For many organizations the promise of big data is the ability to deliver the right tailored offer to customers at the right time. But experts say it has to be done in a way that doesn’t invade customers’ privacy.
Can big data and privacy work together? Yes, says Ann Cavoukian, Canada’s respected privacy expert. And she points to Telus Corp. as proof.
“Telus walks the talk. Bell not so much,” she told a panel discussion Monday at the annual Canadian Telecom Summit in Toronto.
Vancouver-based Telus crunches billions of bits of customer and network data, yet does it in a way that respects customer privacy, she said.
And what some CIOs, CSOs, and CMOs might find surprising is the carrier does it while stripping customer data of all personally-identifiable information and identifying groups of people with similar traits or behaviours to put offers to.
From the early days of its big data project, Cavoukian’s Privacy By Design framework was baked into its processes, said panellist Lloyd Switzer, Telus’ vice-president of network transformation.
“The whole trick to analytics is looking at behaviour over time, or behaviour from multiple sources” to see trends,” Switzer said in an interview. In Telus’ case, that included everything from how much data customers use, and whether they come into the stores or phone call centres.
De-identifying data isn’t easy, he said. Creating a framework at the beginning of the analytics planning process is crucial, Switzer also said. “We know we can never identify a customer, so let’s make sure it’s masked in a way that can never be unwound.”
In the end, “you’re happy because you get the right offer in front of you, but I’m not targeting Howard the individual, I’m targeting the group that includes Lloyd and Howard and behaves in the same way.”
Proof of the carrier’s success, he added, is its low rate of customer turnover, called “churn” in the industry.
De-identification of customer data isn’t essential for data analysis, Cavoukian added. But, she said, if the organization doesn’t do that at the least it has to get each customers’ permission to use their data that way.
In his presentation, Swtizer said Telus [TSX: T] started its big data effort two years ago with the marketing team, the customer-facing team and the technology groups.
“We focused a lot on use cases and business value — this was not an exercise to deploy Hadoop technology or build a process. From the start this was all about how do we reduce churn? How do we produce the best offer in front of our customers? How do we improve the customer experience?”
They identified 14 projects that could be proofs of concept. They also decided on certain capabilities — like privacy, security and making sure the data is clean — that could be used across each of them.
“It’s so easy to get sucked back into ‘Let’s talk about the technology, let’s talk about all the challenges of getting the data, let’s talk about how do you get the data clean,'” said Switzer. “Those are really important things, but you have to start with the premise from the customer: How do we create business value?”
Its data store — including 37 billion IPTV records, 110 billion wireless data session records, 20 billion voice session records, plus social media interactions — is more a data ocean than a data lake, he added.
”I am a huge fan of data analytics, and big data,” said Cavoukian, who heads Ryerson University’s Privacy and Big Data Institute, “but you can’t do it in a way that ignores some basic business tenets like ‘Start with your customer, what does your customer want, what can you deliver.’ If you don’t have that dialogue you’re going to lose out.”
“What drives me crazy is that argument that people think privacy stifles big data innovation. That’s garbage. It’s limited thinking that stifles innovation.”
