Be fundamentally different, not incrementally better

Executive overview

Nubank's explosive growth—80-90% through word-of-mouth, more customers than Bank of America across three Latin American countries—stems from a ruthless commitment to building products people love fanatically rather than products people tolerate. The company achieves this through a distinct culture that prioritizes customer obsession, clarity of strategy, and discipline around product-market fit before scaling. The core insight: fundamentally different products drive word-of-mouth growth; incrementally better ones require expensive marketing.

The five values that drive everything

1. We want our customers to love us fanatically
2. We are hungry and we challenge the status quo
3. We build strong and diverse teams
4. We pursue smart efficiency
5. We think and act like owners, not renters

Operationalizing customer love

• Mock press release first: Before a single engineer writes code, explain why the customer should care in two paragraphs. If you can't articulate it, the work isn't ready.
• Weekly product design reviews: Ask relentlessly: Why is this great? Why is it fundamentally different? Does it redefine the category on quality, complexity, and price simultaneously?
• The Sean Ellis score as a gate: Measure product-market fit by asking customers how disappointed they'd be if the product disappeared. Nubank's threshold: 50% "very disappointed" (adjusted for Brazilian optimism). Most products struggle to hit 20%—Nubank demands 50% before scaling.
• NPS obsession: World-beating net promoter scores (70s, 80s, occasionally 90s) are tracked religiously. A 1-2 point dip triggers investigation.
• Zero tolerance for leaky buckets: Churn destroys word-of-mouth loops. Retention metrics are as sacred as growth.

From survey insights to ruthless iteration

When a product hit the Sean Ellis gate with only 40%, Nubank dug deeper and found a micro-cohort with extreme love (70% very disappointed). The payments assistant product was complex for customers across four Brazilian payment rails. The insight: success came from customers managing multiple bills across multiple rails automatically. After dozens of iterations on consolidation and multi-rail onboarding, the product went from hundreds of thousands to 10+ million monthly actives in 15 months.

The method: Don't rely on surveys alone. Call 10 customers directly. By call five, you'll predict calls six through ten. Brazilians are expressive; tone conveys what statistics obscure. One PM calling 10 customers beats a researcher analyzing 1,000 and summarizing back through layers.

How to push back and say "not yet"

Product managers inherit intense pressure to scale immediately. The job is often to refuse.

• Culture of ownership, not rentership: Ownership means telling the hard truth early, even when it's inconvenient. Your job as a leader is to practice saying "I don't think it's ready" before you get a C-suite title.
• Respect for past pivots: Senior executives who've seen products fail and been pivoted know the cost of scaling too early. They respect clarity and data, not blind agreement.
• The big mess lands on you: If you scale a broken product, it's your mess to untangle later. Pivot when it's small; scale when it works.

User research: Five core disciplines

• Anecdote beats data: Qualitative insights from real conversations trump sophisticated analyses that have lost the original question.
• Crisp hypothesis first: Without a clear hypothesis, research becomes conversation, not validation. Know what you're testing before you test.
• Avoid falling in love with your hypothesis: Be a judge, not a lawyer for your own idea. The research should invalidate as easily as validate.
• Observe more than ask: Ask indirect questions, not leading ones. Watch what customers do rather than what they say they'd do. The Swiffer came from observing the pain of mopping, not asking about mopping.
• Strong hypothesis examples: "Joint bank accounts are 120-150 years old, built when women couldn't have independent accounts. Modern financial life is social; sharing a savings goal should be as easy as a Spotify playlist." Go in with this view, then observe where it breaks.

Strategy: Clarity over correctness

Kevin Systrom's famous line: "We may not be right, but at least we are clear."

A strategy is not a vision, goal, or aspiration. It's a coherent plan: specific customer, specific problem, specific solution, specific advantage.

Nubank's 2014 strategy: Young, middle-class, urban Brazilians hated credit card fees and couldn't access credit due to age. A branchless, digital-first bank could deliver a no-fee card at disruptive pricing. That's specific, locked in, actionable.

When strategy is vague ("become the world's largest neobank"), execution diffuses. When strategy is crisp, focus emerges. Concentration is what builds wealth in startups; diversification preserves it.

To develop this muscle:

• Read frameworks: Good Strategy, Bad Strategy; Play to Win; Where to Play, How to Win.
• Practice making concentrated bets, not hedges.
• Accept that focus requires courage and feels like desperation to outsiders.
• Improve by doing it repeatedly under pressure.

Category design and the new category debate

Some argue: enter categories where customers already spend money. Others: create new categories.

Nubank's answer synthesizes both. The company was first in digital-first branchless banking in Brazil—that was new. But when entering adjacent categories like consignado (secured lending for government employees), Nubank invents within existing markets. The job: be fundamentally different, not just better positioned.

The real principle: incumbents control their own playing field. To break through, find a disruptive vector that's categorically new. Incrementally better doesn't cut through noise anymore; fundamentally different does.

Launching and scaling product lines

Nubank launched from a Brazilian credit card company to a full Latin American bank over five years. New products: bank accounts (primary relationship), lending, investments, insurance, small business services. Now launching into Mexico and Colombia with the full suite.

How to sequence new products: There's less science than desired. Smart people, robust debate, and constant reexamination. Some principles:

• Know the TAM and where customers pay (obvious, but often ignored).
• Use the 50% gate religiously: Ultravioleta (rewards credit card) launched July 2021, didn't scale until July 2023 or later. 18+ months in the lab validating and iterating. Only then aggressive scaling.
• Avoid fixing big problems you scaled too early: This is how you avoid 80% of catastrophes.

The vision: AI-native global banking

Nubank is building toward a world where everyone has a personal banker in their pocket, powered by social mechanics and self-driving automation.

Three principles guiding phase three:

1. Holistic banking across life seasons: Spend, save, invest, borrow, protect—full solution, not best-of-breed.

2. Single global codebase: Most global banks (Citi, Santander) are mosaics of acquisitions on different tech stacks. Nubank is building a single bank across 40+ countries and languages, respecting local regulation while leveraging global infrastructure. Google required this for ads; Nubank is importing the discipline into fintech.

3. Social and self-driving automation: Why remember to pay a bill? Why manually track savings goals? Millions hire personal bankers for guidance. AI can democratize that to everyone.

Customers live "harshly unoptimized financial lives" regardless of income. The next layer: observing what life event you're in (new child, marriage, home purchase), then automating recommendations and decisions for that stage.

Integrating AI into product

Using ChatGPT and standard tools for customer experience improvements and efficiency. But the real work: understanding what "AI-native" means—not appending AI at product edges, but designing products where AI is foundational.

Facebook's mobile migration is the analogy. They didn't add mobile features; they reimagined the entire product mobile-first. Nubank is asking: What would financial products look like if built for AI from day one?

Failure and resilience

Jag's seminal failure: The Google radio and terrestrial TV advertising business. Hypothesis was sound; methodology was different. Broadcast markets have high inventory scarcity, fragmented control, and require brand advertising expertise. Google pivoted: rather than chasing TV, have TV come to Google. YouTube. The small team became "collateral damage."

The lesson: Within weeks of dropping out of two grad schools to join Google, facing layoff while newly married. Song on repeat: "Just because I'm losing doesn't mean I've lost." Bloody-mindedness—persistence while staying clear-eyed about odds—is underrated. A year later, that team built Google's display advertising business, still Google's second-largest.

Core mantras

• We're in the business of fundamentally different, not incrementally better.
• Good enough isn't good enough. Is it great enough?
• We will not scale small problems; we solve them small, then scale the solution.
• Culture of ownership means you speak truth early, even when inconvenient.
• Strategy is clarity first, then correctness follows.