How Scotiabank is tackling data analytics

Scotiabank's marketing data analytics capabilities are becoming an enterprise-wide asset as the organisation seeks to improve customer interactions and operational efficiencies, according to one of its VPs.

Speaking at this year’s ADMA Data Day in Sydney, the Canadian-based bank’s VP of information management customer knowledge and insights, Neil Freyke, explained how it has worked to first unite customer data in Canada, to support a variety of analytics capabilities across its Canadian operations. It is also exploring unifying data globally. The group operates across 55 countries with 21 million customers, and has made a host of acquisitions.

Scotiabank’s analytics team reports into the global marketing function.

Freyke said the division started with a ‘marketing information system’, pulling files off core banking systems in order to segment customers and explore cross-sell opportunities. As demand to know more about customer behaviour and history grew, his division created a ‘marketing data mart’ to pool information that could help drive richer marketing campaigns on monthly basis.

Marketers then began asking to see more transactions from different parts of the business, so Freyke’s team desensitised and joined up cross-functional information in an enterprise data warehouse.

More recently, the success of Scotiabank’s data-driven marketing efforts has prompted other functions to seek data to change how the company operates and seen the group embrace new technologies, he said.

“We now have so much data, and the traditional warehouse is starting to slow us down and is chunky,” Freyke said, adding that Scotiabank’s data pool is upwards of 1.5 petabytes.

“We’re evolving this to an analytics data ecosystem, using Hadoop to better manage data and elicit more insights off this data.”

Freyke said his focus today is not so much the volume of data collected, but how Scotiabank can use this information to support business imperatives.

“My group makes sure those analytics activities are baked into business process where they’ll be actioned and you can find business value out of that,” he continued. “This is an important part of my job as lots of new analytics capabilities come to the fore.”

To drive the enterprise-wide shift to data-driven decision making, Freyke said there has been a convergence of data domains and environments across marketing, treasury, risk management and business functions.

And while traditional marketing was once focused on using data for income growth through next best action campaigns and product pricing, that data is now supporting risk management, governance, fraud detection and operational efficiency, Freyke said.

“This convergence of capabilities is powerful when you look at how you are competing on global stage,” he said. “The capabilities we have built in marketing science now support other areas of the bank and we’re seeing cross-pollination of different areas.”

To help, Freyke has recently deployed analytics ‘SWAT teams’ that go out to business teams and help them leverage and get better value out of data.

The quest to be data-driven is not without its challenges, and Freyke highlighted diseconomies of scale through traditional analytics process and platforms, as well as executing and meeting market demands, as big hurdles.

“We need to be more flexible and engaged and drive value out to different business lines and channels,” he said. “New data technology presents new opportunities but also major changes for staff. A lot of new technology means we have to evolve, shift and change to gain new capabilities. New big data technology creates skills gap/labour shortages, and enterprise standards for big data across the industry are slow to catch up.”

Freyke has also been working with HR to become more aggressive in attracting and retaining talent. To do this, he said the team needed to educate HR execs about the data story, highlighting the challenges the function faced, and getting their buy-in to investigate different ways to retain them. These could include bonuses, different types of incentives, and providing an environment to be more innovative.

One structural change that has helped is the organisation’s move to an Agile approach.

“We found if we took the traditional model of waterfall, people were less interested,” Freyke said. “Now in each of our current initiatives [regulatory and next best action activity], our small team of data scientists act like ‘SWAT’ teams and solve problems fast. That encourages innovation and looking at things differently as well as executing better and faster.

“Those have been successful in attracting talent and where we’re seeing success with retaining staff.” When it comes to prioritising data and insights, Freyke said the key was starting small. “We didn’t look to bring in every transaction into Hadoop, we’re working with different analytic uses cases."

These include working with mortgage business lines to bring in mortgage specific data application and build learnings off that.

“My advice is to start slow and make sure you’re not putting too much risk on a platform that you don’t have skillsets to manage," Freyke continued. "I am worried people have unrealistic expectations around Hadoop – this requires a measured approach, starting small and focusing on something that drives value that you can build on.”

On the marketing front, Scotiabank is also moving to real-time engagement and decision models, which Freyke labelled “a right time model”.

“We’re taking data and looking at all customers and channels, then prioritise that and come up with offers that are relevant at the time of interactions,” he added.

Follow CMO on Twitter: @CMOAustralia, take part in the CMO Australia conversation on LinkedIn: CMO Australia, join us on Facebook: https://www.facebook.com/CMOAustralia, or check us out on Google+: google.com/+CmoAu