Nate Silver, the quant savant who made waves with his accurate election predictions in 2008 and 2012 and who now works for ESPN, came to Wednesday’s 4A’s Data Summit looking to qualify not only his own successes, but also the rush to proclaim data as the magical solution to all future problems.
In an intellectual and entertaining 30-minute address, Silver sought to dispute the notion that “size conquers everything else” — that Big Data, by virtue of its existence, can lead to a “magical solution.”
“All the good data in the world wont help you if you’re using faulty assumptions to process it,” he said.
Silver’s talk was punctuated by three suggestions for advertisers and marketers as they move to more fully incorporate data.
1. Think probabilistically
People generally have trouble grasping probability, Silver said. When something is given an 80% chance of happening, many treat that as if it means 100%. But Russian roulette has an approximate 80% success rate, yet few would give that a try. If you’re confident in certain outcomes, go ahead and project that confidence, but sometimes it’s better to show some humility than to be humbled when things don’t turn out as you expected.
“Thinking probabilistically is a way to reconcile what we don’t know with what we do know,” Silver said.
2. Know where you’re coming from
Drawing on a sports analogy, Silver referenced Michael Lewis’ “The Blind Side,” offering the theory that in football, a weak offensive line can make even the most talented quarterbacks and running backs appear weak.
“You’re defined sometimes, in very competitive realms, by your weakest links, so it’s important to note what things you might be doing wrong — where your ‘blind spots’ might be.”
Considering your assumptions and recognizing your own potential for bias is essential for people in the data world.
3. Try and err
Silver said Google was the most successful user of Big Data, because they are “constantly iterating their products,” experimenting and pulling back when things don’t turn out right. They test their results in the real world, and have the discipline and courage to change course when things go wrong.
“Focusing on processes can give you better results,” Silver said.
He closed by quoting Duke University scientist Michael Babyak, who said that “in science, we seek to balance creativity and skepticism.” Silver said that the same principle could be applied to advertising. Be creative and aggressive in your campaigns and use of data, he said, but don’t be afraid to abandon ideas if they aren’t helping an organization achieve their critical objectives.