DBS Bank Pumps Up the Volume on its Technology

Singapore’s DBS Bank is the largest bank in Southeast Asia, with $401 billion (Singapore) in assets. It is the dominant retail bank in Singapore, and also has a growing presence in China and South Asia.

The bank is using technologies like mobile, social media and analytics to remake its relationship with customers and its operations.

David Gledhill, group executive and head of technology and operations, joined the bank in 2008, and has led the bank’s reworking of technology. He spoke with MIT Sloan Management Review contributing editor Michael Fitzgerald about adapting new technologies and developing the skills needed to use them effectively.

When you arrived at the bank in 2008, observers commented that DBS’s systems had not kept pace with the bank’s rate of expansion. Now it has a reputation for technology excellence. What have you and the management team done to change its reputation?

The most important thing was to get the culture shift right. That’s what we worked on for the first two or three years. Obviously that’s a journey. The cultural messages we gave were some of the most relevant ones to get the thinking to shift. It was a little bit “mother and apple pie,” but we focused on three simple things: we’re going to build world-class technology, we’re going to focus on getting the business moving and driving innovation through technology.

Just the attitude shift that we’re going to build world-class technology was a major component.

What were the most important new technologies you wanted to innovate with?

We’ve done a lot in the mobile space. We’ve really pushed the agenda on analytics.

For example, ATMs are a big, big challenge for us here in Singapore. We have one of the highest volume uses of ATMs anywhere in the world. Just to give you a sense, the average ATM in the U.S. dispenses about 2,000 cycles a month. Our average ATMs do 20,000 cycles a month.

We said, there has to be a better way to schedule the filling of these ATMs to reduce cash out, improve customer services, etc. So, we partnered with SAS. It took about six months of modeling work. And then to train the model probably took another year. Now, the whole island is running on the thing. We already had embedded sensors in the ATMs, which told us about cash flow and dispense velocity, but advanced analytics and big data helped us come up with predictive models that got us to a completely different level of performance. Our cash out dropped by 95%. The number of trips we had to make dropped by 10% or 15%, and customer satisfaction went through the roof.

The point of that whole program wasn’t just about how you fill ATMs, but it was exposing to the company the power of analytics applied to a seemingly inconsequential problem.

Your flagship bank branch is a tech and design wonder. Will features like the MicroTiles be more than gimmicks?

Lots of banks build branches with lots of screens on the wall and make it high tech almost for the sake of high tech.

Our goal was to create a signature customer experience. For example, when you approach our branches, the first thing that happens is, you get met by a person. Why? Because from an Asian service perspective, we believe it’s respectful to have somebody greet you. They don’t stand behind a counter, they stand in front of the counter, and as you’re walking towards the branch, they walk towards you. That’s a signature experience.

We have lots of technology that a customer can use to interact with us. We have mobile; we have Internet; we have ATMs. So when they come to the branch, they probably want to speak to somebody.

Behind the scenes is very high tech. For example, when they go to one of our service pods, we said we’re going to automate 95% of what they want to do, so the teller will be able to [take care of them] without leaving their seat. In the past, if a customer came and wanted to open an account, I’ve got to photocopy this form, I had to walk and go to the photocopy machines and come back. You want to issue a new card, you have to go to the card-issuing machine and come back. I want to deposit some money, I have to go to the teller counter and come back. It was a very choppy experience.

So what we’ve built into these pods is enough technology so 95% of the interaction can be done without the person leaving their seat. That’s where the technology comes in, but we actually hide it away. It’s kind of hidden behind the teller, and it’s something the customer doesn’t see.

One of the issues with analytics for many companies is that they don’t have the talent in-house to do clever things with it. How did you address the skills gap?

We’re doing a number of things. One is that we are building up our own analytics team. The other thing is that we are again going to best-in-class vendor partners to have them help with that resourcing skill set. We’re using some vendors in India, we’ve gone to IBM because they have some advanced analytics capabilities, and using a number of different partners. We’ll bake off different vendors against each other and look at model performance, to see how much uplift — how much of a “hit” did we get — to score how these different vendors are doing.

That’s actually quite effective, because as analytics is developing so quickly, no one firm has all the right answers. To assume we’d get all the right answers just by hiring internally I think is naïve, and to believe that all bets into one vendor partner is the right idea, again, is naïve.

How did you decide how to prioritize analytics versus mobile, versus social, versus other technologies?

I think analytics, we always believed fundamentally, was a huge opportunity. For mobile, it was more of a hygiene strategy. Back in 2008, smartphone adoption had not really kicked off at the level it is now. We were thinking we should provide services on mobile because everybody else is. But we were not really sure how to monetize this. It was probably 2009-ish before we took another look and said, you know what, we need to place some major bets. Now we’ve got 10 or 12 different mobile apps. Still, if I look at mobile, really all we’ve done is digitize transactions. We haven’t created a digital bank in the mobile space.

Describe how the uGoIGo campaign came about. Where does that fit into your social media strategy?

UgoIgo came from our consumer team in Hong Kong. They just had this cool idea that we could use Facebook and other platforms to create this collaboration between investors to drive up the interest rate. It was aimed at [encouraging] wealthy customers to place large deposits with us. The more people that signed up, the more interest rate they got on the deposit. It got all these wealthy people to tell their buddies, hey, jump on in.

It was very, very successful. It exceeded the take-up rate we anticipated. It showed we could use a certain level of social collaboration around some of our products. It’s a difficult thing to scale, but what it did tell us is that there are opportunities to leverage social media-type sites to cross-sell and use likes and other people’s social connections to get our brand awareness up.

Frankly, in all of these things, even in analytics, we’ve come a long, long way, but I still think we’re only scratching the surface of what the possible opportunities are.

The other thing is, we’ve made some bets. For example, we figured back in 2008, 2009, that we should start looking at some real-time capability, event-driven capability. At the time, we didn’t have a business case, but just saw it as a technology we needed to experiment with. Now it is driving a lot of our ATM fraud analytics. It’s driving a lot of our notification services to customers. We’re building marketing campaigns from the top of it.

What does “real time” mean?

Real-time event-driven marketing or other pushes. For example, we load up a campaign to say, if these people use this ATM in this certain area, we can push them an offer for a restaurant that wants to fill tables. If we’re using it on fraud, if these different types of cross channel correlated events happen then do the following.

People talk about cross-channel awareness; some of that, we now use from real-time event engines. For example, if a customer gets a card swallowed in an ATM, that’s an event, not a transaction. So how should the other channels respond to that? Should the RM [relationship manager] be aware? Should the event of this customer, who happens to be a very important wealth customer, drive and alert the RM to say, “hey, your customer’s in trouble”? Or if that customer then calls the call center, should the event awareness be able to tell the call-center agent based on the mobile [phone], “uh oh, this customer just called, his credit card transaction failed, and he just lost his ATM card, therefore do the following for him”?

I know the bank is using voice analytics in the call center. Is that related to this real-time events idea?

A little bit different, a little bit separate. This is a sort of embedded sensoring thing, right? If you think about what we did with the ATMs, we took what was already an embedded sensor, which is the amount of cash and the velocity of withdrawals, [and] based analytics around the output of those thousands of sensors delivering information the whole time. For voice analytics, what we wanted to do is to say, actually, if you think about the power of embedding sensors into conversations, what could you do with that?

Tell me, what could you do with that?

Voice analytics is embedding a sensor into a conversation stream. We can sort of build up models, which really start to get to the core of what’s going on in the call center. How many calls are because we had a certain issue? Typically, you get volatile call volumes, and often it’s driven by campaigns, but we never really quite understand why. You can look at certain patterns of words that get spoken on calls. If you have problems with the way that something’s worded or something is designed, or whatever else, it features back into the question of, can we design it differently, can we change the way a button operates, can we change the way a promotion is done — because what we’re seeing is the shape of voice, the shape of words coming into the call center, which we think is a sign that something is badly done.

The other thing we can use it for is to help optimize our call center agents. We know what part of calls where agents are taking longer than they should, and what kind of things are going on in those conversations, which we need to use to retrain the agents. It’s about embedding sensors in conversations, picking up figures from those conversations and using that to drive optimization and change in the organization.

DBS’s CEO, Piyush Gupta, is known for being a champion of social media usage. He’s even got his own internal portal, the Ask Piyush Portal. What sorts of tools and culture development does the bank have to do to keep him happy?

The Ask Piyush Portal we’ve turned around to say “Tell Piyush,” — we want more people to tell him about things we can improve. Rather than keeping it open the whole time, we open it for periods of time tied into a quarterly town hall he runs. If it’s open the whole time, it tends to get stale. Again, I think we’ve got a long way to go on our internal social mobility. We have a whole team of folks looking at what we describe as the future of work, how we enable mobility, how we enable team participation, how we enable fun things like vote-a-thons, whatever it is. We are not what I would say is world-class right now. We’re on a journey, but we’ve had some good first results.

Externally, banking social media is kind of different than the likes of a Nike or some of these trendy brands. People don’t tend to like banks as much as they would like clothing or record labels or pop groups. You don’t get as much of an active following. What we’re thinking about with social media is to the extent that we can be the anchor point of communities. We have a lot of knowledge and a lot of advice. Some people have tried to create communities around trading activities, around saving, around small to medium type enterprise. Those are the sort of things where we think the bigger play is for us.

DBS has Innovation and Customer Experience councils. Can you talk a bit about their role in adopting new technologies?

So, two very broad areas, but both are fast merging into one. Customer Experience — we go right back to when Piyush came on board. He laid out five pillars he felt could help differentiate us as an organization. He calls these the Asian pillars, [since] we want to do it in an Asian way. One was service, and one was innovation. On the service pillar, we can differentiate by great customer service. The question is how you do that. You can’t just go train people on Asian service and how to say hello. There has to be something driving the whole customer experience agenda. That’s what the council was set up to do. Somebody from my team plus somebody from consumer business helped facilitate the Council; it’s chaired by the CEO.

We figured we needed to come at it from two different ways: “heartware” and hardware. The heartware is how do you get the culture shift going on in the organization. Through a long set of interventions over a 6-month period, we came up with this idea of service as being respectful, easy to deal with and dependable. This concept became the whole driver of what customer service meant — how you become more respectful, how do we become easy to deal with and dependable.

Then we set ourselves a set of hard targets — really important, customer-visible things we want to influence. So, turnaround at the call center, we want to be best in class at how quickly we can deliver a new credit card to a customer, we want to have the best ATM availability in the world. Big sticks; hard goals. The innovation came from the customer experience council defining a set of very hard KPIs [key performance indicators], which you could only achieve through innovation.

There are three or four big areas we want to drive innovation in. Let’s get the business leads for each of those, and the innovation council is where they come together to talk about progress, to talk about milestones, talk about where we want to push. We really wanted to push the point that everybody owns innovation. We have a very small innovation team, but they are agitators.

So you kind of centralized the technology innovation by bringing people from around the firm into one place to meet and talk?

They still own the innovation, so it’s decentralized, but the innovation council brings it together to make sure they are broadly in alignment.

How important is it to have a CEO who gets it and is involved?

Massive. The CEO chairs the innovation council. I don’t think we’d get 10% of the stuff done that we want to get done without the CEO living it, breathing it and believing it. With the CEO, that aligns all the business guys, and it aligns the budgets, and it aligns the Board, and it aligns the investment dollars and everything else you need to drive something like this.

I love that you made this point about technology being an enabler, but not the point. What are the big challenges you would want others to understand about what it takes to make technology be transformative?

The transformation that’s truly taking place is the move from transaction processing to information and knowledge. That requires a completely different thinking in the bank and a different tooling and a different set of skills. The old is important, but the new is where the value is. All the stuff I talk about around ATM analytics, voice analytics, pushing on business analytics, creating these social media apps, is all about information and knowledge transformation.

What do you mean by information and knowledge?

So, customer uses a credit card in a store. The transaction is that I accepted the payment, I paid the merchant, I gave the customer a statement and sent them a balance [report] at end of the month. The information is: where were they, what were they doing, what did they buy, how did what they buy correlate to their investment preferences, how can I use the knowledge and information in that transaction to drive something.

Same thing with the ATM. The transaction is, the customer takes cash out of the ATM. I book the transaction, I take the money out of the customer’s account and I’m done. On an information basis, it’s all again about, where was the customer, what where they doing, what can I upsell them, what did that transaction mean in terms of the velocity of cash moving from the ATM, what kind of analytics can I drive to optimize how I fill the ATM, etc. There’s so much information and knowledge in the transaction that we don’t extract today.