Big Data Not Really About the Data
TORONTO -- Big data can be a big flop if service providers put too much emphasis on just the data itself, not on what they want to achieve with it.
So said a group of data analytics experts at the Canadian Telecom Summit here this week. Speaking on a packed panel, the six industry experts told telecom and cable operators to focus on their core business values and customer needs first, not the tracking, collection and analysis of all the information streaming over their networks. The experts warned that if network operators don't heed this advice, their data analytics initiatives will likely either fall way short of expectations or simply fail altogether, as a number already have.
"This is not about technology," said Sanjay Kumar, general manager of global telecommunications industry for Hortonworks, which promotes a data analytics platform called Apache Hadoop for enterprises and other businesses. "It's about what's driving the technology, it's about where you see the business benefits of it."
Several panelists cited Telus Corp. (NYSE: TU; Toronto: T) as a classic example of how to do things right. The large Western Canadian telecom, wireless and broadband provider has scored by using advanced customer data analytics to reduce its churn rate and enhance services for customers. Due at least partly to such moves, Telus now claims that it enjoys the lowest subscriber churn rate among major service providers in Canada and possibly all of North America.
"We have kept the focus on customer value," said Lloyd Switzer, senior vice president of network transformation for Telus, speaking on the panel. "You have to start with focusing on customers."
In addition, the panelists recommended that service providers focus on optimizing their networks and developing the capabilities to manage and govern all the data traffic flowing through their networks. For instance, Switzer said Telus is now handling data from 20 billion voice sessions and 110 billon wireless sessions a year. With its annual data traffic load having doubled from 6 petabytes to 12 petabytes in just two years, the telco is aiming to boost its capacity to handle up to 30 petabytes of data.
Because network operators are faced with so much data flowing into their networks, panel speakers urged them to carry out other changes as well. Kumar, for instance, suggested that service providers shift from their current service delivery models to a more "data-centric" approach. With this new approach, he contended, the apps would come to the data rather than vice versa, minimizing how much the data moves around the network and making it easier to for providers to leverage the data in real-time.
Rajesh Nambiar, who heads the telecoms unit of TIBCO Software Canada, said service providers could take some tips from banks, retailers and airlines, all of which have taken advantage of data analytics to develop popular customer loyalty programs. "Banks and retailers are ahead of the telcos in big data," he observed, noting that telcos have plenty of customer data to use too.
Other speakers urged service providers to treat consumer privacy and security concerns very seriously. Several, in fact, advised providers to embed privacy and security programs into their data analytics initiatives from the very beginning to prevent network data breaches that could damage their brands and erode consumer trust.
"If you don't address these kinds of concerns, don't address what the customer wants, it's going to come back to bite you," said Ann Cavoukian, executive director of Ryerson University's Privacy & Big Data Institute. "So take this seriously -- the notion of trust is very important."
Citing a prime example of what can go wrong in the consumer privacy area, Cavoukian noted the $750 million class-action lawsuit that was filed against BCE Inc. (Bell Canada) (NYSE/Toronto: BCE) in April for allegedly tracking what its wireless subscribers browsed on the Internet and selling the information to advertisers without their approval. "Telus is great at that [putting the customer first]," she said. "Bell, not so much."
David Ritter, a partner and managing director at the Boston Consulting Group, said service providers should make sure they can analyze and interpret their data in a meaningful way. He noted that a Harvard Business School friend told him that the key is to hire economists, not statisticians, because economists know how to interpret data in the proper context.
"There are good ways and bad ways to interpret big data," he said. "The math is important. But understanding how to apply that math is probably much more important."
— Alan Breznick, Cable/Video Practice Leader, Light Reading
