Skip to contentSkip to navigation

Transforming a traditional bank into an agile market leader

In 2014, Piyush Gupta, CEO of DBS Bank of Singapore, asked his staff to think like the employees of a fintech startup and build the digital capabilities they would need to succeed.

This interview is part of the Inside the Mind of the CEO series, which explores a wide range of critical decisions faced by chief executives around the world.

While big banks in the U.S. and Europe have been grappling for almost a decade with how best to go digital, a modest-sized bank in Asia, DBS Bank Ltd., with assets of SGD 541.5 billion (US$394.7 billion) as of September 2018, has come from nowhere with a solution: Don’t try to figure out how to acquire digital capabilities. Instead, think like a fintech company and transform your business completely. The strategy is working: The compound annual growth of the bank’s net profit from 2014, when the transformation process started, to 2017 was 2.7 percent, and the first nine months of 2018 showed net profits up 36 percent compared with the same period the previous year.

The architect of this strategy, Piyush Gupta, is an Indian-born banker who spent many years at Citibank in Asia before becoming chief executive of Singapore-based DBS in 2009. At the time, the bank, which was founded in 1968 by the government to help fund the young city-state’s industrial development (and which was then known as the Development Bank of Singapore), was slow-moving and bureaucratic, and was running on outdated technology. No central data was available to show which of the bank’s branches were profitable and which were not.

Gupta, a keen birdwatcher, spent some time observing not only what peers were doing but also how Alibaba and Tencent, China’s two largest e-commerce companies, were moving rapidly into banking by using digital payments technology. He also observed that banks tend to face more onerous regulatory burdens than do tech companies, making growth through acquisition of other banks undesirable. He decided that DBS needed to have a fintech mentality, with all the agility and inventiveness that would imply. “Banks are yesterday’s story,” Gupta recently told the Financial Times.

Gupta’s transformation program is regarded by many in the industry as the most far-reaching of any bank’s. Internally, DBS has inculcated a startup mind-set among its 26,000 employees to help them develop and deliver simple and effortless services, enabling customers to — in the words of its latest brand campaign — “live more, bank less.” It uses hackathons to spur product ideas and is also the first bank in the world to develop a methodology measuring digital value creation, which shows the impact of digitization on earnings. Gupta has also looked externally for technology. In 2016 DBS took a stake in Kasisto, a New York–based banking chatbot and AI startup.

DBS operates mainly in what it calls the three key Asian “axes of growth”: Greater China, South Asia — where the bank launched India’s first online-only bank, called Digibank — and Southeast Asia, where DBS is the largest bank by assets.

Strategy+business sat down with Gupta in October 2018 at DBS’s headquarters overlooking Singapore’s busy harbor. The conversation took place a few weeks before Singapore’s annual Fintech Festival, of which DBS is a sponsor; although it’s only in its third year, it is one of the world’s largest such gatherings. Gupta believes that DBS on its 50th anniversary is at a crossroads — and so is banking, thanks to the digital revolution.

S+B: Where and how did you get your “eureka” moment telling you that DBS had to change the way it was offering banking services — and transform its business model?
When we started the journey in 2014, the conventional wisdom was, “It’s hard for legacy companies to change. And the guys in the startup in the garage always win, because they are a different breed.” Three things happened that year. One, I started really watching Alibaba. It was quite clear to me that these guys were going to redefine banking in substantially different ways. And it wasn’t going to be a marginal exercise. It was going to be the core of the company. So we were going to have to attack the core.

The second insight came from my experience at Citi. In the late 1990s, [former Citi chief executive] John Reed had set up something called “e-Citi.” He hired a guy called Ed Horowitz from Viacom to run it. Three or four years and $3 billion or $4 billion later, they shut it down. My insight from that was, if you don’t “mainstream” this, then the core of the company is always trying to shoot it down. Doing things on the side is not such a smart idea.

Get the strategy+business newsletter delivered to your inbox



And my third insight was actually personal. I’d gone back to Delhi in late 2013 and saw that my Dad, who was in his mid-80s, had completely changed the way he lived. He was paying his taxes online. He was making bill payments online. He was ordering stuff for my mother online. And I remember thinking, if a guy in his mid-80s can change in his personal life, what makes us think that people in their 30s, 40s, and 50s cannot change in a working environment? The human brain is the same, and all of us are changing: We are ordering Uber, we are Skyping. If we can do it, why can’t employees do it? The problem is not with individuals — the problem is with the environment.

And in order to fix the environment, you basically boil it down to two big things. You’ve got to create an environment where you teach people and you give people the ability to get their hands dirty, learning by doing. Experimenting.

And the second big thing is the fear of risk. In the professional environment, risk is extraordinarily high. At home, worst case is we get frustrated because some app didn’t work. At the bank, people could lose their jobs, they could lose their bonus. So if you figure out a way to learn by doing and make it OK to fail, then it’s OK to take risks. So how do you get this culture change and become like a startup? You have a central team that creates a culture of experimentation, which gives people an opportunity to work with other people [in a risk-free environment]. I was really surprised that in the first couple of years [of our change in mind-set] we started getting really huge traction. And we made it happen in every part of the company, including human resources, marketing and communications, everywhere.

S+B: And with these departmental initiatives, did you bring in people from the outside? How did it work?
I found that once you give people permission and you give them some training, you unleash this tremendous energy to do things. Most of the groups went and hired a couple of [young people]; the comms people hired one or two people who knew data analytics. HR hired data scientists. It wasn’t orchestrated. So people used people they had and we gave people the freedom to experiment and try.

S+B: That’s interesting. Dan Cable, a professor of organizational behavior at London Business School, has developed a theory that employees need to be given just that — the freedom to experiment and try things. Unleashing the creativity of employees, essentially.
Exactly. I found this in practice. So, when I saw this happening at DBS, it really blew me away how much traction this thing had gotten in the system. My first two years [here] I used to run projects top-down. The last couple of years I’ve been blown away with what people are doing. Everybody is driving this transformation and change in every part of the company.

S+B: You started to talk about the digital transformation at DBS four years ago. What has happened since then?
It wasn’t all chronological. Some of it we stumbled into. We basically focused on trying to build five capabilities. One, we figured we had to learn how to acquire customers differently. Because when Amazon and Facebook acquire customers, they have neither branches nor bricks and mortar nor feet on the street. So they’ve changed the acquisition paradigm.

Two, we had to rethink our transaction model. Our whole transacting thinking was built on transaction in the morning, output in the evening. Can you imagine putting something on Google in the morning and getting a response in the evening? So you’ve got to change your “transaction thinking” to instant. Instant fulfillment has got to be your mantra, which means you’ve got to kill paper.

“I found that once you give people permission and you give them some training, you unleash this tremendous energy to do things.”

The third capability is moving from what I call cross-sell to cross-buy. The idea of cross-sell has been that once you get the customer — which is an expensive proposition — then you see how much more you can keep sending to the same customer. But what the Amazons of the world do is actually create the conditions and use data to give the customer so many choices. So how do you get a customer engagement model, using data, in which the customer chooses to do more and more with you?

The fourth capability was, how do you move from pipelines to platforms effectively? We as banks have always been direct to customer, and we never worked through partnerships. Whereas all the big tech companies work through partnerships. They changed the whole mind-set. The fifth capability was data. We just had to really figure out how we would become a data-first company.

To build these five capabilities, we put in place an overarching program to change our technology architecture. David [Gledhill], my CIO, came up with GANDALF: Google, Amazon, Netflix, Alphabet, LinkedIn, Facebook. The D in that is DBS. And the idea there was not to be vain — it was to change our competitive frame. We’ve got to think like a tech company. So we carry a PowerPoint slide deck around that says: “What would Jeff [Bezos] do?” because we’ve got to be a tech company.

We figured out that most of the companies that we think of today as tech companies started with technology similar to ours. But they all transformed in the 2000s. And we said, “If they can do it, we can do it too.” Like Amazon and Facebook, we have gone into cloud-enabled services, open source applications, [and] commodity hardware, [and have] shrunk the cost of our technology down and in-sourced a lot. We were 85 percent outsourced and now we’re 85 percent in-sourced. We built a tech center in Hyderabad and hired 100 people: design people, data science people. But perhaps most important, we retrained our people.

I saw an interview with the Netflix CEO Reed Hastings, where he was asked, “Where do you get these techies from?” And he said, “I hired the people you fired.” And this can be done. If somebody gives employees the capability to reinvent themselves, gives them some training, they will do it. That’s what we found. We hired some new people but we reskilled our old people and built this whole cadre to re-architect our internal technology.

S+B: Did you follow a particular road map or theory?
We focused on Harvard Business School professor Clay Christensen’s idea of “jobs to be done,” or basically “customer journey thinking.” And the basic idea is that it’s not just technology. Uber was not about technology. Uber was about reimagining what the customer really wanted. So we came up with our own journey methodology that’s called the four Ds: discover, define, develop, and deploy. We trained the whole company with the four Ds. We gave people targets. We’ve done 500 to 600 different journeys across the bank. The power of that is that it’s in my human resources, it’s in my compliance, it’s in my audit, it’s in my strategic marketing and comms — everywhere. So everybody embraces this journey of how to really be customer-centric.

This is all about culture change. How do you become a startup? How do you become customer-obsessed, data-driven, agile, a constantly learning organization? We created a culture of experimentation. We had a central team whose job was to create a set of programs to help change the culture of the company. We went through incubator to accelerator, and started the first hackathons in Asia. We had a KPI for a thousand experiments in the company; we created a program for startups to work with us; we created big centers where we have people who work with startups the whole time. All of it was focused on changing the attitudes and mind-set of our people.

S+B: What forecasting data are you using that your predecessors did not have access to?
We’re still somewhat in the early stages of how we leverage data well. We pull data from the rating agencies and pump it in to see whether our own data is consistent with what the external world has. But some of the data is unstructured. So what I’d really like to be able to do is take all the data from the Financial Times, and take all the data from Bloomberg, and sift it and mine it to ask, “What does it tell me about what’s going to happen to this industry or this customer?” The process is not there yet. We are experimenting with more natural language processing and fuzzy logic to read some of this stuff and make sense of it. But it’s still very primitive.

S+B: Nevertheless, has your company become measurably better at forecasting compared with a decade ago?
I think we have. We have machine learning and AI tools, which take office data and forecast what’s likely to happen to savings in Singapore in the next two years. We take leading indicators — surpluses, macroeconomic indicators — and make those forecasts. Based on that, we can forecast what’s likely to be the shape of our balance sheet. We couldn’t do this even a year ago. These kinds of forecasting tools are new for us.

S+G: Where do you get the expertise to do this?
Almost all internally. We’ve developed an in-house capability [with] data scientists. Across the board, we have trained our people to think “data first” and think about how to be a data-driven company. We’ve trained hundreds of people on how to use data, how to leverage data.

S+B: What is DBS using AI for?
For robotics process automation, which means all the stuff that you have [previously done] manually that you now can…build intelligent rules around. The system can then automatically use those rules to do whatever needs to be done. We’re also using AI to figure out the products we should be offering to somebody and at what point in time. A couple of months ago we launched a front-end chatbot with AI to do first interviews for all those who apply to become wealth planning managers at DBS. We get about 80,000 job applications every year. We also use AI to predict who is going to quit the company. We can tell with 84 percent accuracy whether anybody is going to quit in our sales areas in the next nine months.

S+B: Really? How does your system predict that?
It uses data science. We can track basically everything that signals the employee’s engagement or disengagement: what time do they come to work, how many times do they access email. We send a list to the managers and say, “We think these are the people who are likely to quit in the next year,” and the manager has the choice of whether to engage with the person ahead of time.

S+B: Isn’t that rather Orwellian?
We have an appropriateness committee to assess the appropriate use of data. This is going to be one of the biggest challenges of our time. It’s challenging social mores, and attitudes are changing. What was not appropriate three years ago now everybody thinks is OK, right? And sometimes it seems to be OK and then everybody says, “Hey, maybe you shouldn’t be doing this.”

S+B: Do you have an example?
Let’s look at it through the lens of the interview process I mentioned earlier. We use all this data to decide who to interview, and maybe people find that a bit creepy. But hang on a second, isn’t it my job, our job, to find the right people for the company? In the past, I did this by looking at your CV and interviewing you. And whatever data I gleaned from the written application and from interviewing you was OK, right? But it’s my job to make sure you’re a good fit for the company. If I can find more data to make sure you are a good fit, what makes it not OK suddenly?

Now let’s take credit. The whole job of banking is to assess whether if I give you a loan you will repay it. So I use whatever information I have to ask, “Are you a creditworthy customer or not?” Now, if I can get more information to refine the process and be doubly sure you are a creditworthy customer, does that make it bad? That’s still the same job, and it’s trying to achieve the same outcome.

The problem is that it can lead to financial exclusion. And the most vivid example to my mind is insurance. The whole basis of life insurance, for example, is socialization of risk. Nobody knows who is going to get cancer, so we all effectively pool money in by paying premiums, and then if somebody dies, the money goes to him or her. But once I know, via data, whether somebody is likely or not likely to get cancer, what happens? You can’t socialize this anymore. And if you can’t socialize risk anymore because you know too much, then — kaput — there goes the insurance industry.

S+B: This gets us into determining the appropriate use of data in the provision of financial services.
Yes, this is why I say this is one of the biggest challenges of our time. At some stage, societies will evolve and say, “We don’t want to know, even though it leads to better outcomes in some cases.” I’m not smart enough to know where this is going. All I know is that it’s going to be a big issue. And we’ve got to be very thoughtful about deciding what is the appropriate use of data.

S+B: Getting back to the economy: What do you think are the global prospects in the coming years?
I’m still somewhat sanguine. Global growth will probably continue for another year and a half before the bull sentiment runs out. As long as the U.S. is strong, it can power economic growth. China will continue to slow, but at current rates of slowdown, that’s not going to affect global growth. Southeast Asia is already demonstrating that it has the flexibility to retrain its export engines, having switched from facing the West to facing China in 2009. Things will be somewhat slower, but I don’t see a massive recession coming any time in the next 18 months.

What could throw that off, obviously, is “event risk.” The most obvious event risk involves China. I don’t think trade is really the principal issue. I think leverage and idiosyncratic counterparty events are greater potential risks over the next 12 to 18 months. A lot of bonds have been issued out of China, and given the environment and the uncertainty, I think people will find refinancing not that easy.

While trade itself is less of an issue, I do think the psychology around trade is a risk. There’s a general environment of skittishness, and there could be [repercussions] from that in the general economy. The Fed’s already signaled continuing higher rates; dollars keep moving out of the system, and there is some dollar liquidity tightness. If this “taper tantrum on steroids” continues for a couple of quarters, that could dislocate the markets. 

The third risk is Europe. Brexit is just symptomatic of broader European issues. There are event risks out there, but the economy will stay intact into [2019].

Author profile:

  • Jeremy Grant, a former Financial Times correspondent, is an editor with PwC based in London and international editor of strategy+business.
Get s+b's award-winning newsletter delivered to your inbox. Sign up No, thanks
Illustration of flying birds delivering information
Get the newsletter

Sign up now to get our top insights on business strategy and management trends, delivered straight to your inbox twice a week.