Editor's Note: Noise Pollution: Why more data isn't better data
I'm writing this note on Election Day Eve, and while the world is eagerly reading Nate Silver's political predictions in his excellent New York Times blog FiveThirtyEight, I've been tucking into Silver's new book, "The Signal and the Noise: Why So Many Predictions Fail — but Some Don't" (Penguin Press).
Silver holds an exalted position as a high priest of political statistical analysis; his model for crunching polling data in 2008 proved eerily accurate. As publishing moves into an era of big data, there's much to be gleaned from Silver's cool approach to information. Silver's stance is that, contrary to conventional wisdom, more information is not necessarily a good thing, and certainly not right away.
Silver tells us: "The original revolution in information technology came not with the microchip, but with the printing press." While Gutenberg's invention enabled the Industrial Revolution, the European Enlightenment and the American Republic, Silver notes, "the printing press would first produce something else: hundreds of years of holy war."
Turns out all that new information was creating confusion, and factions around the ways people chose to interpret it. The deluge of ideas—good, bad and ugly—was outstripping our ancestors' abilities to absorb and comprehend it.
It's an apt warning for what's happening in the wake of the recent computer- and Internet-driven information explosions. And it's a good lens through which to view what we're experiencing in publishing.
Publishing's digital revolution, in addition to creating strong factions around the way books and other forms of digital content are bought, sold, marketed and consumed, is dumping an information age's worth of data into the industry. We're overflowing with aggregate data about reading and purchase habits, and very granular data about how far into ebooks people read, how they perform on test prep questions, what passages they highlight and what products they've viewed on the Internet.