Data scientists at Sprint see the company’s use of data, as “a natural evolution from where we started,” more than a century ago, said Von McConnell, executive director of the company’s Innovation & Advanced Labs.
So while the telecommunications giant continues to log transactions and managing systems, to the point where it is storing up to 70 terabytes, it also has scientists developing new ways to derive meaning and business value from all that data.
Speaking at the IBM Information on Demand conference in Las Vegas, McConnell said Sprint uses analytics to find ways to lower operating costs, and in that sense the applications are “a matter of salvation” in an industry which always looks to reduce expenses, McConnell said. Newer use cases of analytics are something else too: an opportunity for Sprint to develop insights about the behavior of its customers, to develop predictive models based on consumers’ collective behaviors.
“For us it’s not just about cutting costs, it’s actually about opening new product channels for us. So in an area where we’re moving from selling handsets and minutes of use, we think it’s a new way to create channels of interest,” McConnell said.
Sprint is not alone in developing new business opportunities for outward-facing applications from its internal use of analytics. The retailer Dollar General uses the 1010data data warehouse-as-a-service to analyze its consumers’ shopping baskets and then offers these insights to its suppliers, consumer packaged goods makers, to examine purchase patterns and identify ways to increase sales. In August, the New York City Police Department and Microsoft unveiled an offering for sale to other public safety agencies, using data-aggregation techniques and real-time analytics to combat criminal activity.
A Big Data Wrinkle for R&D
These efforts to combine and sift through different data sets to yield new kinds of insights take different forms; the Dollar General case is a retailer serving a syndicate of suppliers, while the public safety example is exporting expertise from one major city to a different jurisdiction. Taken as a group, projects that develop into data products and services are not new as a concept, but the big data trend of harvesting vast volumes and variety of data to create these products generates new opportunities, said consultant and author David Loshin, president of Knowledge Integrity, Inc.
“I would suggest that the ability to adapt the existing massive data volumes to addressing opportunistic problems is a paradigm shift when it comes to reporting and analytics,” Loshin said.
Loshin pointed out that taking advantage of such opportunities requires top-notch data management practices. “If I have a good data management strategy, then the materializing of these products becomes another factor of my development life cycle,” he said. The conditions placed on certain kinds of data represent another caveat: companies can’t develop new data services using digital materials that are off-limits because of data privacy terms or other contractual obligations.
Sprint is weighing these issues carefully, McConnell said at the IBM conference. Sprint is in the early stages of investigating whether it can work out the details with regulators and business partners to develop new products and services, McConnell said. One idea in the experimental stages, for example, would use predictive analytics to sift through patterns of consumer behavior to help retailers answer questions such as where people tend to go before and after a visit to a supermarket, or whether certain goods inspire comparison shopping.
At the same time, the telco doesn’t have to wait for potential customers—the demand for such a service already exists, he said. “A large retailer approached us, and said, ‘Can you tell me where customers were the two hours they were before they came into the store and where they went after?’ We can’t share data on individuals, but we can show trends.”
In an experiment, Sprint was able to analyze the data to find that shoppers going into one store and then traveling to another nearby store that sold similar goods, such as TVs. An inference would suggest there was comparison shopping taking place.
“In general, with analytics you can understand patterns of what people do,” McConnell said, adding that his team is asking, “How can we optimize services for our customer, if we know in general where people are going?”
McConnell was careful to say such insights were a long way from becoming a new line of business at Sprint. Privacy policies would need to be worked out, for example. “For any data analytics, privacy is one of the biggest concerns,” he said, adding that the company treats privacy issues “very conservatively” and that he is bringing this issue to industry groups to explore ways to address it.
There are other challenges to overcome. Sprint needs to speed up the process to access compressed data records in storage systems—something the company and IBM are working to improve. And then there is the question of a business model. “What is the price of the data, the value of the data?” McConnell said. “I’m trying to work within the industry so that [the parties can] move forward together…. We’re just discovering new channels, but we don’t know how to price it.”
Cloud-based Offerings on Tap from the New York Stock Exchange
Emile Werr, vice president and head of enterprise architecture at The New York Stock Exchange, was another executive at the IBM conference who said his company is looking to turn in-house expertise into a business opportunity.
Werr said his company, already in the business of spinning off technology offerings, is in the process of commercializing a division to take advantage of its secure infrastructure and systems integration prowess. Prospective customers for the division will be “anyone with big data problems,” that is, not just players in the financial services industry but others including life sciences, pharmaceuticals and retail.
“All those industries are dealing with the same types of challenges,” Werr said. “How do you get analytics done quickly, and how do you leverage the plumbing in an efficient manner. It’s all about elasticity, it’s all about agility. Our model is really going toward the whole cloud strategy. Anybody who is looking at integration being a key problem, and then empowering you to do the type of transformations they need to do, they will all be potential customers.”
Michael Goldberg is editor of Data Informed. Email him at firstname.lastname@example.org. Follow him on Twitter at