Why ANZ separated its data platform and analytics teams

ANZ Bank has separated its enterprise-wide data technology unit from its business unit-based analytics teams as part of a digital strategy introduced by chief executive Shayne Elliott.

According to the bank's general manager of data technology Darren Abbruzzese, the appointment of Elliott in January led to an increased internal focus on digital innovation and data as part of ANZ's wider strategy.

“Previously, data in the bank hasn’t been treated as a first-class citizen. It’s been seen as something on the side [like] ‘yeah it would be great if we could mine x, y, z to get some customer insights into customer behaviour or credit quality’,” Abbruzzese told the Connect Expo conference last week.

“We also have those external threats in the fintech space. There’s .. a lot of start-ups that want to eat banks’ lunch, especially in the payments area.

“In six or seven years’ time, as [Elliott] hands the baton over to the next CEO, he wants to be remembered as the CEO who digitised the bank, and not just in the experience we provide to customers, but also our internal processes. So that has a major impact on technology and how we operate.”

As part of this digital focus, the bank created an enterprise data office, headed up by Abbruzzese, to centralise the bank's use of data as well as its quality and governance.

“One of my early jobs was to bring together a lot of different assets… I really want the business to be enabled through really solid platforms where the data is sourced into an appropriate lake [a Teradata platform], and there’s various technologies set up underneath that enable the particular use cases,” Abbruzzese said.

“Bringing it all together is a challenge, it’s something we’re trying to solve through a very clear platform strategy.”

But what the bank has opted not to bring all together is its enterprise data office and the individual analytics teams assigned to business units.

Abbruzzese said it was not the technology team's role to extract value from the data; that was the role of the business. The use cases for data also differ markedly across the company, he said.

“Some, within our financial markets division, will want to use data that is updated at sub-second level for pricing instruments, whereas our customer marketing data will be on a batch basis, so we have very different requirements and we work closely with the analytics teams,” Abbruzzese said.

“We have to ensure sure we meet a diverse range of needs, but also at the same time, we build a capability where as new needs emerge."

In order to make the new structure work, Abbruzzese said there needed to be a clear delineation between responsibilities as well co-operation between analytics teams and the enterprise data office.

“It’s really about realising you work in a network and not a hierarchical structure where it’s about having a culture of real collaboration, and agreeing upfront that ‘this is what you’ll do because it’s what you’re good at, this what I’ll do because it’s what I’m good at’," he said.

“If the Teradata platform falls over because I forgot to buy enough disk space, that’s my problem. But if the business isn’t getting enough value out of analysing its mortgage data, then that’s their problem.

“If we can create a network between those two where they can work together and appreciate what everyone else wants achieve, you get a much better outcome.”

Real-time analytics

ANZ is moving away from an earlier view that a centralised enterprise data warehouse can solve all its needs to one where big data is complemented with other technologies, most notably real-time analytics.

“Moving past big data, which is still by and large batch based, and getting into real-time data platforms is one of our challenges and one we really want to focus on,” Abbruzzese said.

“We think that being able to engage with customers in a more real-time way based on their behaviours, spending patterns or the way they go through our website with live support, we feel that will provide a lot of value and they will appreciate that."

One example of where the bank is looking to apply real-time analytics is in its website search results.

“So we know you’ve logged on to internet banking, we know who you are, and as you’re starting to go through our website, how can we tailor that in real time for a pattern and the type of customer that you are?"

Andrew Sadauskas attended the Connect Expo as a guest of Intel.