Data and Donald’s triumph – the math of elections
As you read this, the world is yet to recover from the US elections. Everyone – journalist, analyst and layman – is still attempting to figure what went wrong (or right.) The first reaction, is to trash experts and these new-fangled analytics algorithms. But as a practitioner of this arcane science, I beg to differ.
I believe this happened because humans prefer to hear what they like; the Democrats, transferred that prejudice to math. So here’s my take on Hillary, Donald and math.
The change – from Cleisthenes to Trump.
I love elections. Not surprising, as I grew up in the Calcutta of the pure whites of Jyoti Basu and the poor reds of communism. We marched our hearts out in those innocent days, because we believed that elections hinged on the motivations of the electorate, on the zeitgeist of the day. But that’s not what elections are, any more.
When Cleisthenes sold direct democracy to the ancient Athenians, he led with an idea and then asked for votes. But what President Obama had, was Narwhal. Narwhal was not a senior Democrat but a data-platform employed by the 2012 Obama campaign – integrating data for functions such as customized emails for fund-raising, identifying likely-voter clusters and using them via social media as influencers. In short, first come the numbers, then personalisation of ideas.
Now meet Hillary’s Ada. Ada (romantically named after Ada, Countess of Lovelace) is the most sophisticated electoral dataplatform, ever. Ada was so precious, it had a dedicated server – not what the rest of the campaign was using (no private server jokes, please).
Ada decided everything. From Hillary’s frequent visits to Pennsylvania to Jay Z and Beyonce. So what went wrong? The data? Or the interpretation?
If you ask me, the humans went wrong.
Man bites math.
Ada was sophisticated, yes. But what we don’t realise is that most data-sets and forecasting models often carry their creator’s predilections. And, they often interpret what they unconsciously want to hear.
Trump, uncomfortable with the numbers his analysts were sharing, kept pushing back (10 days to elections, they were still rewriting sampling methodology). While Hillary, happy with the numbers she saw, kept getting more of the same.
We refuse to challenge results that mimic our opinions; we find comfort in data. Unfortunately, that might just be something we fabricated ourselves.
The politician as analytics-driven marketer.
Today’s campaign analyst uses math and statistics to predict voter behaviour. They are armed with behavioural psychology, social media strategies and the results of randomised experiments. The information we provide on social media helps them analyse public mood – to decide everything from candidate selection, ideological stances, campaign ideas, personalised targeting and predicting voter turnout. And then they use this data for targeted campaigns
– the politician has evolved into a marketer.
Today politicians employ analytics-led customer marketing – like any toothpaste brand manager. They focus on branding, market research, voter segmentation, use of imagery, targeted and personalized communication.
Today’s politician uses the merging of cognitive and behavioural patterns. The citizen in an advanced economy is trained to think as a consumer and this thinking spills into other life situations that require decision-making – like voting.
But is voter, toothpaste?
The marketer offers customers choice – to take an informed, individual decision. What elections offer citizens, is a responsibility that is collective. With micro-segmentation, citizenship could be eroded by selfish, individualistic choices; ‘customer behavior’ could kill ‘citizen behavior’.
Also, with every citizen participating in social media, we could be getting the very idea of representative democracy wrong. We could degenerate from representative democracy to a direct democracy – neither ideal nor practical for complex democracies. (Proof, in one word – Brexit.)
The future will definitely bring uncomfortable questions. About (manipulative) math and elections. It would help to remember then, that democracy has never been about meritocracy. It’s just about numbers. Apologies, but it’s true.
Democracy has always been about numbers. The difference between the ancient Greeks and us is that with science, we can play numbers better. But, as the Hillary campaign now knows, the numbers should never know of our biases.