Course Map

Route: the three assumptions of a card → what rails grew in the places where those assumptions fail → three ways to turn a local rail into national infrastructure → the same rails wearing other form factors → why the aggregators all integrate everything → whether the card networks are actually in danger

1. The One Question This Course Answers¶

In 2025, credit cards were 20% of global e-commerce volume and debit cards 10%.

Thirty percent between them. The other 70% does not run on cards.

That is not a forecast. It has already happened. If your picture of payments is "swipe a card, Visa, Mastercard, interchange," that picture does not reach a third of the world's online payments.

This is what the course answers: what that other 70% is, where it came from, why it won in some countries and not in others, and whether Visa and Mastercard are really in trouble.

One rule has to be set before anything else, or the names will fool you: the real subject of this course is rails, not QR codes. A QR code, a phone-number alias, a bank redirect, a one-time six-digit code — those are form factors, meaning how a user initiates a payment. The thing underneath that actually moves money from one account to another is the rail. The same form factor can sit on completely different rails: under Singapore's codes is bank instant clearing, under China's codes is a wallet's own ledger, and the card networks issue codes too — scan one of those and a card transaction is still what runs underneath. Never conclude anything from the form factor; look at what runs beneath it. Appendix A turns this rule into a four-layer model, and every chapter in the main line uses it.

The answer to the last one will surprise you. Before you read on, hold on to one fact: while Pix, UPI, iDEAL, and Blik were rolling out across their markets, Visa's and Mastercard's revenue kept growing at more than ten percent a year. "Cards are being displaced" and "the card networks are earning more" are both true at the same time. An explanation worth having must explain both, not pick one.

2. What This Course Does Not Repeat¶

If you have read Payment Systems, this course uses the following directly and does not unpack them again: the difference between clearing and settlement, the push and pull directions, the economics of interchange, the money hierarchy, why NetsUnion exists, payment licensing.

If you have not read it you can still read this, but you need three words:

• Clearing: working out who owes whom how much.
• Settlement: actually moving the money across.
• Interchange: the fee an acquirer pays an issuer — the economic engine of the card business.

If those are new, go look at the first six chapters of Payment Systems. Half an hour is enough.

3. The Thread¶

The course runs in one line: the assumptions → the assumptions failing → each region's own answer → the new problems those answers create. Every chapter answers the question the one before it left open.

Part I · Why an Alternative Exists at All (Chapters 1–2)

Cards run on three assumptions: the merchant has a terminal, the consumer holds a card a bank issued, and an interchange fee feeds the whole chain. All three hold in developed markets, which is why cards won there.

The question is what happens where they don't. A vegetable stallholder cannot afford a POS terminal, and is not worth installing one for.

Chapter 2 covers that answer: replace the terminal with a piece of paper. The QR code is not a technical innovation. It is a redistribution of cost — the hardware cost on the acceptance side is pushed to near zero, and the price is that the cost moves onto the consumer's phone and onto one centralised ledger. But the thing that chapter has to settle at the end is this: the piece of paper only solves initiation. The money still needs a rail to move on. The twelve chapters after it are all about that rail.

Part II · Three Ways to Turn a Local Rail into National Infrastructure (Chapters 3–7)

Three countries built three rails of completely different character — even though what the user sees is a code in all three. This is the best available demonstration of "same form factor, different thing underneath." These chapters are the body of the course, and the place where "contrast reveals the variable" gets used hardest.

Chapter 6 is Southeast Asia, your market. It is a fourth case: nobody dominates, so every country builds its own rail first and drapes a code over it after. That chapter walks the markets one country at a time, rail before code, and explains why there is so much interlinking news and so little interlinking volume.

Chapter 7 puts five rails on one table and asks a single question: who pays this bill. It is the course's first place where everything closes.

Part III · The Same Rails, Wearing Other Form Factors (Chapters 8–10)

QR is only one of the form factors on top of these rails, and not the most common one — Australia's entire NPP has essentially no code, Poland uses a one-time six-digit code, the Netherlands uses a bank redirect. These three chapters ask what happens to the rail and to the bill when the form factor changes. Chapter 8 has to dismantle the most common misunderstanding first.

"Digital wallets are 56% of global e-commerce" gets quoted everywhere, but it lumps together two completely different things: what sits inside Apple Pay is a card; what sits inside Alipay is not. The first is a card wrapper, and the card network still gets paid. The second is a genuine alternative. Miss that distinction and every judgement you make about this industry will be wrong.

Chapter 9 covers the methods that neither scan a code nor run on cards: the Netherlands' iDEAL, Poland's Blik, Brazil's Boleto, Japan's convenience store payments. It uses one very simple piece of arithmetic to show why the Dutch charge a flat per-transaction fee and the Brazilians charge ad valorem — at what ticket size a flat per-transaction fee beats an ad valorem one.

Chapter 10 covers buy now, pay later. It contains an odd fact: BNPL charges merchants roughly double the card rate, and merchants take it anyway. Why they take it is the only question that chapter answers.

Part IV · Why the Aggregators All Integrate Everything (Chapter 11)

Stripe and Adyen are not charities. They integrate a hundred payment methods for three reasons: coverage, cost, conversion. This chapter prices the three separately, and points out which one the industry systematically overstates.

Part V · Are the Card Networks in Danger (Chapters 12–14)

This is the climax, and the only part that requires everything that came before.

Chapter 12 gives only the bad news: India's debit card transaction count fell by two thirds in four years, and Brazil's Pix passed both card types combined on count three years after launch. These are not trends. They are displacement that has already happened.

Chapter 13 explains why none of this shows up in Visa's and Mastercard's financials. There are four reasons, and two of them are moats — credit and dispute handling — neither of which has anything to do with whose technology is better. It also settles another account: the card networks were never on the opposite side from scanning. EMVCo, which sets the QR format, is owned in equal shares by six card networks, and Visa and Mastercard run codes of their own — mVisa, Masterpass QR, BharatQR, UnionPay's code.

Chapter 14 turns the previous thirteen chapters into something that predicts: under what conditions can a country's local method displace cards? We test the model on seven markets: Brazil, India, China, the Netherlands, Poland, the UK, the US. The UK and the US are the counterexamples — they have had fast, cheap transfer pipes for decades and cards were never dented. Only a model that explains the counterexamples is worth using.

The last section lands on a conclusion rather different from "A2A versus cards."

Appendix A: the fourteen chapters wrung into one model. The main line goes country by country; the appendix cuts across horizontally by layer — form factor, scheme, rail, settlement — sets the four layers against eleven markets in one table, and fills in what the main line skipped: Hong Kong (CHATS and FPS), Australia (NPP and its overlay services), and Mexico (on one SPEI, the QR-based CoDi never took and the phone-number DiMo did). To place a local method you have never seen before, go straight to the five questions in the appendix.

4. Chapter Table¶

Previous: (none — this is the starting point) Next: Chapter 1 · The Three Assumptions of a Card: A Terminal, a Credential, and an Interchange Fee