Follow Datanami:
April 11, 2013

Sprint Taps Streaming for Predictive Capabilities

Ian Armas Foster

Telecommunications company Sprint has over 53 million people using their networks. With the help of IBM, Sprint is hoping to apply analytics to data culled from those millions of users.

“One of the things that’s really intriguing us now is big data analytics,” said Von McConnell, Director of Innovation and Advanced Labs, Sprint. “The goal is to look forward in a predictive mode and make adjustments in real time.”

A 53-million person sample size, each with their various interests, transactions, and resulting data points, is a mammoth undertaking even in an age where big data initiatives are more commonplace.

The first part to that undertaking involves collecting the data—a  task that is not insignificant when considering those millions of users spread out over the multitude of systems that Sprint employs. “The challenge that we face historically is that we have so many systems,” McConnell said.

“We’re trying to take all these systems and understand how all these systems work together. So what we’re trying to do with IBM is work with their streams product and first be able to capture some of this data. We’ve never been able to do that before,” McConnell explained in the IBM video below.

According to McConnell, Sprint has the potential to track and analyze billions of transactions. With IBM’s stream computing technology, they may be able to do that. In the effort to achieve that goal, Sprint plans on introducing real time analytics to their core processing network, such that the time between data collection and analysis is reduced.

“We’re coming up on working with an architecture model that allows us to look across these systems and coordinate the data appropriately.”

The goal, as is often the case with big data analytics initiatives, is to eventually shift from a reactive to a predictive model. In other words, Sprint would like to understand their customers’ transactions before they happen to the extent where they can nudge people toward such transactions.

There are some intuitive applications to this, such as offering more customized plans that further appeal to individuals and families. Further, this technology could be used to dig into payments, as many companies waste money tracking down those who pay their bills late, but consistently late. Understanding those patterns would allow a telecommunications company to leave those customers alone for the most part and focus their resources elsewhere.

“We’re trying to use these systems to understand behaviors, understand intents,” McConnell said.

Sprint isn’t the only company tying to apply big data analytics to help them understand their customers’ behavior. Retail vendors with a significant online presence like Target and Amazon frequently use similar analytics to point customers toward products they may be likely to purchase.

With as big a user base and as diverse as data that Sprint will have to collect, the effort will be a significant challenge. Early returns are, promising, however, for as a result of IBM smarter computing, Sprint has seen a 90 percent increase in data capacity.

Related Articles

Preventing Brain Injuries with Predictive Analytics

IBM: Five Use Cases Driving Big Data Adoption

Big Data – It’s About Real Time

Datanami