Predictive Analytics Can Become Flim Flam


By Samuel Greengard

Nary a week goes by without news of some remarkable breakthrough in predictive analytics. Last week, Google announced that it has developed an algorithm that can guess movie box office hauls with 94 percent accuracy. The paper, Quantifying Movie Magic with Google Search, notes that Google can accomplish this task by examining the timing and category of Google searches and paid clicks.

There's no question that predictive analytics is the wave of the future. It holds enormous promise for organizations in a wide swath of industries, including financial services, retail and healthcare. It could revolutionize customer service by helping businesses understand when to replace components like mechanical parts, motors and batteries.

But, like any technology tool, predictive analytics can also undermine creative thinking—if it's not used effectively. Movie industry executives, for example, trot out sequel after sequel of predictable and mindless pap bursting with atrocious dialog—or, worse, no need for dialog. Within a few years, I'm sure we'll be at Star Trek 509 and Fast and Furious Part 47.

My point? Once bean counter executives get their hands on a predictive analytics algorithm that offers 94 percent accuracy, you can bet the feedback loop will become more entrenched than "You look fabulous" comments at a Hollywood gala. Executives will further engineer films for maximum marketing response and box office potential rather than any shred of creativity.

There's a problem with that. Greatness doesn't come from engineering products and services to meet the existing marketplace and what consumers think they want. It comes from thinking and acting in innovative and unconventional ways. Think Apple (iTunes, iPhone, iPad), Amazon (1-Click buying and the Kindle) and Starbucks (seamlessly integrated loyalty, e-payment and digital content channels over Wi-Fi). These companies excel at breaking new ground and taking business and IT where it has never gone before.

In Hollywood, it's the occasional exception—the film that barely makes it past the marketing gatekeepers or bypasses the entire studio process by taking the Indie route—that often emerges as the quirky blockbuster or the film that achieves a level of artistic greatness.

So, by all means embrace predictive analytics. Just don't use it to play things too safe—and justify preconceived notions or reverse engineer results, products and services. Otherwise, you will likely wind up with Disaster Movie Part II. You can bank on it.