Chapter 5 — Who's Who in Payments: Card Networks, Platforms & Protocols

WhiteBottle Coffee has been open for three months. The line out the door every morning tells you business is good. But then the owner sits down with the first monthly processing statement, and the math stops making sense.

That $5 latte? WhiteBottle didn't keep $5. The statement shows a merchant discount of about 2.4% — roughly 12 cents — split between parties the owner has never heard of. There's a fee labeled "interchange" going to something called the "issuing bank." There's a smaller "network assessment" fee going to Visa. And the payment processor — the company WhiteBottle signed up with to accept cards — takes a cut on top of all that.

Twelve cents doesn't sound like much. But multiply it by a few hundred lattes a day, and suddenly you're looking at a meaningful chunk of revenue flowing to companies you've never met, for services you can't quite see.

So who are these people? Why do so many hands touch a single payment? And how did this system — where a coffee shop in Brooklyn pays a bank in South Dakota every time someone taps their card — come to exist in the first place?

To answer those questions, we need to understand how card networks were built. Not as technology products, but as coordination systems — elaborate agreements among competitors that made universal card acceptance possible. The technology came later. The economics and the rules came first.

The Four-Party Model: How Open-Loop Networks Work

When you tap your card at WhiteBottle, four distinct parties are involved in making that transaction happen. This is called the four-party model, and it's the architecture behind Visa and Mastercard — the two largest payment networks in the world.

Here's who does what.

1. Cardholder: The cardholder, also known as the customer, is you — the person buying coffee. You have a relationship with your bank, which issued you a card.
2. Issuer: The issuer (or issuing bank) is the bank — Chase, HSBC, DBS, whoever gave you the card. The issuer extends you credit (or debit access to your account), approves or declines transactions, and takes on fraud risk. In return, the issuer earns interest on credit balances, annual fees, and — crucially — interchange on every transaction you make.
3. Acquirer: The acquirer (or acquiring bank) is on the merchant's side. This is the company WhiteBottle signed up with to accept card payments. It might be a traditional bank, or it might be a modern payment processor like Stripe, Adyen, or Worldpay. The acquirer's job is to route WhiteBottle's transactions to the card network, settle the funds into WhiteBottle's bank account, and handle disputes. The acquirer earns a fee from the merchant — the merchant discount rate — which bundles interchange, network fees, and the acquirer's own margin.
4. Network: Also known as the rails. The card network sits in the middle. Visa and Mastercard don't issue cards and they don't sign up merchants. What they do is provide the rails — the messaging infrastructure, the rules, the brand, and the dispute arbitration framework that lets any issuer's card work at any acquirer's merchant, anywhere in the world. The network earns fees from both sides: small per-transaction assessments that add up at massive scale.

Table 1: The Four Parties at a Glance. Every card transaction involves this cast of characters — even if the cardholder and merchant never see most of them.

The glue holding this together is interchange — a fee that flows from the acquirer to the issuer on every transaction. Think of it as a balancing mechanism.

Issuers need incentive to put cards in people's wallets (marketing, rewards, fraud protection). Merchants need acquirers to accept those cards. Interchange is the price the merchant side pays the cardholder side to keep the whole system in equilibrium.

If interchange is too high, merchants revolt. If it's too low, issuers stop promoting cards. Every card network walks this tightrope.

Introducing Closed-Loop Payments

Now, not every network works this way.

American Express operates what's called a three-party model (or closed-loop network). In a closed-loop system, the network is the issuer and the acquirer. Amex issues the card directly to you, signs up the merchant directly, and processes the transaction end-to-end. There's no separate interchange — Amex simply charges the merchant a service fee and keeps the economics on both sides.

The advantage of closed-loop? You own all the data. Amex knows who the cardholder is and what the merchant sells, which makes fraud detection and marketing significantly more targeted. The disadvantage? Acceptance is narrower. Visa and Mastercard have millions of independent issuers and acquirers spreading the network to every corner of the world. Amex has to do all that distribution itself.

This distinction — open-loop vs. closed-loop — is one of the most important architectural choices in payments. It shapes everything from fee structures to fraud strategies to global expansion playbooks. And as we'll see later in this chapter, the same open-vs-closed tension shows up again when internet platforms and crypto protocols enter the story.

The four-party model distributes roles across independent players connected by the network. The three-party model consolidates everything under one roof. Dashed lines show fee flows; solid lines show transaction and settlement flows.

How did it all begin?

Well, the four-party model is elegant on paper. But it didn't spring into existence fully formed. It was invented, reinvented, and fought over across decades — and the origin stories of the major card networks are really stories about solving one of the hardest problems in commerce: how do you get millions of strangers to trust a shared system that none of them controls?

Visa: From a Fresno Experiment to a Global Operating System

In 1958, Bank of America tried something audacious — and, in hindsight, a little reckless. The bank mailed 65,000 unsolicited credit cards to residents of Fresno, California. No applications. No credit checks for most recipients. Just cards in envelopes, ready to use.

The idea was simple: if enough people had cards, merchants would accept them. And if enough merchants accepted them, more people would use them. Classic chicken-and-egg, solved by brute force.

The results were predictably chaotic. Delinquency hit 22%. Fraud was rampant — cards stolen from mailboxes, used by people who never applied for them. Bank of America lost millions in the first year. But the underlying bet proved right: once a critical mass of cardholders and merchants existed, the system became self-reinforcing. Distribution creates adoption, but networks survive on risk management and reliable settlement.

An image of the first credit card (BankAmericard).

From 1966 to 1970, Bank of America began licensing the BankAmericard program to other banks, then spun it off into an independent entity. Under the visionary leadership of Dee Hock, the organization became a new kind of institution — owned collectively by its member banks, governed by shared rules, competing fiercely with each other in the market while cooperating on the rails.

The technical foundation came in 1973–74 with two systems that quietly changed everything:

1. BASE I for electronic authorization (replacing phone calls to approve transactions)
2. BASE II for electronic clearing and settlement (replacing paper-based batch processing).

These systems meant a transaction could be approved in seconds and settled between banks within days — at scale.

In 1976, the network rebranded as Visa. By fiscal year 2024, Visa had 4.6 billion credentials at over 150 million merchant locations worldwide. What started as a reckless mailer in Fresno became the largest payment network on earth.

Mastercard: Cooperation Among Competitors

Visa's origin was a single bank that expanded outward. Mastercard's was the opposite: a group of banks that banded together from day one.

In 1966, a consortium of banks — led by efforts in Buffalo, New York — formed the Interbank Card Association explicitly to compete with BankAmericard.

The core problem was clear: individual banks wanted to offer card products to their customers, but no single bank could create universal acceptance on its own. The solution was collective action — shared rules, shared brand, shared rails, individual competition.

The branding evolution tells the story of the network finding its identity: Interbank Card Association (1966) became Master Charge (1969), which became MasterCard (1979). Each rebrand reflected a shift from back-office banking infrastructure to consumer-facing global brand.

Today, Mastercard operates in over 220 countries and territories, supporting more than 150 currencies. But the fundamental design hasn't changed: Mastercard is a coordination layer that lets thousands of competing banks interoperate.

The interchange balancing act is central to how Mastercard (and Visa) think about their ecosystem. As Mastercard has described it in regulatory filings: interchange must be set high enough that issuers find it profitable to promote and support cards, but low enough that merchants continue to accept them. If interchange tips too far in either direction, the network loses one side. Every basis point is a negotiation between the two halves of a two-sided market.

The Closed-Loop Challengers: Amex, Diners, and the Data Advantage

Before Visa and Mastercard built their open-loop empires, the very first general-purpose payment card came from somewhere else entirely.

In 1950, Diners Club launched the world's first multipurpose charge card. The legend goes that founder Frank McNamara was embarrassed when he forgot his wallet at a restaurant — and decided no one should ever face that problem again. Whether the story is entirely true or not, the product was real: a card you could use at multiple restaurants, with a single monthly bill. It was a closed system — Diners Club signed up the restaurants, issued the cards, and handled everything in between.

American Express followed in 1958 with its own charge card, initially focused on travel and entertainment. The key difference from bank-issued credit cards: Amex required full monthly payoff. No revolving credit. This attracted a higher-income customer base and allowed Amex to position itself as a premium brand — a positioning it maintains to this day.

The closed-loop model has a genuine structural advantage: you own both sides of the data. When Amex is both the issuer and the acquirer, it knows who the cardholder is, what they bought, where they bought it, and how often. This makes fraud detection more precise and marketing more targeted. Amex has consistently positioned this data ownership as a competitive moat.

But closed-loop carries a trade-off. Without independent issuers and acquirers spreading the network, acceptance grows more slowly. Visa and Mastercard rely on thousands of banks and processors to evangelize their networks around the world. Amex has to do all that distribution itself. This is why you'll still encounter merchants — particularly smaller ones — who accept Visa and Mastercard but not Amex.

Other networks carved their own niches. JCB, founded in 1961, was Japan's first credit card and remains the dominant domestic brand there, with partnerships extending its reach to 34 million merchants worldwide. Discover, launched in the mid-1980s by Sears, entered as a closed-loop challenger to the bank-card duopoly, innovating on rewards (no annual fee, cashback) to win mass-market American consumers.

The following table lists the main card networks (both open and closed loop) today:

Table 2: Card Network Comparison — Origins and Architecture. Each network solved the coordination problem differently, and those early design choices still shape their competitive positions today.

The Duopoly by the Numbers: Visa and Mastercard in FY2024

Those origin stories — Fresno mailers, bank consortiums, charge-card dinners — might feel like ancient history. But the empires they built are generating staggering numbers right now. Here is how the two open-loop giants looked in their most recent fiscal year:

Table 3: Visa vs Mastercard — FY2024 Financial Comparison. Sources: Visa FY24 Annual Report (Sept 2024), Mastercard FY24 Annual Report (Dec 2024), MacroTrends, Statista, Capital One Shopping Research.

A few things jump out:

1. Both companies operate asset-light, insanely scalable models — they don't lend money, don't hold deposits, and don't take credit risk. They simply move messages and collect fees on every transaction that flows through their rails. That is why net margins above 45% are possible with relatively modest headcounts.
2. Visa leads on absolute scale, profitability, and operational efficiency (revenue per employee). But Mastercard actually generates slightly more revenue per card — $8.92 versus Visa's $8.02 — which suggests a different pricing or volume mix.
3. Core business logic is kept simple. More transactions equals more fees equals more revenue. That means growth comes from two directions — issuing cards to the unbanked (teenagers, developing markets, the 1.4 billion adults worldwide who still lack a bank account) and increasing transaction frequency per existing card (subscriptions, micro-transactions, in-app payments, wallet tokenization).

There is one thread that connects every growth strategy: as long as the transaction runs over the Visa or Mastercard network, the network earns its fee.

We will return to this question in the next few chapters: What happens if merchants start accepting stablecoins or other blockchain-native payment methods that bypass the card rails entirely? If a customer can move money wallet-to-wallet with direct settlement, interchange disappears. Whether that scenario is imminent or decades away is debatable — but it is the structural threat that keeps network executives up at night. We will cover this in the chapters that follow.

What's remarkable about these origin stories is that every one of them — Visa, Mastercard, Amex, Diners, JCB, Discover — was fundamentally solving the same problem: how do you create a payment instrument that works beyond a single bank, a single store, or a single country?

The answers vary — open vs. closed, consortium vs. single-entity, premium vs. mass-market — but the underlying challenge was always coordination.

And the networks that solved coordination at the greatest scale won.

But winning the rails was only the first act. Once the networks existed, a new generation of players started building on top of them — stretching what a card could be and do. That's where we go in Chapter 6: the gateways, processors, and platforms that made cards flexible enough to power everything from ride-hailing apps to one-click checkout.