The sport of data science

How many analysts does it take to solve a problem?

That may sound like the start of a bad joke, but no one was laughing in 2006 when Netflix offered $1 million to anyone who could come up with a collaborative filtering algorithm that improved the performance of Cinematch (Netflix’s in-house software) by at least 10 percent. Cinematch predicts which movies Netflix customers like and makes movie recommendations to customers based on those predictions. The goal: boost customer satisfaction and retention along with sales.

Three years later, after receiving several thousand entries from more than 100 countries, a winner was announced, the $1 million prize was awarded and a cottage industry – online marketplaces for business projects where companies post challenges, provide data and offer prizes for the best solutions – took off.

While Netflix reportedly performed no formal cost-benefit analysis on the Netflix Prize, the company was clearly pleased with the results. At the time, Netflix CEO Reed Hastings said, “You look at the cumulative hours and you’re getting Ph.D.’s for a dollar an hour.”

In this month’s cover story, Margit Zwemer, a data scientist and community manager at Kaggle, provides an inside look at “crowdsourcing” – the concept that turns complex analytical problems into a competitive sport open to analysts, astrophysicists or anyone else who cares to submit a solution. As Zwemer notes in her article, the concept is not new; as far back as the 18th century, the British government offered more than £100,000 in prize money to anyone who could come up with simple and practical methods for measuring longitude to assist maritime navigation.

The Netflix Prize, however, helped turn crowdsourcing into a modern-day, mainstream corporate strategy. “Data research competitions are a resource-efficient way for organizations to solve complex data problems, and they create a meritocratic market for talent that changes the way analysts work,” Zwemer writes. Kaggle, an online platform for predictive modeling and analytics competitions, was one of many companies that jumped on the “crowdsourcing” bandwagon in the aftermath of the Netflix Prize. According to Zwemer, Kaggle boasts a worldwide online community of more than 40,000 data scientists and predictive analysts, competing under the slogan “making data science a sport.”