The Margin of Human Error
What product teams should learn about data from Trump’s surprise victory
--
The world was shocked this week as Donald Trump emerged victorious from the 2016 presidential election. How could this be, when polls consistently showed Hillary Clinton ahead throughout the entire election? It’s clear now that the Clinton campaign missed major insights — despite all of their data — that should have warned earlier of an impending defeat.
Misreading data to this degree would be incredibly costly for a product team. And while Donald Trump’s campaign was not analytical in the traditional sense, his unique strategy for gathering insights nonetheless resulted in a surprise victory that product teams should examine closely.
It’s a mistake to disregard statistically insignificant data
Trump’s confidence in his ability to win can no longer be dismissed as mere bravado. He recognized and tapped into something that the Clinton campaign, media, and much of the world completely overlooked.
Trump leaned on insights hidden in plain sight
Despite piles of money, larger staffs, and scientific polling methods, many organizations missed a key understanding that Trump was able to intuitively recognize: that the electorate was deeply frustrated with the political establishment. The New York Times, who had calculated Clinton’s odds of winning at 85%, observed in retrospect:
“The misfire on Tuesday night was about a lot more than a failure in polling. It was a failure to capture the boiling anger of a large portion of the American electorate.” — Jim Rutenberg, New York Times
Ironically, Trump’s winning insight-engine was hardly a closely-guarded secret. In fact, Trump routinely cited the source that repeatedly gave him rich observations into the electorate: his rallies.
Pollsters disregarded what Trump embraced
Trump is notorious for ignoring process and going with his gut. When he presented observations about the size and emotional energy of his rallies as evidence for the campaign’s success, he was quickly dismissed by pollsters. The data wasn’t statistically significant. The sample was biased. The sample wasn’t representative of the greater electorate. But while there were obvious issues with the data’s integrity, that didn’t stop Trump from accurately sensing something more meaningful than what the pollster’s data could provide: that “the silent majority stands with Trump.” While their methods may have been more correct, pollster analysis was dramatically inaccurate while Trump’s gut instincts proved dead-on.
Hillary’s campaign didn’t know what they didn’t know
In Silicon Valley we have a deep faith in data’s ability to guide business and product decisions. However, bias and oversight of key variables can lead to imperfect data that paints an inaccurate picture. By the time Hillary Clinton’s constituents realized they had underestimated the insights Trump intuitively seized upon, it was too late.
It’s easy to disregard data that isn’t gathered with scientific rigor, but intuitive insights can be a great sanity check for raw data. At DoorDash, we have a business operations team that slices and dices all types of product data, but we also rely on less-official qualitative methods to surface new trends. Feedback from local teams operating on the ground is an invaluable barometer for the health of each market. When even a single Dasher on social media surfaces a frustration with our ecosystem that could lead to churn, we pay attention.
Data has a unique ability to build confidence in strategies that may ultimately prove ill-informed. Don’t underestimate the value of keeping your finger on the pulse of your audience through these types of unscientific sources. They can produce significant insights unavailable through analysis alone, the subtle nuances of which could make a huge difference in how you solve product challenges, approach a market, or win an electorate.
If you found this article of value please click the ♥ below so others can find this post.
Ryan Scott is a Senior Product Designer at DoorDash. He’s worked for Salesforce.com and Palantir Technologies, and has launched products at TechCrunch Disrupt, AWS re:Invent, and Dreamforce. Opinions are my own and do not necessarily reflect that of DoorDash.
