Why Eliminating Profitable Features Builds Stronger Trust

The hidden architecture of trust: What Max Levchin saw that everyone else missed

Max Levchin didn't build Affirm because he saw a gap in the lending market. He built it because he saw a gap in how the whole system was wired. And he realized that stripping away the industry's most profitable features would actually create a stronger, more defensible business. This conversation shows why the obvious fix (charging less) is often the wrong frame, and how the real advantage comes from redesigning incentives so that doing the right thing becomes the only rational choice. Anyone building a product where trust is the currency (and that's most of you) needs to understand how Levchin mapped the full causal chain from predatory fine print to talent acquisition, and why the most uncomfortable decisions are often the ones that compound over decades.

Why the most profitable features are the most dangerous

Conventional banking wisdom says late fees and revolving debt are profit centers. Cut them, and you cut your margins in half. Levchin heard this from every banker he pitched. But he saw something they didn't: those "profit centers" were actually liabilities disguised as revenue.

"We've never charged a penny of late fees, never charged a penny of revolving interest."

Here's the systems-level insight: when you eliminate the ability to profit from customer mistakes, you force yourself to get good at something harder. Underwriting. You can't rely on the crutch of punitive fees to cover bad lending decisions. So you build better models. You attract better talent. And over time, that underwriting capability becomes a moat that competitors can't replicate because they're still addicted to the easy money.

The downstream effect is beautiful. Most lenders optimize for the wrong thing: maximizing interest income per customer. That leads to products that are deliberately confusing, with fine print designed to be ignored. Levchin realized that if you optimize for transparency instead, you don't just win customers. You win the kind of customer who stays. The millennial generation, scarred by the 2008 financial crisis, is actively looking for a reason to leave traditional banks. They don't need to be convinced; they need permission. Affirm's structure gives them that.

The talent arbitrage nobody talks about

The second-order consequence of building a transparent lending product is that it attracts the people who make the product work. Levchin noticed something strange: brilliant mathematicians and computer scientists avoid the lending industry. Not because the problems aren't interesting (they're deeply mathematical), but because it's embarrassing to explain at cocktail parties that you're optimizing late fee collection.

"I'm going to get my unfair share of really brilliant mathematicians because they're not going to go to Wall Street. They'd rather come to work for me and build underwriting models trying to help people in normal America borrow money and not get screwed."

This is a classic systems move. Most companies compete on features or price. Levchin competed on moral clarity. By making the product pro-social, he unlocked a talent pool that was otherwise off-limits. The result? Underwriting models so good that even at $50 billion in transaction volume, loss rates remain consistent. The transparency isn't just a marketing gimmick. It's the engine that powers the underwriting flywheel.

The implication for any founder: the most talented people in any field are often avoiding the most important problems because those problems have been captured by industries with bad reputations. If you can clean up the reputation, you get the talent. If you get the talent, you get the execution. If you get the execution, the growth takes care of itself.

The friction that creates certainty

Silicon Valley worships frictionless experiences. Tap and go. One-click checkout. Levchin argues that for high-stakes purchases (anything above a sandwich), friction is actually a feature. The problem with credit cards isn't the speed; it's the uncertainty. You don't know what your debt will look like in six months. You don't know what hidden fees will appear. That uncertainty creates anxiety, which creates hesitation.

Affirm's friction (opening the app, seeing your purchasing power, getting explicit approval) is actually a trust-building mechanism. It replaces uncertainty with certainty. The system responds by making customers more willing to transact because they know exactly what they're getting into. Over time, this creates a feedback loop: more data on good behavior leads to better underwriting, which leads to more approvals, which leads to more transactions.

This is where the "whenever there is any doubt, there is no doubt" quote becomes operational. Levchin applies it to hiring, to product decisions, to marriage. The principle is simple: if you're analytically trying to convince yourself of something that doesn't feel right, you already know the answer. The hard part is acting on it before the doubt compounds.

The marriage model for co-founder dynamics

Levchin's advice on marriage ("don't go to bed angry, stay up and fight") isn't just relationship advice. It's a systems insight about how to handle conflict in any high-stakes partnership. The natural tendency is to let small resentments fester because it's uncomfortable to address them directly. But in a system where you're overlapping professionally and personally, those small resentments compound faster than they would in a normal relationship. The only way to prevent the system from breaking is to force resolution quickly.

The deeper insight is the "marry up" dynamic. Levchin says the secret to his marriage is that he's still trying to impress his wife every day. Both partners secretly think they got the better deal, which creates a mutual incentive to keep growing. Apply this to co-founders: if both people feel they're lucky to be working with the other, they'll invest more in the relationship. If one feels they're carrying the other, resentment builds and the system degrades.

Key action items

Eliminate the profit centers that depend on customer failure. Over the next quarter, audit your revenue streams. Ask: does any part of our business model rely on customers making mistakes, forgetting, or not reading fine print? If so, redesign it. The short-term revenue hit will be offset by long-term trust and talent acquisition.

Build the product that attracts the talent you want. This pays off in 12 to 18 months. The best people won't work on problems they're embarrassed to talk about. If you're struggling to hire, the problem might not be your compensation. It might be your moral clarity.

Add friction where it creates certainty. Identify the moments in your customer journey where uncertainty causes hesitation. Add explicit, transparent steps that replace ambiguity with clarity. The friction will feel uncomfortable at first, but it builds trust that compounds.

Apply the "doubt" rule to hiring decisions. Over the next quarter, whenever you feel analytical doubt about a candidate, trust it. Levchin's experience: "It's always I should have done that a lot sooner." The cost of a bad hire is far greater than the discomfort of passing on a marginal candidate.

Institutionalize the "stay up and fight" rule. Within your team, create a norm that disagreements are resolved quickly, not deferred. This is a longer-term cultural investment (6 to 12 months to embed), but it prevents the compounding resentment that destroys partnerships.

Track what matters, not everything. Levchin evolved from tracking everything to tracking resting heart rate and HRV as the two anchors. Identify the 2 to 3 metrics that actually predict your performance and stop measuring the rest. The marginal gain from tracking fingernail clippings is negative.

Recognize that the best competitive advantage is doing the hard thing others won't. Affirm's no-late-fee policy seemed insane to bankers. That's precisely why it worked. Over the next quarter, identify one "obviously wrong" decision that would actually make your product better if you had the courage to implement it. Then do it.