Predictive analytics predates big data, of course, but the two complement each other nicely. According to Gartner, more than 30% of analytics projects by 2015 will provide insights based on structured and unstructured data.
Being tipped off to the future is always helpful, but what's the best course of action once you get a prediction? That's where prescriptive analytics comes into play. An emerging technology that goes beyond descriptive and predictive analytics, prescriptive tools recommend specific courses of action and show the likely outcome of each decision.
To its proponents, prescriptive analytics is the next evolution in business analytics, an automated system that combines big data, business rules, mathematical models and machine learning to deliver sage advice in a timely fashion.
One of these proponents is Ayata, an Austin, Texas, developer of prescriptive analytics software. The company began as a research and development effort 10 years ago in Toronto, Canada, and incorporated in 2009. Its customers today include such major tech players as Cisco, Dell and Microsoft.
[ Learn more about Big Data's Surprising Uses: From Lady Gaga To CIA. ]
With predictive analytics, "I tell you what will happen, and I leave it up to you to figure out what to do with it," said Ayata CEO Atanu Basu in a phone interview with InformationWeek. "But if you knew what to do with it, you wouldn't be in this predicament, anyway."
Prescriptive analytics, by comparison, addresses a broader set of questions.
"It answers what will happen, when it will happen and why it will happen," said Basu. "And then (it tells you) how to take advantage of this predictive future."
Despite its potential, prescriptive analytics is barely a blip on the business analytics radar screen. According to Gartner, it's used by only 3% of organizations today. "And in that 3%, we are talking mostly structured data, mostly numbers," Basu added.
The rapid expansion of unstructured data -- including video and audio feeds, machine data and social media streams -- may create a growth market for prescriptive analytics. And leading big data vendors are touting the technology as well.
IBM, for instance, released a brief promotional video last year that called prescriptive analytics "the final phase" in business analytics. The video also mentioned IBM's fledgling efforts with prescriptive software in the areas of smart grids and traffic control.
As Basu sees it, the IBM video "is a ringing endorsement of what we have been doing."
In addition to tech firms, energy companies are showing interest in prescriptive analytics. "Oil and gas is looking into it very heavily," said Basu. The technology, for instance, could help petroleum companies predict and prevent failures of electric submersible pumps that extract oil from the ground, both onshore and off.
Prescriptive analytics is based on a mathematical discipline called operations research (OR), which works in conjunction with business and domain rules frameworks.
"Operations research is what UPS uses to get you the right package at the right time," Basu said. "That's one of the reasons why UPS trucks only make right turns."
In "Five pillars of prescriptive analytics success," a recent article he wrote for trade magazine Analytics, Basu emphasized the critical importance of a synergistic relationship between OR and business rules. He also claimed that prescriptive analytics is crucial to the development of the Internet's big data-focused future.
"For the Internet of Everything (or the industrial internet) to reach its true potential, prescriptive analytics -- and the resulting decision automation -- has to play a pivotal role," he wrote.
An automated system that not only predicts events, but also suggests courses of action, is better suited to today's industrial (rather than consumer) markets, said Basu.
"There could be several consumer applications, but we're staying with the industrial market today for obvious reasons ... the value of the decisions are much higher," he added.
Returning to his oil industry example, Basu pointed out that a failed electric submersible pump could cost a company vast sums of money in a very short time.
For instance, if a busted pump means a loss of 2,000 barrels per day, and the price of oil is $100 per barrel, the company is losing $200,000 daily.
"So the impact (of prescriptive analytics) is immediate, and the value is immediate," Basu noted.
Companies want more than they're getting today from big data analytics. But small and big vendors are working to solve the key problems. Also in the new, all-digital Analytics Wish List issue of InformationWeek: Jay Parikh, the Facebook's infrastructure VP, discusses the company's big data plans. (Free registration required.)