When push comes to nudge: Finance gets personal

As digital banks gear up for launch, incumbent banks must go on the offensive by investing heavily in technology. Their goal: to help customers improve their financial decisions, by using AI to deliver personalised advice.

How intelligent is your bank? If it recently sent you an email or push notification that you quickly dismissed as irrelevant, it's probably not as smart as you want it to be.

It's also likely nowhere near its true potential. Today's intelligence is powered by data, and financial institutions have more personal data on their customers than almost any other organisation. Your bank knows where you live, how old you are, and how much you earn. It knows what you spend on dining out, whether you recently visited a hospital, whether you own a home and how much you still owe on it. Your bank can predict other details about you, such as whether you're a parent, how much investment risk you can take on, and how prepared you are for retirement.

In fact, the quality of customer data that banks possess far surpasses that of most other organisations, since their data points are "deterministic" - based on verified personal data and actual transactions. In contrast, many social media companies observe customers' digital behaviour and infer their hobbies and interests, which is known as "probabilistic" data.

Yet fewer than 1 in 10 customers receive personalised messages from their bank, according to a survey of 5,000 American customers reported recently in banking publication The Financial Brand. Why don't banks seem as intelligent as Google or Facebook in giving personalised and useful recommendations? The simplest answer is that most banks don't have a centralised view of its customers. The credit cards team knows your spending patterns, the mortgages team has your home ownership data, and the deposits team can see your monthly salary - but each of these data sets likely sits in a different system. That might be why you could receive an email promotion about a mortgage offer even though you don't save enough each month to buy a house.

But even if a bank centralises all its customer data, it may not have a full view of your entire financial situation, because most people use more than one bank. In Singapore, each consumer has an average of two to three bank accounts, and about five credit cards. This is where a platform like SGFinDex, developed by the Monetary Authority of Singapore (MAS) and the banks in Singapore, comes in. Customers can use this to safely sync their financial data and get a full overview through a single platform. According to MAS figures, more than 150,000 users signed up for SGFinDex in the first year after the platform was launched in November 2020, linking some 290,000 bank accounts in total. But this is just a drop in the pond, given that many banks in Singapore have more than 2 million customers using some form of financial planning tools.

Those worried about data privacy should be comforted that all the necessary data safeguards have been put in through regulation. The financial industry is much more regulated than other consumer sectors, and there are strict restrictions to ensure that banks are using customer data for the right reasons.

Moving up the "hypersonalisation curve" with the help of intelligent nudges 

While regulations are a big reason that banks have been slow to make the most of their data, they are now starting to get wise to their intelligence gaps. This is in no small part jolted by the emergence and success of fintechs, who have it in their DNA to use data to create delightful customer experiences. In the UK, digital banks Monzo and Starling are outperforming traditional High Street banks in customer satisfaction and their mobile apps usage, thanks to their AI-powered tools that can auto-budget, split bills, and even allow customers to discover relevant deals in their location based on financial transactions.

Against this competitive landscape, banks are quickly moving up the intelligence curve. Since 2019, the Commonwealth Bank of Australia has been sending alerts, powered by artificial intelligence and data analytics, to customers to help them avoid late credit card payments and suggest how best to use their annual tax refunds to achieve their financial goals.

More than half of banking customers want more help from their bank to manage their money, according to The Financial Brand report. This includes recommendations for cheaper credit cards or loans, and reminders of subscriptions they might have forgotten to cancel. To get there, banks need to sharpen their focus on personalised and intelligent communications.

This means three things: ensuring the best use of data across the bank; giving customers control over their data and notification preferences; and making sure nudges reach the right people and show them useful insights and relevant products.

At DBS, we've created an internal data mart with 15,000 customer data points that stitch together a customer's profile, including the pages he/she visits, the content he/she is interested in, the frequency of his/her visits, the amounts he/she spends on different categories, etc. We convert the data into personalised, actionable insights - "nudges" - to guide the customer in his/her transaction or investment decision.

Where we have data gaps, we've been supplementing our information with new data sources such as SGFindex, and location data, where customers give us permission to identify where they are so we can send them location-based offers and insights. Unifying this data has allowed us to deploy more than 100 artificial intelligence and machine learning models that pinpoint which messages are most relevant for which customers. This in turn enabled us to send more than 30 million nudges a month to 3.5 million customers in Singapore.

Nudging toward better financial health: A force for good?

By analysing your past spending patterns, banks can predict upcoming payments and alert you to make sure you have enough in your accounts to avoid being charged fees for the shortfall. They can even detect if your free trial period for a subscription has ended, and notify you that you're starting to be charged for it, in case you were too busy to notice. If banks can tell that you are a busy traveller, the models might recommend that you replace your existing credit card with one to help you maximise miles. Another model would identify the right time to send this message to you, by looking at the timing of your transactions. For a first-time investor, you can calculate your risk appetite and get a list of relevant investment products. After you've made an investment, you would then be able to trigger automated updates about your portfolio or other market opportunities based on the latest real-time prices.

The upshot? Customers who receive such nudges are likely to make better financial decisions - they tend to invest more and be better covered by insurance than those who don't.

At DBS, we observed a four-fold increase in the number of customers who completed their investments, after the launch of our AI-powered digital investment advisory feature that can determine individualised investment risk profiles and make specific recommendations. We've also launched a platform that nudges relationship managers to chat with clients, which resulted in 16 per cent growth in client engagement. And consumers have control over how their data is used. They can choose which channel they prefer to receive each type of message sent - perhaps they want transaction alerts through SMSes, but card promotions to be sent via in-app notifications, and investment product offers through email.

We believe this is the future of banking: to be both customer-obsessed and data-smart. Combining financial expertise and the data at their disposal, banks can give customers hyperpersonalised advice and politely, but effectively, nudge them into better financial standing.

Sanjoy Sen is group head, Consumer Bank, DBS Bank. Evy Theunis is head, customer segments and customer science, DBS Bank


BT is now on Telegram!

For daily updates on weekdays and specially selected content for the weekend. Subscribe to t.me/BizTimes