Big Data ROI Will Take Time, Clear Goals and Talent

To put it bluntly, big data is a big business, one that Wikibon expects to grow at a 31 percent annual clip and approach $50 billion in 2017 (up from $11 billion in 2012).

So far, though, a lot of that money has been wasted. Earlier this year, an InfoChimps survey found that 55 percent of big data projects fail.

Tips: How to Avoid Big Data Spending Pitfalls

10 Hot Big Data Startups to Watch: List and Final Rankings

Ron Bodkin, CEO of Think Big Analytics, a San-Francisco based analytics service and solution firm, says big data project failures can be traced to several root causes: No business goal, no alignment with business outcomes, insufficient budgets, poor planning and failure to recognize project scope. (Don’t forget about the data analytics skills shortage that will only get worse over the next five years.) Even those who do succeed can fail if the benefits of the big data project aren’t realized outside of IT to, say, improve operational efficiency.

Big Data Spending Focuses on Sales, But ROI’s in Logistics and Finance

A recent Tata Consulting Services survey, The Emerging Big Returns on Big Data, finds that firms around the world are focusing their big data investments on sales, marketing and customer service more than any other business function. The so-called “gold in big data”, Tata says, “exist[s] in numerous corners of a large, global company,” with the highest potential benefits attached to examining data about customer value and needs, product quality, campaign effectiveness and inventory tracking.

How-to: How to Use Big Data to Stop Customer Churn

Commentary: Data Analytics Will Fail If Executives Ignore the Numbers

Companies expect the best ROI, though, from big data projects for logistics/distribution and finance. Marketing ranked last among the eight business functions Tata tracked; marketing executives were most concerned about how their firms would organize and optimize data from disparate information silos, “reskill” IT to use big data technologies and, above all, handle the volume, velocity and variety of big data.

Bodkin sees several industries primed to benefit from investments in big data. Manufacturers can use test data to drive efficiency and improve cycle times. Financial services, an early adopter of big data, can integrate “less traditional,” unstructured data sets and conduct intraday trade analysis. Online services such as Quantcast—where Bodkin was vice president of engineering prior to leading Think Big Analytics—can measures all sorts of information about visitors of all sorts of websites.

Internet of Things, Healthcare Present Big Data Opportunities

Two fields in particular interest Bodkin. The first is the Internet of Things, where the capability to collect data from smartphones and a whole host of connected devices can improve sales, drive project management decisions, improve efficiency, reduce waste and drive what firms such as General Electric deem the “Industrial Internet.”

The second is healthcare, where “pockets of innovation” in wellness, genomic research and wearable tech are poised to make data processing a core part of the way that teams of physicians treat patients. The “engine of sequencing,” Bodkin says, “is going to require a radical shift.”

Related: 6 Big Data Analytics Use Cases for Healthcare IT

There’s lots of great wearable tech, but the challenge is integrating data from disparate devices and “trying to create a more blended picture” of a person’s health, Bodkin says. As if that isn’t tough enough, the next step is taking this (often unstructured) heart rate, diet, exercise, sleep pattern or geolocation data and painting a larger picture about a person’s overall health.

Healthcare IT’s struggles with interoperability and integration will make this a tall task, but the capability to provide individualized wellness recommendations—and to do it far more often than a patient’s annual physical—will clearly demonstrate the value of healthcare big data, Bodkin says. That, in turn, will encourage others to opt in to data collection initiatives.

Whats Next for Big Data?

The next steps for big data projects will involve a move from simply storing data to connecting it with the business, Bodkin says. This means predictive analytics, automated business decisions and, as the Tata survey respondents also say, using data as an asset to create newer, better product offerings.

This, in turn, will drive customer engagement, thanks to a more consistent experience across channels; customers will appreciate a company’s 360-degree view of their online, mobile and brick-and-mortar activity as much as the company does, Bodkin suggests. Making this work will require an agile approach that makes data available to the right power users and data scientists, he adds.

More: 5 Tips to Find and Hire Data Scientists

As big data technology matures over the next five years, the need to integrate largely single-service big data “point applications” in order to boost business value will raise the bar for business analysts. They’ll need a greater level of mathematical sophistication than is currently expected, Bodkin says. Fortunately, he adds, this sort of knowledge should become increasingly common—just as basic computers skills, once a major differentiator among employees, are now expected.

Brian Eastwood is a senior editor for CIO.com. You can reach him on Twitter @Brian_Eastwood or via email. Follow everything from CIO.com on Twitter @CIOonline, Facebook, Google + and LinkedIn.