Latest / Business Strategy Talks with Fexingo: Competitive Advantage, Market Positioning, and Growth Planning / How Nubank Built a Moat on Data and Disciplined Growth
Transcript
- Lucas: So, Luna, we've spent the last few weeks talking about moats built on everything from distribution networks to intellectual property. But today I want to look at a company that built a moat in one of the most competitive spaces there is: consumer banking. Luna: And you're not talking about a traditional bank, right? You're talking about Nubank. Lucas: Exactly. Nubank, the Brazilian digital bank. It's a name that might not be a household one outside Latin America, but it's a case study in how to use data and product design to break into a market that was basically a cartel for decades. And I think the moat they've built is genuinely underappreciated. Luna: I'm intrigued. And if today's episode is useful to you, the way we keep these conversations going ad-free is listener support. If you've ever gotten something out of these talks, you can buy us a coffee at buy me a coffee dot com slash fexingo. It genuinely helps. Lucas: Yeah, it does. And we appreciate everyone who does that. So back to Nubank. The question isn't whether they have a moat — they have 90 million customers across Brazil, Mexico, and Colombia. The question is how they built it, and whether it's defensible. Luna: And the obvious answer would be, well, they're cheaper. No branch costs, no legacy IT. But that's not a moat, that's just a margin advantage. Lucas: Right. And plenty of digital banks have tried that playbook and burned through cash. The thing that separates Nubank is how they've leveraged data — specifically, data on their own customers' behavior — to drive better credit decisions and lower acquisition costs. And that's a flywheel that's tough to replicate. Luna: So walk me through the data moat. Because I've heard that term thrown around a lot, but what does it actually mean for a bank? Lucas: Let's start with credit. In Brazil, the traditional banks had a massive information asymmetry. They had decades of relationship data, and they used that to charge sky-high interest rates — we're talking some of the highest in the world. Nubank came in with no data initially. But they started with a simple product: a no-fee credit card with a mobile app. Luna: And that gave them a foothold, but how do you build a moat from that? Lucas: The moat comes from the flywheel. Each customer interaction — every transaction, every payment, every card swipe — generates data that feeds into their proprietary machine-learning models. They use that to assess creditworthiness more accurately than the incumbents, especially for the underbanked. That means they can approve customers the big banks rejected, and they can price risk more precisely, which lowers their loss rates. Luna: So it's a classic data network effect. More customers, more data, better models, better products, more customers. Lucas: Exactly. And the credit card was just the entry point. From there, they expanded into digital accounts, then personal loans, then insurance, then investments. Each new product feeds back into the same data engine. And the switching costs start to compound. Luna: Because once you have your paycheck, your savings, your investments, and your credit card all in one app, it's a hassle to leave. Lucas: Right. And that's a different kind of moat — the classic switching cost. But there's another layer I think is even more interesting: their customer acquisition strategy. Nubank didn't spend billions on marketing like some other fintechs. They grew through referrals and word of mouth. And the reason word of mouth works is that the product is genuinely better — the app is simple, customer service is famously good, and there are no hidden fees. Luna: That sounds like a cult brand, almost. But how does that make a moat? Anyone can have good customer service. Lucas: But not everyone can have a customer acquisition cost that's effectively a fraction of the industry average. Nubank has said their cost to acquire a customer is significantly lower than traditional banks, and it's even lower than many other fintechs. And when you combine that with the data-driven credit models that keep default rates low, you get a unit economics story that's pretty compelling. Luna: So they're not just growing, they're growing profitably. I think they hit profitability a couple of years ago, right? Lucas: They did. And that's rare in fintech. But here's where the moat gets tested. You can't just grow in Brazil forever. They're now pushing into Mexico and Colombia, and those are different markets with different incumbents and different regulatory regimes. Luna: And the data moat doesn't automatically transfer, does it? You have to start from zero in each new country. Lucas: Exactly. The models Nubank built in Brazil are trained on Brazilian behavior patterns. In Mexico, the credit bureau data is different, the consumer behavior is different, even the fraud patterns are different. So they've had to rebuild that data engine from scratch. And that's expensive. Luna: So is the moat weaker in Mexico? Or is there something else that gives them an edge? Lucas: I think there are two things that carry over. One is the brand and the playbook — they know how to launch a product, they know the customer service model, they know the regulatory dance. And the other is capital. They've raised a lot of money, and they're public now, so they can afford to invest in building that data engine in new markets. Luna: But that's a moat of attrition, not of innovation. Anyone with enough money could do that. Lucas: True, but not everyone has the operational discipline. And there's another angle I want to bring in: the regulatory environment in Brazil. The central bank there has been a pioneer in open banking, and that's actually created a different kind of moat for Nubank. Luna: How does open banking help Nubank? You'd think it would help the new entrants, not the established player. Lucas: It's counterintuitive, but because Nubank has the scale and the data infrastructure, they can actually ingest open banking data more effectively than the incumbents can. They can take a customer's data from other banks, run it through their models, and offer a better product. So instead of being a threat, open banking becomes another data source that feeds their flywheel. Luna: That's a really interesting point. The incumbents have the data but they don't have the agility to use it, and Nubank has the agility but needed the data. Open banking kind of levels the playing field, but Nubank is better positioned to exploit it. Lucas: And that's the thing about moats. They're not static. Nubank's moat is a combination of data, brand, cost structure, and regulatory savvy. But it's also worth asking, who's trying to attack it? Luna: The big banks aren't just sitting there. They've launched their own digital apps, and they have the existing customer base. Lucas: Sure, but the incumbents have a fundamental conflict. If they make their digital experience too good, they cannibalize their branch-based revenue. And their legacy systems are a huge drag. Nubank doesn't have that problem. But the bigger threat might be other digital banks, especially in Mexico, where there's been a wave of neobanks. Luna: So the moat is not unassailable. But it's strong enough that they've become the most valuable financial institution in Latin America by market cap, right? Lucas: I think they're up there, and that's a testament to how the market values their strategy. But let's talk about a concrete number that shows the moat: their return on equity. I don't have the exact latest figure, but they've been trending towards mid-teens, which is respectable for a bank, and they're growing revenue at a double-digit clip. That combination — growth plus profitability — is what investors pay up for. Luna: So the moat isn't just about defense, it's about allowing them to invest in growth without destroying shareholder value. That's the ideal. Lucas: And that brings me to a question I want to pose to you, Luna. When you look at Nubank's expansion into Mexico and Colombia, do you think they're making a smart move, or are they diluting their moat by spreading too thin? Luna: I think it's a risky but necessary move. Brazil is a huge market, but it's not infinite. And if they want to be a truly regional player, they need to prove the model works elsewhere. The question is whether the data flywheel will spin fast enough in those new markets before the competitors catch up. Lucas: And that's the core challenge. It took them years to build the data engine in Brazil. Can they compress that timeline? They've said they're using some of the learnings, but ultimately they have to collect enough local data to make their models work. Luna: One thing that gives them an edge is their brand. In Mexico, they're starting to become known, and they have the capital to sustain a long ramp. But there are homegrown neobanks like Klar and Albo that have been at it for years. It's not going to be a walk in the park. Lucas: No, and that's what makes it interesting. A moat is only a moat if it holds against determined attackers. And we're seeing the real test of Nubank's playbook right now. Luna: So what's the one takeaway you want listeners to hold onto from this episode? Lucas: For me, it's that data moats aren't just about having data — they're about having a flywheel that turns data into better products, which brings in more customers and more data. Nubank did that in Brazil, and the challenge now is whether they can recreate it elsewhere. And I think that's a lesson for any business: the moat is the system, not the asset. Luna: That's a good way to put it. And I'd add that you have to be willing to invest in the system before you see the returns. Nubank spent years in the red to build this. It's not a quick win. Lucas: Right. And that patient capital is part of the moat too, because not every competitor has the stomach for it. So next time you hear about a fintech, ask yourself: is this a company with a system that compounds, or just a company with a cheap product? The answer tells you a lot about whether they'll still be around in a decade.