# The four core models: payments, lending, neobanks, and embedded finance
You buy a $100 pair of sneakers online. Behind that single click, three different fintech businesses could be fighting to make money off you: the company that moves your payment, the one that lets you split it into four installments, and the app where your money actually lives.
Each earns in a completely different way. Each breaks in a completely different way. If you understand these four models, you understand roughly 90 percent of what the fintech industry does.
Let's dissect them one transaction at a time.
Payments companies move money from a buyer to a seller and take a small cut. Think Stripe, Adyen, or PayPal.
The classic example is a card transaction. When you pay that $100, the merchant does not keep all of it. A slice gets carved up among several players in what the industry calls the interchange system (the fees set mostly by card networks like Visa and Mastercard).
A rough, commonly cited breakdown for a US online card payment:
Stripe's headline US price is famously about 2.9 percent plus $0.30 per transaction. On your $100 order, that is roughly $3.20. But Stripe does not keep $3.20. Most of it passes through to the issuing bank and the network. Stripe's actual gross margingross marginGross margin is the share of revenue left after subtracting the direct cost of producing goods or services, expressed as a percentage of revenue.Voir la définition complète → is a thin sliver of that fee.
This is the key insight: payments is a volume game with tiny margins per transaction. Stripe might net well under 1 percent of the amount processed. That only works at massive scale.
The model breaks when:
For a deeper mapmapUsing software to automate repetitive marketing tasks and campaigns, enabling personalisation at scale across channels like email, web, and social.Voir la définition complète → of who takes what, the Federal Reserve's payments research is a solid free reference.
Lending companies give you money now and get paid back more later. The difference is their revenue. Klarna, Affirm, and most Buy Now Pay Later (BNPL) firms live here, alongside consumer and small business lenders.
Take a Klarna "pay in four" plan on your $100 sneakers. You pay $25 now and $25 every two weeks. The customer often pays zero interest.
Two main ways:
1. Merchant fees. The store pays Klarna a cut (often estimated at 3 to 6 percent) because BNPL increases sales and average order size. This is Klarna's biggest revenue line for interest free plans.
2. Interest and late fees on longer term loans, where the customer pays over months, not weeks.
Here the enemy is credit risk, meaning the chance the borrower does not repay. That is called a default, and the loss on it is the credit loss.
Lending margins look fat compared to payments, but they are not free money. Out of that merchant fee and interest, the lender must subtract:
The model breaks when credit losses spike. In a recession, more borrowers default at the same time, and a lender that priced its loans for good times can suddenly lose money on every cohort. This is why lending is fundamentally a risk-underwriting business: the whole skill is predicting who will repay.
Same sneakers. Totally different business.
A neobank is a bank with no branches, delivered entirely through an app. Nubank in Brazil, Revolut in Europe, Chime in the US. Some hold their own banking license; others partner with a licensed bank behind the scenes.
Here the product is not one transaction. It is you, as a long term customer, using many products.
Revenue comes from a mix:
The magic number is CAC versus LTV:
Neobanks win when a cheaply acquired customer (often via referral, which is why Nubank grew so fast in Brazil) sticks around for years and gradually adopts more products. A user who only holds a free account and never borrows or pays a fee can actually be unprofitable.
The model breaks when:
Nubank's story matters here because it showed a neobank can reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.Voir la définition complète → profitability at very large scale by relentlessly cross-selling credit, but many rivals in richer markets have struggled to convert users into profit.
Vérification des acquis
1. Why does the payments model require massive transaction volume to be viable as a business?
2. In a typical card transaction, which player captures the largest slice of the fees through interchange?
3. A founder proposes a payments startup targeting a niche of low-frequency, high-touch transactions with modest total volume. What is the core structural risk with this model?
4. Select ALL correct answers about the economics of the payments (toll road) model.
Sélectionnez toutes les réponses correctes.
5. Select ALL correct answers describing what the four core fintech models illustrate about the industry.
Sélectionnez toutes les réponses correctes.
Embedded finance means putting a financial product inside a non-financial company's app or checkout. The ride-hailing app that offers drivers a debit card. The e-commerce platform that lends its sellers working capitalworking capitalWorking capital is the difference between a company's current assets and current liabilities, measuring short-term liquidity and the funds available to run daily operations.Voir la définition complète →. The software tool that lets you invoice and get paid without ever visiting a bank.
The provider making this possible is often a Banking-as-a-Service (BaaS) firm: it holds the license, the compliance, and the plumbing, then rents it out through an APIAPIApplication Programming Interface: a standardised interface that lets applications communicate and exchange data without knowing each other's internal workings.Voir la définition complète → so any company can offer banking features.
The economics split between two parties:
The appeal: the platform already owns the customer relationship and the data, so acquisition costacquisition costCustomer Acquisition Cost (CAC) is the total sales and marketing spend divided by the number of new customers gained in a period. It measures how efficiently you grow.Voir la définition complète → is near zero and it can underwrite lending better than an outsider. A software company that sees a merchant's daily sales knows exactly how much that merchant can safely borrow.
The model breaks when:
Embedded finance blurs the line between "tech company" and "financial company." The sneaker store from our opening does not want to become a bank. But with embedded finance, it can offer the checkout loan, the branded card, and the payment processing without building any of it. Finance becomes a feature, not a destination.
Run the same $100 sneaker purchase through each model:
Most large fintechs eventually blend these. Nubank does lending and payments inside its neobank. Stripe now offers lending and embedded banking on top of payments. The categories are lenses, not walls.