Challenging DBS: how do you measure digital innovation?

My friend Emmanuel Daniel, founder and Chairman of The Asian Banker, ran an interesting video dialogue recently, in a near hour long interview with Chng Sok Hui, chief financial officer of DBS. The whole video, audio and transcript of the interview can be found over here, but I liked this edited abstract of the core of DBS’s digital journey. It begins by discussing the digital journey of DBS.

It took us four years to change our whole legacy architecture. We “Fix the basics” from 2009 to 2014, these were the foundational years.  A lot of banks have only done the front end. What we’re saying is that we’ve done the back end, and we did it during the period 2009 to 2014, and we built resiliency. We revamped the data centre. We started insourcing, and we built engineering bench strength. We have a security operations centre. We have a monitoring centre. So, to me, this is all basic infrastructures

And then in 2009 to 2014, when we first did it, our stack was all legacies. We said we wanted to sell. “Sell” means all these are not where we want them to be. The green ones were, “Yeah, we are where we want them to be,” and blue ones were “hold”

By 2014, we had fixed the basics, so this was the chart where we said everything was done, and we were then ready to build the digital bank. So, that was very key in 2014.

In Hong Kong last year we moved to the same core banking system in Phase 2. There’s a Phase 3 this year. Core banking systems are very hard to do, but we’ve done it in Hong Kong. We successfully cut over.  And then, we did the ANZ integration (DBS acquired ANZ businesses in Taiwan and Indonesia). Taiwan needed a new cards system because we didn’t have a cards business in Taiwan. ANZ gave us half a million cards in Taiwan. They gave us half a million cards in Indonesia.

Prior to ANZ, we had VisionPLUS in Singapore and Hong Kong. We didn’t have one in Taiwan. We didn’t have one in Indonesia. And we didn’t want the same VisionPLUS that was already in Singapore and Hong Kong because it’s not cloud-native. So, again, for this, we really wanted to bring it to cloud-native as the next stage, so we rebuilt the whole VisionPLUS on cloud, working with the vendor. “You will be our first customer on cloud platform. Do you want to do it for your integration?” We said yes, and we cut over on the cloud. Cloud platforms are much, much cheaper. They’re a fraction of the traditional cost.

We successfully cut over for ANZ Taiwan and the half million cards are on the cloud platform. I don’t think there’s any other bank out there that’s got the VisionPLUS on cloud yet. And in Indonesia, we also successfully cut over one or two weeks ago, to the VisionPLUS cloud platform.

And they’re all done in-house because we also have the in-house Hyderabad capability, and we have about 2,000 people there now. Previously, we used to outsource to software vendors like IBM, NCS. Now, we insource 80 %. So, we have our own tech people who have this capability.

So, the legacy stack has been tackled. That’s why we started building the new sort of capability, and that was our next journey from 2014, when we looked at a number of these players. Google took four years (to move to cloud-based systems), Amazon took six years, Netflix took seven years, and we actually started our journey later, but we learned from all of these, and we believed we could do in three and a half years.

So, today, the bulk of our systems are already on cloud. And then, we try to make sure that APIs can be linked through a middleware layer, and that’s why we think the technology part is very key. So, what I mentioned earlier, initially, in 2009, we were 85 % outsourced. Now, we are 85% insourced, so that again is a critical capability.

So, the transformation that you see is the technology transformation that I told you about, 2009 to 2014. We overhauled all of our back-office systems. We introduced a layer so that, today, all our systems are already in the end state. Then, we moved to a cloud platform. So, there’s our first transformation, a technology transformation.

The second transformation that we are on is a customer journey-thinking transformation. All our products – and I hope you’re users of our products – whether it’s iWealth, Remit, Paylah!, are actually designed very differently from the past. The way we designed them is different - we have scrum meetings, people come together to collaborate. People walk through the journey. If you talk to people, they’ll often tell you about the four Ds. You discover. You define. You develop. You deploy. And you roll out cadence every ten weeks. So, the journey-thinking is, “I’m thinking on behalf of the customer. I’m the customer. I’m thinking how to solve the customer’s problem.” So, there is transformational thinking. Journey thinking is big.

And finally, it is culture change. We want to create a 22,000 start-up bank, and that is not a few people doing it. Its 22,000 people, all having this culture change and mindset and that is what we actually do.

 

The interview then moves into a discussion of who is digital and non-digital and how do they analyse their customers:

(This results in the bank being able) to adopt different strategies for different segments of our customers. We call it unbundling the bank. This part, which we say is 44 % of the pie, is consumer banking and SME in Singapore and Hong Kong only. So, this portion is where we know it’s stable. It’s annuity income. And here, we are being disrupted. Alipay is here. Everybody is here. That’s the market share that they want. And here, our intention is to pre-empt disruptors, so we want to make sure our market share stays.

Then, we have a slice that’s very small, and that’s consumer and SME in growth markets. So, that’s about 4%. Now, this sliver (opportunity)is just India, Indonesia, China, and Taiwan, and today is not profitable. But we believe that we have a business model that will work. We launched digibank India. We launched digibank Indonesia. We launched a partnership with Tally, so for the SME side, they can open the account and do everything digitally. We expect that this segment will break even in a couple of years, and that it will contribute at least 10 % of our pie over time. It’s 4 % today. This is a revenue pie.

Then, we have the remaining pie that is a bit noisy. It’s very hard to model because it contains Institutional bank income, and Private Bank income, which tend to be market driven. Some years are better for Capital market activities or Treasury & Markets activities. Last year, T&M did very badly. So, this is the segment where there’s bit more noise.

And then, we unpeel this 44% segment of the pie into two segments - the digital segment and the traditional segment. And the whole idea is this: that within this annuity business of the consumer banking and SME in Singapore and Hong Kong, there are actually two segments. Those who exhibit the three behaviours that I said are under the digital segment. The rest are under the traditional.

And what has been happening is that customers are increasingly adopting digital behaviour. So, from 49% in 2015, such customers accounted for 60% of revenue in 2017. The other thing that is happening is that this digital segment is actually growing faster. It’s got an income CAGR growth of 23%, whereas the traditional is flat.

We’re trying to say that you have to unpeel it and get to the granularity that you’d be able to understand the underlying drivers better. So, the digital customers’ ROE is 27%, but the ROE for the traditional segment is 19%. The cost-to-income ratio in the digital customer segment is 34 % while it is 55% for the traditional segment, giving a blended 43% for the whole segment. When you do everything blended, you can’t see it.

So we will be showing, with our annual report – the finalised 2017 data.  We show first half annualised [on Digiday], so that people can follow the progress. We have now shown the market 2015 and 2016. We show first half 2017 annualised. We’ll be showing 2017, and we’ll keep updating this process. And the whole idea is that people and the analysts can see that, yes, this digital segment is growing. The pie is growing, and as this pie grows, the 44% will grow. The overall ROE will be a better ROE. The ROE improves, and the cost-to-income ratio comes down. So, that is why we embark on this methodology, to segregate the digital segment to explain the underlying drivers of ROE and cost to income ratio.

 

After the interview, Emmanuel wrote a critique of the bank on Facebook that went as follows:

As many of us know, Singapore’s DBS Bank. Living, Breathing Asia has been doing a great public relations job on just how "digital" it is.

I had some very hard questions to ask the bank about the way in which it measured its so-called “digital” contribution to its bottom line and performance, and spent an afternoon going through the numbers with CFO @Chng Sok Hui.  I came away with the following impressions:

1. Digital measurements and benchmarks should be set by a neutral party, and not by the institution doing the drum beating. In this regard, there is no one authoritative matrix out there, and so customers, investors and employees may well subscribe to different ones.
2. The Asian Bankerhas been assessing retail banks across more than 30 countries using our new scorecard in the past two years. Our metrics is based on the tracking of the institution's own digital journey ("how digital is one institution over the others") and "what does the customer see".
   a. Our finding was that from a retail finance perspective, China Merchants Bank (CMB), the Australian banks such as Commonwealth Bank even Singapore’s OCBC Bank were far more “digital” than DBS. Our ranking scorecard is published annually.

As far as rankings go, there is no bank anywhere in the 50 or so countries we cover in Asia, Middle East and Africa who are near the new non-banks like China's Ant Financial Services Group, Tencent 腾讯 and even newcomer Korea's @Kakao Bank, and understandably so, as these are mostly "only digital" and operate under less operational, regulatory, and capital requirements.

Having said the above, DBS claims are not inaccurate or misleading. All scorecards developed by the institutions themselves are always self-serving – to tell a story to investors and internally as a guide to their own employees as to their own journey. So we have to ask them questions based on their own terms. This was something completely missed by investors at the annual general meeting of the bank recently.

1. While institution can communicate its own measurements, the fact that the institutional investor community is so mesmerized by DBS Bank. Living, Breathing Asia, as reflected in the spike in its share price since November, is amazing indeed.
2. In my conversation, the big questions that I had were as follows:
   a. Why work on the assumption that only 50% of customers are digital? 50% raises the question of double counting for some metrics. She disagreed. But I still ask this question. In my view, digital is digital and should be all-encompassing.
   b. What is the overall “plot” in its digital journey? Where do you want to take your employees, customers and shareholders?
   c. Customer value? So, what are the savings, ala Amazon.com, that are being passed back to the customers instead of to the shareholders? I told one of my journalist friends that the question to ask any bank that is boasting is “what does the customer get?”
   d. Most of the income from digital appears to be old-fashioned FX fees and NIM. There are no original digital driven fees such as in APIs and competitive pricing for trade and transactions. These are not easy to derive, but they should be the holy grail. The expectation should be set for banks by an independent observer, and not the banks to sidestep the difficult processes.
   e. What are the targets that you have set for yourself. DBS has wisely avoided stating targets.

I personally think Emmanuel is being a bit harsh on the bank, but his point is a valid one. If DBS or any other bank is truly digital, then there need to be independent metrics that can assess and show this, not internal measures and PR. What do you think?