Allied Irish Bank blends industrial and artisanal analytics to gain edge

Bank has bulked up its data analytics capability, partly with SAS, to gain a more intimate understanding of its customers

Allied Irish Bank in Dublin has been muscling up its data analytics capability over the past three years, as part of a drive to gain a more intimate understanding of its 2.3 million customers.

CEO Bernard Byrne put the bank on a path to being more data-driven three years ago, and Peter Swan, Allied Irish’s head of customer contact and data modelling, says: “We had an existing capability in data, but we’ve now turned that more towards getting a better understanding of the customer, and away from anecdotal decision-making.”

The vast majority of the bank’s customers are in the Irish Republic, with some in Northern Ireland. It is one of two banks that the Irish government refers to as a “pillar bank” – the other being the Bank of Ireland.

Allied Irish has had an enterprise data warehouse, from Teradata, for about two decades. “We’ve had a very good single customer view, a good basic infrastructure that collects our data,” says Swan. “The next step was to take much more advantage of that.”

The data team within the bank’s customer analytics department has almost doubled in size in recent years, from 37 to 72 people. Their data analytics efforts combine artisanal, data scientist-type analytics with a more industrial version.

“We have data scientists in the team who are coming up with ideas in an ad-hoc mode, but then those ideas progress into a robust, fully functioning, production piece of code,” says Swan.

“We have brought in a new generation of people coming out of college. We are also encouraging different ways of working, more agile ways. And we are developing a more agile mindset, so we stop thinking like bankers and become more creative.”

The Dublin and wider Irish IT recruitment market is busy, he said, with Allied Irish having to compete with Facebook, Google and Airbnb for local talent.

Read more about data analytics in banking

Recruiting so many young graduates led the bank to consider open source as well as SAS technology because so many millennials are used to R and Python as languages for data analytics work. “The SAS functionality and the thought leadership swung it,” said Swan – away from open source and also IBM.

“We were looking for a partner, not just a technology provider,” he said. “We did debate open source at length, but because we wanted to scale up and run everything in a robust environment, we preferred to use a vendor-supplied tool. So, if I lose a staff member, things will still work. It’s not something made up of bits that people have built [in-house].”

The team is composed of a mix of people, he said, drawn from across the British Isles and continental Europe.

The bank has what it calls an “analytics factory” where it deploys SAS tools that enable the team to take a business query and explore available data to provide insights. They then create, deploy and maintain analytical models.

These then run continuously in the background, indicating what a next best action could be, says Swan. “For example, if you walk into a branch, we already know to talk to you about a personal loan. Or when you access your account on a mobile app, you’ll be told where the nearest branch is. It’s less about being a ‘digital bank’ and more about being omni-channel. We’re all digital banks at this stage.”

Allied Irish’s data management estate comprises the Teradata data warehouse and an IBM implementation of Hadoop on the back end. The modelling tooling is from SAS, and visualisation tools from the same supplier are used to work with stakeholders to “get to the nub of a problem”, says Swan.

From a reporting point of view – for branch managers and call centre workers, for instance – the team use Qlik. And for campaign management, they use some Teradata software. “I’d say from SAS, as well as the tools, we get thought leadership too,” says Swan.

Read more on Business applications