"I'm finding you can attract really great data scientists to ideas where you're applying the technology they've been working on to a great idea," Jobanputra says. "The technology definitely has to be there, but making sure you have the right expertise and are addressing the right pain points is really important."
Mason adds that big data startups need to make sure they remember the 'science' part of 'data science.'
"Most companies say they're data driven, but they're actually using their data very poorly," she says, noting that they collect data but then optimize it, subverting the scientific method. "They don't really sink their mind into the scientific method. They lack an experimental process."
Sign up for CIO Asia eNewsletters.