Updated: Feb 27, 2019
by Martin Lindstrom
n 2013, after analyzing billions of data, Google thought they’d discovered a relationship between how people searched for flu-related topics and people’s actual flu symptoms. The discovery sent shockwaves through the U.S. healthcare sector, which had been struggling with the balance between supply and demand. Once an outbreak occurred, doctors were often left without enough pharmaceutical products on the shelf. Suddenly, thanks to big data, Google promised them a flu-outbreak prediction four days in advance.
But this story took a surprising turn. The truth is, as the flu epidemic spread, fear spread, too. People became curious to learn more. The conclusion of the Centers for Disease Control, based on data from millions of doctors? Google’s prediction was a two-fold overestimate.
Google had focused on correlating data, but they missed causation. We have tons of data, but little information. Remember Hans Christian Andersen’s iconic fairytale, the Emperor’s New Clothes? Everyone blindly believes that the king, though naked, is wearing the most exquisite clothes.
“Our perception of the world is almost always local”
Get the magazine to read full article