Data Science Darling Kaggle Cuts One-Third of its Staff


Kaggle—the San Francisco startup that has spent the past five years trying to push hardcore data science across the business world—has cut about a third of its staff as it tries to find new ways of making money.


The company has cut seven of its approximately 20 job positions, according to a person familiar with the matter, who requested anonymity because he or she is not authorized to speak with the press about the company’s situation. Kaggle CEO Anthony Goldbloom declined to comment on any job cuts at the company. But he says that Kaggle is shutting down its energy-industry consulting business, first launched about a year-and-a-half ago.


Kaggle will continue to develop its popular competition platform, he says, and it plans to announce a re-focused business plan sometime in the next few weeks. “We’re not ready to unveil just yet where the company is heading,” he tells WIRED.


The news raises questions about the commercial viability of the company and the popular data science competitions it runs on its website. Data science has attracted huge attention in recent years, but some question just how useful it is, and Kaggle’s struggles may point to a larger hole in this movement.


Kaggle is best known for running competitions that invite anyone and everyone to contribute their own algorithms, solving a variety of problems. The startup Jetpac used code that had won a $5,000 Kaggle prize to power its vacation travel app, and Microsoft is currently offering $16,000 to whomever can help it to a better job of classifying malicious software families.


The competitions typically attract a lot of participants. Microsoft’s malware contest, for example, already has 167 entries. But insiders say that they haven’t been lucrative enough for Kaggle. And the algorithms that win them aren’t always general enough to be useful to the company sponsoring the competition.


So a year-and-a-half ago, Kaggle pivoted and started developing building out an energy consulting business. The idea was to leverage Kaggle’s big science expertise into lucrative contracts with oil exploration companies and the like—and to eventually build out more lines of business serving other industries.


“This decision was motivated by the fact that we believe there’s a large market for end-to-end solutions that can’t be satisfied by the competition product alone,” Goldbloom in a Quora post at the time.


But Kaggle’s energy business suffered from a downturn in world oil prices and a slowdown in oil exploration research and development, Goldbloom says.


With $11 million in VC funding, Kaggle has struggled sometimes to line up big contracts from large corporations who are often reluctant to subject their data to the kind of open scrutiny that comes via the Kaggle contests. “The major concern is even if you can obfuscate data, you will have a hard time from your legal department,” says Alexander Linden, a research director with the Gartner analyst firm.


Kaggle’s main problem right now may be that it is simply ahead of its time. “The idea, I think, is sound,” says Linden. “You just cannot expect a 200 percent growth rate in this area.”



No comments:

Post a Comment