28 Feb 2020
We hear more and more data is the future, that data will play the role in industry this century played by oil in the last. But what does that mean in practice?
Data is now central to our biggest decisions at ANZ and that means asking some key questions: What do our customers want? What do our stakeholders and shareholders need? What does the data need to do differently to enable us to provide those services and outcomes?
"In any crisis, you need to have the most current data available at the fingertips of decision makers to enable quick decision making.”
This shift needs to be made efficiently and effectively; it must address the cost of our legacy infrastructure and it must be done in a manner that gives confidence to all authorised users at any point in time.
Successful data projects conducted in banks tells us they're not technology projects. If we view data with just a technology lens, we risk missing the boat. Data projects should actually focus on outcomes for customers, then understand how processes need to adapt to allow the data to flow where it needs to. Finally, projects must understand where we need technology to support.
Data in a crisis
In any crisis, we need the latest data at the fingertips of decision makers to enable rapid decisions. The COVID-19 pandemic and recent bushfires in Australia graphically demonstrate that need. But really crises just emphasise what should be happening with data anyway.
Data needs to be available ‘end to end’ (from when a customer enters our channels, to when the issue is addressed, to the follow up required with that customer to understand how that intervention has improved their outlook).
Embedding data is about much more than a reporting function churning out reports - the world is way beyond that. Increasingly, all executives need to understand what behaviours we want to drive, what products we want to build for customers, how we enable those from a process standpoint and how we enable them from a technology standpoint.
Data needs to be consistently available to different decision makers and stakeholders (including regulators) which means we must ensure we have confidence in the source of the data and calculations made with it, how it’s assembled and stored.
With that, we can have a constructive dialogue focused on decisions we can make. Without it, management teams can sink into ‘my data vs your data’ traps that frustrate progress.
Ongoing investments in data quality drive faster decision making, faster reporting and automation. It all fits together. We want to be more proactive about knowing what levers the bank and our customers can pull in a crisis. It’s a combination of customer-driven propositions and data processes backed by technology which can help us accelerate.
Useful data and trust in the brand
My team at ANZ has two main customers: the external customer whose financial wellbeing we’re trying to help improve. And the internal decision makers who we’re trying to help make faster decisions which are transparent to the customer.
Before we can help a customer, that customer must first trust we will use that data for their benefit and they will be engaged with us in the process. That means building the muscles that pull data out from legacy ‘locked boxes’, engage with customers on propositions, and ask permission to engage on it.
“Open banking” is spurring both established banks and new challengers to do more. It will drive initiatives like comparison sites which will encourage competition. At the heart of open banking is a way of using data which is actually what banks should have been doing anyway.
At the heart of that is trust, given that open banking will happen at a customer’s request. At ANZ, to build that trust, we have a very strong data governance function. We have a set of ethical principles which we use to gauge each of our use cases.
We ask a series of questions for each use case: who wants the data? What is the data? What do they want to do with it and why? We gauge the answers against these ethical principles to make sure we're really comfortable that we're protecting customers and that the data is being used for good. These are questions we were already asking so as we accelerate we can embed these principles into future technologies.
Pandas in the snow
One of my favourite anecdotes about data came from a Google AI expert at a conference I attended.
They were explaining the power of AI by teaching a computer to recognise pandas. To do so, they threw hundreds of pictures of pandas at it. After a while, the computer started to recognise the pandas. But when the creators looked closely at the pictures of pandas being fed to the computer, they realised most of the pictures had snow as a backdrop. After the process was finished, the computer couldn’t recognising a panda unless there was snow in the background.
Now this might sound a little bit silly but what it tells you about the world of data is that you must look broadly at the quality and readiness of your data to make sure it can be thrown into an engine or program and actually produce something of use.
When it comes to open banking, say a customer wants a proposition from a fintech. They must first give permission to that fintech to request specific data from their current bank to satisfy that use case. The data will be pretty specific and the ‘share’ will be governed by specific protocols set by the Australian Competition and Consumer Commission and the government.
On receiving that request, the bank would adhere to these protocols and the fintech would be responsible for then providing the customer with the requested proposition.
This puts the customer in control. If they decide the proposition is valuable and they want more, they're in control. And if they don't want it, they're also in control. If the customer decides against the proposition, the fintech will be obliged to securely delete the data given to them under this regime.
There's clear liability set down by the regulations as to who's responsible for what. So if a bank gives data to a fintech on the customer's behalf, the fintech has liability for making sure the data is safe and secure.
Open banking really is a system which relies on everyone ensuring they can adequately protect the data. Customers will provide consent and they can take that consent away.
It raises the bar for all of us and has elevated the focus at all banks on ensuring their data quality, processes and degree of automation hit the mark.
Emma Gray is Group Executive – Data and Automation at ANZ
Since joining ANZ as the bank’s first Chief Data Officer in 2017, Emma Gray has been leading its data strategy, including how data is defined, gathered, managed and protected.
In April, Gray was promoted to the bank’s Executive Committee as Group Executive – Data and Automation. In the expanded role, she will continue to lead the transformation of the strategic use of data, as well as creating new customer insights and driving automation to ultimately improve customer experience.
The views and opinions expressed in this communication are those of the author and may not necessarily state or reflect those of ANZ.
28 Feb 2020
24 Jan 2020