The path to profit for digital banks

OF THE approximately 200 digital banks of all forms, both startups and within incumbent banks, I found only 4 that were profitable in 2019: Kakao Bank, WeBank, MYbank and Tinkoff Bank. My research showed that the first 3 had a huge benefit in the form of low cost of acquisition and activation due to their association with their parent companies, which had lots of potential customers and data on which the banks could leverage.

When I asked Kakao Bank during a visit in February 2020 what factors were most responsible for its success, it named 2: its unique shareholding; and its timing, citing that when it launched in 2016, there was no truly credible Korean mobile banking app. In its first month of operation, it signed up a million customers. As at March 2020, it boasted 12 million users (23 per cent of the population of South Korea), of whom 10 million are active.

Kakao Bank was listed, and debuted with a valuation of 25.5 trillion won (S$29.7 billion) at the opening bell on Aug 6 this year. The country's largest banking group, KB Financial, is valued around 22 trillion won.

From its inception, Kakao leveraged the immense popularity and trust factor of its parent KakaoTalk, the mobile messaging app used by over 90 per cent of Koreans, and thus they were able to attract a wide spectrum of customers. Featuring Kakao Friends characters on the bank's app and debit cards helped immensely in attracting customers in the early stage. While Kakao has stated that it targets 30 to 50-year-olds as they are more profitable, Maeil Business News Korea reported that as at June 2019, "32.1 per cent of Kakao Bank customers were in their 20s, 31.2 per cent in their 30s, and 21 per cent in their 40s". It added: "When looking at Korea's entire population, this means 46.4 per cent of all 20s and 42.8 per cent of all 30s are using Kakao Bank."

In my view, most digital banks will not have Kakao Bank's good fortune. They will not become profitable so fast as the incumbent banks in their countries have digital banking that perform reasonably well, and there is no platform like KakaoTalk to help rapidly accelerate customer capture.

At the margin

GET BT IN YOUR INBOX DAILY

Start and end each day with the latest news stories and analyses delivered straight to your inbox.

Sign UpVIEW ALL

Instead, for most digital banks, the key to profitability starts with understanding marginal and fixed cost accounting. Businesses lose money if their marginal profitability is negative.

The easiest way to think about this is to do the maths on the next customer you acquire. If that customer's current annual revenue is less than his cost to acquire and his annual servicing costs, then the marginal profitability of the business is negative. This means that as the business expands and acquires more customers, it incurs more and more losses. In a traditional bank, new product profitability is mixed with the profitability of the bank's existing cash cows, so one rarely has to think about marginal profitability the way a start-up has to go about it.

Definitely, there will be a temptation to think about large revenue generation potential early. This would be a mistake. All routes to a successful digital bank will require obtaining a positive marginal contribution first. This also means keeping your acquisition costs and servicing costs low and containing your annual fixed costs within the ability of your customer base to recoup.

Even WeBank and Kakao Bank would have passed through this stage first - only in their case, they managed to reach this point much sooner. In the case of Kakao Bank, it did so within 18 months of launch.

If there are patient investors who are willing to trade short-term profit of a smaller business for long-term market capture to build a bigger business, then the company's appetite to sustain a negative marginal contribution could allow it to take these losses in its stride.

This patience would also depend on the possibility and probability of share capture. If the population is large and there are a lot of unbanked customers, and scale would lower the cost to acquire and serve and spread the annual fixed cost to be incurred over a bigger base of customers, then there may be light at the end of the tunnel for sustaining losses in the hope of building a bigger business.

More often, the patience is not there, and in that case the cost incurred needs to be commensurate with the revenue generation ability.

There is a distinct connection between the path to profit and the revenue generating potential of the chosen segment. If the team has decided to develop the younger, more digitally savvy segment, then there is a need to contend with their lower revenue generating potential. Therefore, from the outset, the focus will be on costs - in particular the cost of acquiring an active customer and the annual cost of servicing this customer.

The first decision is to select how long you are prepared to allow the digital bank to sustain negative marginal profitability. The higher the focus on share capture, presumably due to a large pool of unbanked customers or an anticipated upswing of revenue in later years due to a focus on young professionals, the bigger the appetite must be to sustain a negative contribution. You cannot have your cake and eat it too.

There must be a focus on active customers. In my experience, it is better to have a smaller base of highly active customers than a large base with a very small proportion of active customers. The numbers speak for themselves: If you have 2 million customers with an active rate of only 10 per cent, then you really have only 200,000 active customers, and it is the same as having 500,000 customers at a 40 per cent active rate.

But from a cost standpoint, things may be very different. If 80 per cent of your 2 million customers are eligible for the $25 reward you dangled to attract them to apply, then you might have spent $40 million to acquire 200,000 customers at $200 per active customer. But in the case of 500,000 total customers at the same 80 per cent eligibility, the spend is only $10 million, or $50 per customer.

If the segment's immediate revenue generating potential is small, and if there isn't a large appetite to sustain many years of losses for the eventual prize at the end of the rainbow (a situation that will be faced by almost all incumbent banks), then the key is not to pay a lot for acquiring active customers.

Net promoter score

This goes back to concentrating the spend on the initial build to ensure that the experience is very good and sells itself, rather than saving on the spend and not being able to achieve the level of advocacy needed to generate sufficient active customers at lower cost.The minimum viable proposition is really the features and experience required to ensure that the advocacy measure, eg net promoter score (NPS), is high enough so that the cost of acquisition falls to the required price point to produce a positive marginal profit in the target year.

So the first and most critical key performance indicator is the customer experience differential between the existing competitors and the service you are about to launch. For example, the 2019 NPS scores for the top 7 retail banks in Singapore do not exceed the legal age for buying cigarettes; and so to make an impact, I would recommend setting an NPS target of double that. If you were to launch with an NPS that was, for example, +20, you would have to spend more to attract the customer to switch over and stay active.

Bankers operating in Singapore would also know that among the Asean countries, customers in Singapore are best at extracting maximum benefits from various banking products. When I was head of consumer banking, I recall that Singaporean customers often knew the terms and conditions as well as - or even better than - the bank's product managers. So, the tactic of incentivising them till your NPS becomes much better might cost you more or not even work in Singapore.

We launched TMRW in Thailand in August 2019. Over the following 6 months, we worked very hard to improve the NPS many times over to +33. We verified the inverse relationship between NPS and the cost to acquire an active customer.

Acquiring customers just by using Google or Facebook is an expensive exercise. One study showed that in 2018, for every A$100 spent by Australian advertisers, A$49 went to Google and A$24 to Facebook. I would recommend all digital banks to design their solution such that you are only paying for active customers from Day 1. Certainly, a CPA (cost per acquisition) model is more expensive than CPC (cost per click), but it makes more sense as you end up with active customers.

Eventually we improved the system so that we could distinguish between customers who only signed up for a deposit account and those who signed up for a bundle of deposit account and credit card. Since the latter generated more income, we were prepared to pay more to acquire them.

Achieving positive marginal contribution is the first sign of success. After this significant milestone, the subsequent strategy will depend on the potential size of the segment you have chosen and what customer share you desire. In a larger market, where the larger customer numbers can make up for low individual ticket sizes, the design needs to defray the higher cost required to launch a more superior experience at a higher level of NPS needed to acquire more customers. To ramp up the level of acquisition and continue to have positive marginal contribution, the NPS must thus move even higher over time.

This simple set of maths gives you the perspective. If you make $5 annual revenue per customer, and your cost of acquisition was $10, spread over 5 years your annual cost of acquisition would be $2. Let's say you add $1 for annual servicing costs, then you would make an annual marginal contribution of $5 - $2 - $1 = $2 per customer. That would amount to a $10 million total contribution if you had 5 million customers.

Thus, your annual fixed cost needs to be less than $10 million if you want to break even on a total cost basis, and this assumes you could acquire 5 million customers. So, the more customers, the higher the annual fixed cost you can afford. Likely at the point where your marginal contribution is zero, the total customer base is much smaller than 4 million, say 1-2 million customers.

So, you need to scale further, but without increasing your cost of acquisition further, as this might cause you to regress to a negative marginal contribution. Your NPS must increase.

In a digital bank, the annual fixed costs would comprise mostly people and technology costs. If your annual fixed cost footprint is much larger than $10 million in the example above, then you would not be able to turn total profit positive even if your marginal contribution is positive, assuming it would be too much of a challenge to target more than 5 million customers.

In a market with better demographics and a steeper increase of customers' disposable income as they age, a case can be made that sustaining more losses gives you a bigger business in the end.

If the market is small, as in Singapore, then you may not be able to afford a $20 million cost base, as you do not have 10 million customers to spread your fixed costs over, even if you capture all of the market. In this scenario, a design that allows you to create and launch a bank at a low initial investment and at the same time achieve the NPS breakthrough lift required is essential. Once that initial experience difference is established, then more care and prudence are needed to improve or initiate only what has a good payback in terms of cost vs additional NPS lift.

At this stage, we are not even talking about recovering the initial investment put into starting up the bank. Starting a new digital bank is a long-term strategy. It is a business where you put in a lot of costs initially, in the hope of building an annuity stream of income for many years to come.

• Dennis Khoo is the former group head of UOB's TMRW Digital Group, and was responsible for the strategy, growth and delivery of TMRW digital bank. The article is extracted from his book Driving Digital Transformation: Lessons from Building the First Asean Bank