30th May 2024

Better business. Better community

Business Industry and Financial

Helping banks think like tech companies

As traditional banks grapple with how to compete in a world of digital-native start-ups and changing customer desires, they cannot avoid the problem of aging core technology. The architecture of banks’ legacy systems does not lend itself to achieving the rapid pace of innovation that is becoming the norm for technology businesses. To help with this challenge, Mambu, the SaaS cloud banking platform, offers a backbone on which incumbents (and others) can create an ecosystem of vendors and in-house talent, allowing a more gradual approach to core technology transformation. Mambu co-founder Eugene Danilkis spoke recently with McKinsey senior partner Vik Sohoni, who leads our global digital and analytics work for banking, and with banking technology partners Srinivas Ramadath and Henning Soller. This is an edited transcript of the conversation.

McKinsey: You’re a technologist at heart. Did you aspire to the work you do when you were growing up? You could have become an astronaut. You could have become a soccer player. What made you choose this path, to create the next-gen cloud core?

Eugene Danilkis, Mambu: I don’t know how you knew those were both exactly my childhood dream options—being a soccer player or an astronaut. Seriously, though, I was always really passionate about technology. What I loved was that it felt like a blank canvas. There’s so much possibility for what you can create and the immediate feedback, when you put it in the hands of users, see what they like, and iterate on that. And there’s also the element of reach. When I was at university in the early 2000s, the internet wasn’t as huge and omnipresent as it is now, but you could still feel the potential, and I was incredibly excited by that.

My very first job out of university was at a company that developed mission-critical software for NASA and the International Space Station. It was amazing. We were building technology to enable research and science that could potentially impact millions of lives, if you think about what could happen in space exploration. It was the impact you could have with technology, whether directly or indirectly, that always made me excited about it.

McKinsey: Clients ask us about the rapid evolution in the retail and commercial banking landscape. What’s your perspective on the disruption that comes with it?

Eugene Danilkis: We’re just going into a fundamentally different era of how financial services need to be built, provisioned, and provided to the end users. If you think of the big epochs of financial services, you had an era where there was a small number of financial service providers. The technology that was available to them to create new financial products and services and to improve the user experience was relatively static and built for a relatively slow pace of change.

Over the past decade, there has been much more new technology in the financial services—not just how you design a mobile app or integrate something into a website, but also the back-end technology, including payments, machine learning to help with decisioning, and how we price and customize different products on a micro level for each person. This new level of personalization and automation has only recently become possible.

Whenever new technologies come out like that, someone is always going to put them together to create a great value proposition for the customer. The iPhone, for example, came out because many individual pieces of technology were put together in an amazing package, to the point where it was almost inevitable that it would come together as it did.

I think the same thing is true in the world of financial services. With the most recent developments in artificial intelligence and everything that’s happening around payments and open banking, there are fundamentally different ways to reimagine creating products for different types of customers, whether they are retail or commercial banking customers. You have so much opportunity to take all these different technologies, combine them, and then continuously recombine them to keep creating better financial products and services. And given the speed at which new technology is coming out, that’s only going to continue to accelerate. You have to almost rethink your operating model and technology to be able to keep up.

McKinsey: What is Mambu’s approach to participating in the evolving digital financial services scene and core banking technology transformation? How did you think about building your offering?

Eugene Danilkis: Our entire philosophy was built around questions such as, “What do banks really need? What are the mission-critical services that allow you to design and serve a specific set of financial products and services in the retail and corporate spaces? And what’s essential to that?” Then we asked, “Which of those pieces have existing providers, and which require creating something new that could be a differentiator?” For example, maybe you don’t just use one provider for credit scoring, but you do it differently, market by market. Or maybe you use a mix of options to build your pricing or credit decisioning model—maybe a traditional provider plus one or two new fintechs and an in-house team of five people that custom codes your model.

There’s huge opportunity around all these different areas, so we decided to focus on providing the cloud infrastructure, service levels, reliability, and performance that enables all these domains of opportunity and to build an open platform that other pieces can plug into.

McKinsey: The field of cloud-native core technology offerings for traditional banks is growing rapidly. How does Mambu’s offering distinguish itself in this increasingly crowded and fast-paced field?

Eugene Danilkis: We’re not necessarily differentiated in the sense of the business features of core banking—interest calculations, fees, accounting rules, and so forth. The differentiator is in the how. With Mambu, you can create a financial product, whether it’s a mortgage, or credit card, or personal loan, literally while you’re having a coffee in the morning, including the complete behavior of that product, the application side of it but also the whole system—security, performance, and everything else. And you have the full APIs to have your developers build a proof of concept in the next week and then bring that to market in the next few months. It’s that speed that you’re enabling. And you’re derisking, because the systems are self-contained and talk to each other via APIs. It’s very easy to test the entire product. All of that allows you to continuously deliver distinctive and delightful financial experiences for your customers. The real trick is not necessarily in the pure feature you offer, but how customers receive it, how you embed it into your operating model, and what that unlocks for the rest of your business.

We’re not necessarily differentiated in the sense of the business features of core banking—interest calculations, fees, accounting rules, and so forth. The differentiator is in the how.

McKinsey: Core technology transformation is deeply difficult work complicated by the twin pressures of aging infrastructure and growing competition from fintechs and other disruptors. Talk about what your customers are going through.

Eugene Danilkis: What we’ve seen historically is a lot of different decisions to buy, build, or customize in-house that created bloat around the core, to the point where it became numerous different services, functions, and capabilities all kludged together into one system. It made sense at the time because there weren’t many other choices.

What makes this change right now difficult is that you have a legacy system that’s very tightly coupled together. You think about changing one process, and you suddenly have a lot of other things that are closely integrated into it that then require a lot of change. Or a change around one system potentially puts a lot of risk around a different part of your technology. And a huge amount of any such project could simply be testing to be sure you don’t completely break everything with one new feature.

If you look at the speed at which fintech companies release new products and update existing ones, it is literally an order of magnitude faster than traditional companies. They can afford to make all those changes because they cause less disruption than legacy tech changes do.

McKinsey: What should banks do now, though? Should they build or buy the new core technology that will let them keep up with their digital-native competitors? At a time when they are continually bombarded with new tech options, how can they really understand their needs and the appropriate technology for meeting them?

Eugene Danilkis: As you rightfully said, someone has to decide what those new pieces are going to be and plug them in together in a way that works. A lot of it is scouting the market, identifying the different technologies out there. Then banks have to ask how they can put those various technologies together to create a better proposition for the customer and do it in a way that allows the products, services, and technologies that run them to change and adapt often. And then they have to ask how much of the new technology will be done with in-house technical staff, versus third-party help.

The answers to these vary case by case. It’s a philosophy that banks have to own the conceptual model and stack, taking responsibility for deciding which pieces are right for them, because the technical choices and decisions you make determine how you continue to innovate and differentiate in the market. Banks can’t simply trust one vendor to say, “Here is your whole stack,” because it’s impossible for any single provider, no matter who it is, to be able to provide everything you need to innovate and combine those pieces in a way that’s appropriate for your particular business as a bank.

That curation of partners and technology to build an ecosystem that works for a specific bank is a huge part of the job and the challenge today. This is also true of actual technology companies; they don’t build everything themselves. Rather, they specifically ask, “Okay, what do we not have to build?” They try to build as little as possible, because they know there’s so much other technology they can cultivate from.

We almost fell into this trap ourselves at Mambu once, when we started to build a little piece of technology, and our CTO said, “You know, this is actually a product that one of the cloud providers has.” Not only that, but the cloud provider might have hundreds of people working on that product, whereas we have a skunk works team of three or five people. Obviously, it doesn’t make sense for us to compete with that.

McKinsey: Can you share some examples of how Mambu works with banks to dive into core technology transformation?

Eugene Danilkis: A large traditional bank wanted to shorten their time to market, so we helped them build a “digital speedboat.” It was a specific digital-only product targeted at a specific subset of new or existing customers. They created a fundamental concept and minimum viable product [MVP] in just 13 days and got it into the hands of customers within three months. Our platform makes integration easy, so they could spend all their energy on designing their product and experiences and on creating other services around it. After launching the new product, they kept that 13-day cycle going, continuously tweaking the financial products they could offer, the in-app user experience, and the onboarding journey, and optimizing for costs.

McKinsey: Is there a cautionary tale or pitfalls to be aware of when engaging in this level of change?

Eugene Danilkis: We had one established bank come to us after achieving an unanticipated level of success with a digital product they had built onto their legacy core system. Because of the technology they used on top of it and the stronger-than-expected customer adoption, they were struggling to keep up with the success. Their costs started to increase as a result. They started to have performance issues and trouble scaling. They actually had to pull back some marketing campaigns and expansion because they were getting hit on the technology, infrastructure, and cost side.

They moved over to Mambu and thought through their supporting systems, operations, and processes. They think they saved about 50 percent of run-rate IT costs this way, in addition to getting back up to signing up close to 100,000 new customers a month. And now they have multiple-million customers and are thinking about international expansion.

McKinsey: Given the complexity of core tech transformation, the challenges banks face in the process are myriad. How should institutions contemplating it think about the shift they’re about to make?

Eugene Danilkis: Really understand what you’re going through: you should think about your entire end-to-end operations and design for your target state. You’re not just replacing one piece of technology with another one. The technology, fundamentally, is a different enabler in terms of how you run your operations, which other systems you use, and how you integrate things. You’re redesigning to fundamentally work and act much more like a tech company in terms of your organization, your processes, your speed, et cetera.

You should think about your entire end-to-end operations and design for your target state. You’re not just replacing one piece of technology with another one.

And you want to really align your organization and culture so everyone understands this is not just a core system replacement. It’s stating clearly, “We want to be this kind of company. We want to release improvements to our customers every week. We want to be able to use a new technology that comes out in the market and plug it into our business quickly. And we also want to be able to unplug things from our business within that same time frame when they’re no longer fit for purpose or best-of-breed solutions for us.”

You might say, “We want to be a company that’s not spending 80 percent to just maintain our current customer base. We want to be a company that invests over half of our resources to continue to innovate around new value to customers, reducing our costs and creating new products.” You’re painting the picture of where you want to be and using technology as an enabler.

It’s a mindset of trying to continuously identify the quickest wins from a business perspective. You’re looking across your portfolio of business lines or products or markets and thinking, “Where can we have the most immediate impact with potentially the least amount of risk from an execution perspective? Let’s achieve that. And then let’s do another one and another one.”

You have to be comfortable with the unknown in that journey. You know what kind of company you want to be, but you don’t necessarily know what technology you will eventually have, but you’re comfortable with that because the goal is to be able to evolve continuously.

McKinsey: What are the technological considerations involved in building a strong innovation capability while still maintaining regulatory compliance?

Eugene Danilkis: It’s about having systems and processes that are flexible enough that you can shift without having to take on massive projects and risks. In that sense, the importance of the element of flexibility and agility of technology and organization applies equally to innovation as it does to responding to these external requirements such as regulatory change or even a crisis such as the pandemic, because all of these require making changes to the core systems.

The challenge we’ve seen in the past of building some new technology alongside the legacy systems is that operating at two speeds is hard. If there are new regulations or compliance requirements coming, it becomes difficult to just apply the change to the product or process involved without having to change many other things around it. So having things siloed and moving in two gears doesn’t work very well in systems that are supposed to play together.

The challenge we’ve seen in the past of building some new technology alongside the legacy systems is that operating at two speeds is hard.

McKinsey: What holds banks back from implementing core technology transformation, and what enablers can help them move past those barriers?

Eugene Danilkis: I think they’re not quite sure where and how to start, because it feels like a huge, intractable problem to them. The trap that a lot fall into is they say, “Well, if our technology’s going to be in the cloud in five or ten years, then let’s start looking at our technology.”

That’s not actually the right place to look. The technology and the cloud are just enablers. The real key is your customers—those you have now and those you want to have in the future, those that love you and those that don’t like you but whom you want to retain over the next five or ten years. You want to start designing around the questions of what products and services should look like to those customers and what your cost base would need to be to build them.

Then you can start to think about what technology you need to support all of that. Will the new products and services use system A or system B? Will it run in the cloud or on my old data center? And that’s where you have to come back to that first question of what kind of organization you want to be. If you want to operate like a tech company that runs at speed, that has the same sort of customer satisfaction, margins, and customer experiences as a tech company, then you’re going to be moving to the cloud. And if you want to be in the cloud, then you should be choosing cloud solutions as much as possible, starting now.

Consistency in these types of smaller decisions regarding the business lines and individual projects eventually compounds and creates the biggest overall impact over time. It’s like a fitness program. One bag of chips or one apple doesn’t necessarily make or break your program, but if your goal is to be at a different state from where you are now, then you have to make more and more choices that are consistent with that goal along the way.

McKinsey: High-quality engineering talent is key here, so how can banks make themselves more attractive to the engineers and developers they need to pull off core technology transformation?

Eugene Danilkis: Banks have a lot to offer to tech talent, conceptually. They have a direct impact on the lives of millions of individuals, and the change they’re going through now has a huge impact on how commerce happens, which affects everything.

The fact that core banking systems are mission critical and that it’s a regulated industry does not mean there isn’t room for creativity and innovation. Technologists are fundamentally creative individuals, and they are completely aware that working on an app that puts filters on photos has a very different risk profile than working on products related to people’s life savings or mortgage or to corporate accounts.

McKinsey: Banks face an unprecedented mix of pressures at present, including the need to resolve their tech debt. Old systems are nearing their end of life, the talent who run those systems are retiring, and somehow they need to innovate at the same time as overhauling their core technology. Solving such challenges typically involves long, multiyear lead times. How do the financial services executives you interact with decide where to invest and when?

Eugene Danilkis: If you’re starting with an assumption that anything is going to take five years plus, then you won’t do anything, because you’re not even sure if you’re going to be in the same job at that time, so why bother starting something? The first task is to start shortening time cycles to be able to create results and outcomes in well under 12 months. Three to nine months is ideal.

Then you have to reduce the scope. There’s no other way around it if you’re going to have fast outcomes. Scope reduction means you’re thinking in terms of where you can find quicker wins. Rather than trying to solve everything, you take one slice of a business line and get results with that. You’re still going to have legacy tech issues to work out in years four to five, but at the end of year one, you’re going to have higher customer satisfaction, higher margins, and you’ll be building the ability to innovate externally or internally in a matter of weeks instead of months. It starts to become a mentality of “What can we achieve in the next couple of quarters?” and not “What can we achieve over the next couple of years?”

Of course, you need to be clear on the vision for the next couple of years, because otherwise you start to lose the context of why you’re overhauling individual pieces of a business. You have to stay focused on the fact that you’re going to sunset the legacy tech over the next three to eight years but you’re not replacing it this year. You’re innovating at the same time and changing the culture and ways of working, so that new tech talent is excited to work on these challenges, and you’re not worried about them retiring as you do the legacy tech talent. Your customer satisfaction scores are higher, your cost to serve customers is going down, and over three, five, or eight years, you retire your legacy systems. That’s going to be the end state for everything. We’re not doing five-year projects; we’re starting here, now.

McKinsey: Five to seven years down the road, what do you think will be the defining aspects of the change we’re undergoing?

Eugene Danilkis: I think we’re going to see a rethinking of the role of human contact and which experiences should be digital. With all the focus on the benefits of automation, there’s a potential risk of the pendulum swinging a little bit too far. There are particular points in the customer experience where human interaction can have a really strong impact on improving satisfaction. I think we can combine automation and human contact in a way that makes the experience smarter and more personalized.

With the amazing progress in machine learning over the last couple of years, there will be more and more opportunity to streamline transactional pieces of the customer experience where nobody benefits from a human being involved and to understand better which are the right human-to-human interactions, when they need to occur, and how to make automation and human interaction operate to complement each other. I don’t think the world of financial services has quite got the balance right yet. It’s still on the horizon.

McKinsey: Eugene, thank you so much for sharing your story with us.

Eugene Danilkis: Thank you. It was a pleasure.