumers.  But, as the President himself has said many times, we are supposed to be a collective society, not a group of 120 million individual voters, each who received their own special, personalized promise from the government.  That’s not sustainable. A business might be able to personalize a computer or car to a customer's exact specifications.  However, the government can't make 120 million personalized laws and polices.

Big data and analytics provide a tremendous opportunity to provide better products and services to customers in more efficient and effective ways for companies and governments.  However, as with any tool or technology leaders still must challenge themselves to use these tools in morally, ethically, and socially responsible ways.

Brad Kolar is an executive consultant, speaker, and author.  He can be reached at

[1] Sasha Issenberg, How President Obama’s campaign used big data to rally individual voters, December 16, 2012 (
[2] Zeynep Tufekci, Beware the Smart Campaign, Wall Street Journal Op-Ed, November 16, 2012 (