Disillusionment is a danger for the data science community in many organisations. A lack of understanding at senior levels of capabilities, timelines and dependencies can create a dangerous cocktail, when mixed with continuous hype in mainstream media.

As expectations peak, we can overcome this business concern by evidencing bottom-line impact of short, successful analytics projects. Euro returns buy more time for your data scientists to research and commit to longer term transformation.

Data Infrastructure

Bank of Ireland is on a journey to becoming a data-driven organisation.  We invested in Teradata infrastructure in 2013 to create a shared, accessible and integrated data source for all business use cases. Beginning with regulatory projects, over 60 Bank systems were fed in.

Tableau was selected as our data visualisation tool to provide enterprise-wide access to automated, standard reports with drilldown capability. The platform also enabled self-serve efficiencies and made information accessible to a larger, managed audience.

Value Realisation

Our CFO setup a Data Value Realisation initiative in 2017, to measure a return on this investment. Over the subsequent two years, Data Value Realisation delivered €58m of value, through using data to better meet the needs of our customers, and enabling internal efficiencies.

We also introduced analytics to new corners of the Bank and supported a collaborative, agile work environment. We now have over 800 recipients of fully automated data visualisations, and a significant backlog of requests for new products.

Our success was recognised by an Analytics Institute Business Value award nomination this year.

Challenges to Adoption

Like most organisations in financial services, we have numerous, static, Excel- or PowerPoint-based management reports which are distributed by email.

While some areas have introduced an element of automation or centralisation to these processes using tools such as VBA, SAS or local SQL servers, the static backward-looking descriptive statistics failed to sufficiently support strategic management decisions.

Hard copy reports were reviewed at decision forums, leading to requests for diagnostic analytics to understand what was driving the trends. Expert resources spent far too much time on data wrangling, report production and ad-hoc query responses.

Three Pillars Approach

The Group Data Office analytics team at Bank of Ireland is founded on three pillars which define how we provide our product to the business:

  • value prioritisation
  • design thinking
  • agile delivery

Every project that we take on must have value – anything worth doing has a value. Examples include cost saving, cost avoidance, new income generation or risk reduction. Our definition of value does go beyond the monetary though, we also take on projects because it is the right thing to do from a customer experience or brand perspective.

Once the idea is established, we understand the why and the relative priority; we turn to the design of the analytics solution. Data scientists tend to begin with a tools and data mindset, design thinking helps frame the problem and keep the customer at the centre of our focus. We develop an understanding of our customer needs, and create paper prototypes of data visualisations before any data exploration or programming.

Development is governed by the scrum framework. This allows us to prioritise workflow on a fortnightly sprint basis by value realised, and to take an iterative approach to product development with constant customer interaction. The development team estimates the size of each user story and has full control over how a given business outcome will be achieved.


Analytics adopted a Hub and Spoke operating model, with Group Data Office at the centre, as part of our recent organisational redesign. In addition to producing our own analytics, we coordinate collaboration across all of the other analytics functions.

To support our ambitious analytics goals, we must develop our workforce. We have a training programme for data professionals with the ultimate goal of Analytics Institute certification, but also aim to improve broader data literacy.

Bank of Ireland established a Data & Analytics Community of Practice in late 2018, hosting monthly lunch and learn events with internal speakers and evening events with invited external speakers. More than a dozen events have been held to date, each attracting a crowd of 70 or more.

The audience includes business people as well as analytics professionals. We have already seen this interest translate into career progression for individuals, and new value realisation ideas for the analytics teams. Our external speakers from companies as diverse as Microsoft, Facebook, Leinster Rugby and the Irish Times have generated discussion and visions of the Bank’s future capability.

Next Steps

In 2018, the Group Data Office 2.0 was established, which incorporates advanced analytics with data management and data infrastructure. Through a business partner team, we plan to step up the value/complexity curve and take on more advanced projects to yield an even greater return.

The data value realisation achieved in 2017-18 was vital to our business case for further investment. We are adding a Teradata data lake and Cloudera’s Data Science Workbench to our infrastructure in 2019. This will enable us to seamlessly integrate datasets and unlock the value of the data with the state-of-art machine learning and big-data solutions.

A key member of this new Bank of Ireland team is Lei Xu, nominated for Data Scientist of the Year at the 2019 DatSci awards. Lei is breaking new ground for us, taking on predictive modelling projects outside of traditional banking areas.

Author: Conor Sayles, Group Advanced Analytics Lead – Bank of Ireland