Big data in big companies

By Jill Dyché, SAS

This past summer my report with Tom Davenport, Big Data in Big Companies, was published by the International Institute for Analytics. When Tom and I first discussed writing the report, we agreed that we needed to transcend the definitional chatter so common in today's big data conversations. We decided to focus instead on how companies were using big data. We wanted fewer platform comparisons and more delivery approaches; fewer definitions and more value propositions; less ideation and more execution.

We hunted down companies with big data stories, interviewing C-level executives or vice presidents at more than 20 large firms. Each of them had a compelling vision for big data’s potential, and each of them – and this was our only stipulation – had at least one active big data project in the works. 

We heard a lot about diverse value propositions for big data across industries and market segments, and what was happening on the ground. We also learned a lot about what companies weren't doing with big data, or where executives who were supporting big data initiatives were still on the fence.

Dyche sascom
Jill Dyché is the Vice President of SAS Best Practices, and the author of three books on the business value of information technology

What's happening with big data

Tom and I discovered that the companies who were underway with big data efforts were, despite high expectations, starting with focused implementations. Stories featured in our report included:

  • UPS, where telematics sensors in more than 46,000 vehicles help the package-delivery giant optimize its routing. In 2011 UPS saved 8.4 million gallons of fuel by cutting 85 million miles off of daily routes due to more efficient routing, saving millions in fuel costs.
  • UnitedHealthcare, which uses its Hadoop platform for high-performance text analytics on call-center data, allowing the health care provider to monitor service levels and determine which customers may need additional support.
  •, whose Customer Insights group uses its big data infrastructure to enrich personalization, advertisement and email marketing business processes for the fast-growing online channel, with an eye toward supporting omnichannel capabilities for the company at large.
  • GE, where sensors on turbine engine blades can reveal patterns in part degradation or breakage, allowing the business to foresee repairs and adjust turbines or replace parts before they fail.
  • Sears, which is making significant investments in real-time data acquisition and integration using big data solutions. Much of retailers' big data strategy focuses on the economies of scale resulting from replacing legacy ETL code and accelerating new and existing business processes.

The executives launching these and other projects explained the appeal of big data's low barriers to entry. They were satisfied that they had the support of their companies' top managers in their big data endeavors.

According to Bill Ruh at GE, a number of senior executives, including GE CEO Jeff Immelt, "saw the benefits of the marriage of machines and analytics." They were also careful to mention the deliberate coexistence of new big data technologies with their incumbent IT infrastructures, resulting in a set of capabilities that Tom describes in the report under the heading "Analytics 3.0."

We were impressed with both the speed of big data technology adoption and its delivery, as well as the ongoing support these sponsors were getting from colleagues both above and below them. Without exception, our interviewees acknowledged that their nascent big data projects represented the tip of the iceberg. One top-five property and casualty insurer even considered its big data program to be of sufficient competitive advantage that it would not agree to be mentioned in our report.

What we didn't hear

A few companies were experimenting with simultaneous quick-hit big data projects, inviting new data exploration activities where "fail fast" had become part of the management lexicon.

For most, though, big data projects were being viewed as test cases for larger or successive efforts that would use newly acquired unstructured data and previously untested platforms. Big data was fresh-baked but nevertheless topped with a large dollop of wait-and-see.

This, in turn, drives new conversations about process changes, organizational structures and job roles. More than one executive acknowledged that it was easier to acquire new technologies than it was to acquire new skills. This was particularly acute for data scientists.

"I'll be honest," shared one V.P. "It's hard to find data scientists on the open market. So we're figuring out who we should train." Others acknowledged that the boundaries and skill sets necessary for data scientists were still unclear. They were balancing the industry's evolving definitions with their own internal knowledge and practice gaps.

We didn't hear much about enterprise commitment to open-source software. While foundational to most big data efforts, managers made it clear that open-source projects were specific to big data and not a systemic IT standard. Inarguably a growing IT trend, open source seemed to be confined to isolated IT efforts at most of the companies in our study, with big data efforts serving as an on-ramp.

Nor did we hear about big data program plans. Initially I thought this was a bad sign, indicating lack of firm commitment to long-term big data investment. But during my chat with the Top-5 insurer mentioned above, it hit me: Rigorous program planning for quick-hit discovery projects could, in fact, thwart the agility many early adopters considered one of big data's big selling points.

Indeed large companies get a bad rap these days for lack of innovation, as startups are celebrated for their nimbleness and collective commitment to institutionalized ideation. But big companies ingesting device and sensor data, setting up advanced analytics labs, embedding data management into deployment, and formalizing decision-making processes IS innovation. And, increasingly, big companies have the business results to prove it.

Bio: Jill Dyché is the Vice President of SAS Best Practices, and the author of three books on the business value of information technology. Download her Big Data in Big Companies report, co-authored with Tom Davenport, here:

sascom magazine logo 50% gray

Read more:

Back to Top