What this covers

Patrick and Alex trace Visa from its 1958 origins to its current position. Bank of America launched BankAmericard by mailing about 60,000 unsolicited cards to Fresno residents and signing merchants such as Florsheim Shoes, which sidestepped the chicken-and-egg problem. The network then matured for decades inside a bank-owned nonprofit consortium, free of competition. Patrick frames Visa as an information router for banks rather than a finance company. On the economics, most of the roughly 3% a merchant pays goes to the issuing bank as interchange, and the network itself keeps only about seven and a half cents per transaction.

The middle of the conversation covers how interchange pays for consumer rewards, and why competition between networks pushes merchant fees up rather than down, since networks court issuers with richer rewards. Target's 5% REDcard shows how costly it is for a merchant to dodge the system. The open-loop vs closed-loop networks distinction explains why Visa cannot easily change the consumer experience the way American Express can, and why protocol changes such as T+2 settlement or the 20-character merchant name limit take a decade.

The more contested turns are policy and power. Australia's interchange cap is presented as a case of concentrated benefit, diffuse harm: large merchants gain while consumers lose rewards. Other topics include the 2006 IPOs as an antitrust maneuver, Visa's enforcement of sanctions after Crimea, Mexico as the endpoint of issuer-acquirer consolidation, and the failure of Visa Checkout.

Sharpest takeaway

Visa is a consortium-born protocol router whose moat comes from network effects and banks' dependence on it, and its main threats are issuer concentration, regulation, and geopolitical control of networks rather than fintech challengers.

  • Visa earns a small per-transaction fee while issuing banks capture most interchange, so its economics depend on keeping issuer relationships
  • Fragmentation across many banks is what makes a central router valuable, so concentration of issuing and acquiring is the biggest risk
  • Open-loop networks are hard to change or disrupt because changing the shared protocol requires agreement from all participants

The argument · threads12 threads · 79 claims
0.78

Visa's shared open protocol and network effects make it a durable, hard-to-replicate router of commerce.

10 pointscentrality 5/5
  • Business models like Visa's, with a shared open protocol among many competitors, are rare because normal network effect businesses centralize both the client experience and the economics, so few companies can build a protocol with pricing power.

    “it's always the centralized player that controls the experience and controls the economics”

  • Visa's main growth driver is the ongoing shift from cash to digital payments, including debit, and the pandemic accelerated that shift, which has been overshadowing other problems.

    “the eradication of cash. It is this bridging transaction.”

  • Visa's product development mostly consists of card tiers and benefits like wine tastings, purchase protection, and price-drop refunds, because deeper changes to the protocol are hard and changes that compete with issuers would destroy Visa's distribution.

    “There's not that much that they can really build or do.”

  • Visa was birthed by Bank of America but now has a far higher market cap because it is the central router where all commerce flows, while Bank of America is only one endpoint participant.

    “it's one of these weird cases of, it's almost like patricide”

  • Visa's management is good, but the network effect is so strong that long-term structural risks such as geopolitics and fragmentation versus concentration matter more to investors than fintech challengers' press releases.

    “it's the reputation in the market that remains intact. It's the opposite here.”

  • Networks should farm early-stage fintech startups, as Stripe did by signing up Y Combinator companies with insignificant volume, because today's two-payments-a-year startup can become a major processor like Instacart, and capturing them early creates a lasting volume advantage.

    “that's actually probably the farming motion that both Visa and MasterCard have to get used to”

  • Network businesses often start with a comically small base, as Facebook began with Harvard students, and Visa followed the same path by dominating Fresno before expanding.

    “Facebook started off with the comically small example of Harvard students.”

  • A new payment network must solve a real problem for both consumers and merchants, since solving only the merchant side by cutting fees does not give consumers a reason to adopt it, and buying consumer adoption with rewards eats the savings.

    “You have a solution to one side of the network, but just not the other side.”

  • Visa stays valuable in the long tail of bank pairs that would never connect directly, such as a Colombian bank and a Jordanian bank, because building direct links for rare transactions is too costly, so networks persist even if big banks route their own traffic.

    “But if it's the long tail of like, 'Hey, I'm visiting Jordan from Colombia,' Those two banks, Banco de Colombia isn't going to go connect with the Bank of Jordan directly, too much work.”

  • A Buffett-style principle holds that when a brilliant management team meets a disaster-prone market, the market reputation stays intact, but Visa is the reverse: management is good, yet the network effect is so strong it is hard to disrupt.

    “when a management team known for brilliant meets a market known for disaster, it's the reputation in the market that remains intact.”

0.64

Bank routing leverage and dependence on issuing banks constrain Visa's control, pricing, and product choices.

10 pointscentrality 4/5
  • Concentration of issuing relationships is the biggest risk factor for Visa, because a major issuer such as Citi or Chase could switch networks and cause a large drop in Visa or MasterCard stock.

    “It's the biggest risk factor, and the reason why”

  • In Mexico almost no transactions run over Visa because acquiring and issuing banks are largely the same entity, showing the endpoint of issuer-acquirer consolidation.

    “almost no transactions in Mexico go over Visa”

  • Chase renewed with Visa but is also the largest online acquirer, so it asked to stop paying Visa for transactions it routes to its own cards, which shows how an issuer with acquiring can bypass the network.

    “Why the heck am I Chase paying Visa to talk to myself?”

  • Because Visa and MasterCard are open-loop networks that do not control the customer relationship, they cannot easily improve the consumer experience: a change to interchange can cause issuers to switch networks, so product changes are constrained.

    “So it actually makes improving the product and improving the experience so much harder”

  • When regulators squeeze interchange, Visa's own fee is only mildly affected because it bears no credit risk, while most of the loss falls on issuing banks, which earn most of the economics.

    “normally Visa's fee doesn't get squeezed as much because they're not taking any credit risk”

  • Protocol-level improvements are slow not because they are technically hard but because every participating bank must agree, so if five large parties disagree nothing changes for a decade, as with T+2 equity settlement and the 20-character merchant name limit.

    “it's not a hard tech problem, it's just you got to get everybody to agree”

  • Visa's pricing committee cannot easily cut rates for trusted e-commerce merchants, because issuers that fund rewards from those rates would threaten to switch to MasterCard, so product improvements that lower merchant costs are stalled.

    “if Visa says, 'Oh, it's a good point. We'll cut interchange for ecommerce transactions for trusted consumers to 1.7%.' That's going to really off the issuing banks”

  • When a bank both issues cards and acquires merchants, it can route transactions internally as On-Us transactions without paying the network, so a bank like Chase could avoid fees by threatening to route its own transactions in-house, which pressures networks to renew at lower rates.

    “I don't want to pay you anything, if I'm talking to myself,' which is a pretty reasonable request”

  • Visa cannot build a consumer-facing product that competes with issuing banks because issuers are its distribution channel, and it has no direct customer relationship, so it can only offer tools that issuers choose to roll out.

    “we don't want to compete with the issuing banks because the issuing banks are how we get our business.”

  • Visa, MasterCard, and American Express form an oligopoly where competition exists, but it is not a monopoly with only one player.

    “Visa and MasterCard, American Express, like it's an oligopoly, but it's not just a monopoly”

0.57

Merchants ultimately pay interchange that funds richer rewards, and fees ratchet upward as networks compete.

13 pointscentrality 3/5
  • Interchange is not one rate: it varies by card type, international status, merchant category, and card-present versus card-not-present, with debit usually far below 3% unless issued by a bank with under $10 billion in assets.

    “So there are so many different rates”

  • Of the roughly 3% a merchant pays per card transaction, the vast majority goes to the issuing bank as interchange, a smaller share to the acquiring bank, and the network itself keeps only about seven and a half cents per transaction on average.

    “the vast majority goes to the issuer.”

  • Visa and MasterCard compete primarily for issuing banks, because acquirers must process all cards and merchants cannot be excluded, so card networks try to win issuers by offering higher interchange that funds richer rewards.

    “Visa and MasterCard compete over issued contracts.”

  • Issuing banks use the interchange they collect to fund consumer rewards such as cash back, so merchants effectively pay for the rewards that induce customers to use the card and avoid cash.

    “the issuer uses most of those economics to then, I don't want to say bribe”

  • Merchants cannot easily avoid card interchange because rewards cards pull customers toward using cards at their stores; a merchant that pushes cash or alternative payments must offer enough incentive to overcome the customer's reward, so the cost stays embedded.

    “how do they induce me to sign up for a new payment mechanism or go pay them with cash?”

  • Unlike ordinary markets, competition between card networks raises merchant fees rather than lowering them outside regulated regions, because issuers can use higher interchange to fund better rewards that win their customers.

    “outside of regions like Europe or Australia where regulatory issues have forced down fees, fees go up”

  • Visa's fees can be framed as a tax on commerce or as an enabler, and Visa justifies higher card-not-present pricing through value-based pricing, based on how much value the internet economy gains from the network.

    “It's just classic value based pricing.”

  • Visa and MasterCard have a double-sided moat: the classic network effect of merchants and consumers both needing the card, plus an incentive loop where interchange paid by merchants funds rewards that keep consumers using the card at the merchant.

    “It's not just a network effect in that everybody uses one of these cards because everybody else accepts that card.”

  • Fees rise over time rather than falling under competition, because a higher interchange rate lets issuers offer richer cash back that wins consumers, and merchants have no way to refuse the network, so the rate ratchets up.

    “fees go up because that is more compelling for the issuing banks to go get their cards into market”

  • In the early BankAmericard program, the merchant rate was reportedly around 6 percent, and the program's fraud problems came from issuing cards without credit checks.

    “I think the rate was something like 6%.”

  • Visa Signature cards carry benefits such as travel perks, purchase protection, and price protection, and Visa sells these features to issuers much as enterprise software is sold against competitors, with checkboxes that issuers compare at renewal.

    “it's almost like classic enterprise sales where I'm a bank, let's say I'm Barclays”

  • Europe capped interchange at about half of Australia's rate, with exemptions for some small business cards, showing that regulators in many regions have concluded interchange is too high.

    “Europe did the same thing, but even lower, half the rate of Australia”

  • Visa Signature cards carry a higher interchange rate, and that premium is used to incentivize banks to issue the card because they keep more of the money.

    “the higher interchange rate is actually used to incent banks to issue that card because then they get to keep more money.”

0.53

Visa is a bank-owned shared protocol that routes transactions among all banks.

5 pointscentrality 4/5
  • In the 1960s and 1970s, two bank consortia formed: BankAmericard became Visa, and Master Charge became MasterCard. Both were nonprofits owned by member banks, so they operated as shared protocols, not profit centers.

    “there was Master Charge, which was one. And then the BankAmericard became Visa.”

  • When a card is swiped, the acquiring bank reads the card's leading digits: a prefix of 4 routes to Visa and 5 to MasterCard, and Visa then reads the BIN (first six digits) to identify the issuing bank that must approve the transaction.

    “this card starts with a four, that means it's a Visa”

  • A card transaction involves five parties: the consumer, the merchant, the issuing bank that owns the consumer, the acquiring bank that owns the merchant, and the network (Visa or MasterCard) that routes messages between them, which neither merchants nor consumers directly deal with.

    “There are actually five parties in a transaction.”

  • Visa's moat was built slowly over decades as a bank-owned nonprofit consortium that operated the router for all banks, so the network effect matured without competition before the company went public.

    “it was a communication protocol”

  • China UnionPay is the largest domestic network in China and interoperates with Visa, so foreign-issued transactions route over Visa while domestic ones route over UnionPay.

    “The largest payment network in China... is called China Union Pay”

0.46

Fintechs and phone-based permissioning could bypass card networks by controlling issuing and default placement.

8 pointscentrality 3/5
  • Companies like Uber or Lyft may issue cards to their drivers to replace their banks, because controlling financial infrastructure lets them retain drivers and monetize them more deeply.

    “There are all sorts of benefits to actually controlling the financial infrastructure.”

  • Chase should have dominated online merchant acquiring by offering lower rates through its own issued cards, but it did not, while Stripe could do so because it is a tech-first company with better merchant relationships.

    “Chase honestly, should have run the table on online, acquiring merchants”

  • Plaid is a data aggregation network between financial services companies and does not compete with Visa or MasterCard because it does not handle transactional payments, so it is analogous to Visa only for non-commerce banking data.

    “They don't really compete with Visa or MasterCard at all.”

  • Moving cards into phone wallets threatens the default-card advantage that issuers enjoy, because software can reorder cards and permission new financial products instantly, changing which issuer a consumer uses by default.

    “It's the power of the default. That goes away because now it's software.”

  • Stripe could eventually become both a major issuer and acquirer and self-clear transactions, bypassing Visa's rails, though Visa would tolerate this because customer cards still route through Visa for many merchants.

    “they could self-clear and self-settle, and they don't have to pass anything through Visa”

  • Merchants cannot send itemized purchase data back through the card chain, such as Walmart line items appearing on a card statement, because the legacy protocol has no field for it and changing the protocol requires all parties to agree.

    “there's literally technologically no way for Walmart to send back to First Data, to send back to Visa”

  • Financial products will be discovered and activated inside phone app stores through permission prompts, as apps are now, which replaces junk-mail card solicitations and lets consumers add a new card instantly, making card loyalty far more fluid.

    “instead of getting junk mail to go open up a new credit card, just there's an app store for apps that use my phone”

  • Phone-based permissioning and direct connections from Apple devices to payment terminals could allow financial products to be adopted and routed without traditional card networks, though the timing is uncertain.

    “there actually is a way of potentially bypassing these networks”

0.41

Interchange caps shift value to merchants, removing consumer rewards and leaving consumers worse off.

6 pointscentrality 3/5
  • Interchange caps produce concentrated benefits for large merchants like Walmart and diffuse harms for consumers, which makes the policy hard to reverse or rebalance, because each consumer barely notices a tiny per-transaction saving.

    “it's very, very hard to fix these problems of concentrated benefit and diffuse harm”

  • Many countries are experimenting with regulated interchange because it is better for merchants, but this is largely true only for merchants, and it ends up worse for consumers than expected, since rewards vanish when fees are capped.

    “it ends up being a little bit worse for consumers, which I think was unexpected”

  • Australia capped interchange at under 50 basis points, which removed rewards and transferred value from issuers to large merchants, leaving consumers worse off and annual fees rising.

    “it turned out that it's bad for consumers because the rewards went away”

  • Target's 5% REDcard rewards show how expensive consumer incentives become when a merchant tries to avoid network fees, because Target pays more to reward customers than the 2% interchange it would otherwise pay.

    “Target has something called the Target REDcard, and you get 5% cash back”

  • The Durbin Amendment caps debit interchange for banks over $10 billion in assets at about 5 basis points plus 21 cents, which backfires for quick-service restaurants because low-ticket debit purchases become relatively more expensive than unregulated credit transactions.

    “So ironically, the non-regulated form is cheaper for them.”

  • Walmart-backed MCX failed as a competing payment system because merchants wanted zero fees while consumers needed rewards to adopt a new payment method, and zero interchange removes the rewards that drive adoption.

    “MCX was a complete disaster.”

0.41

Centralizing the routing hub let one owner capture the economics and charge tolls.

6 pointscentrality 3/5
  • Had Visa been designed from scratch as a decentralized ledger hosted by all banks, no single party could capture the router's economics, so the centralized hub that became a for-profit toll-collector would be prevented.

    “all the banks are nodes that hosts the ledger. Nobody can control it”

  • Joe Nocera's book A Piece of the Action, about Visa's history, is the best resource for studying Visa, and Dee Hock's book One From Many is also recommended.

    “There's a great book called A Piece of the Action.”

  • Visa was originally designed as a chaordic system that no single person could control, but a lawyer converted it from nonprofit to for-profit, which concentrated the central routing data and let the owner charge a toll.

    “Visa was meant to be this chaordic thing that nobody could control, but a lawyer figured out how to rejigger that”

  • Twitter began as a protocol with many third-party clients such as TweetDeck, but it moved to control the client experience and charge for protocol access, as a company that centralizes the router typically controls both ends of the experience.

    “Twitter realized, 'Wait a minute, we can't serve as effectively... we want to control the client experience on both sides.'”

  • BATS, the Better Alternative Trading System, grew because high-frequency traders who owned equity in the exchange routed their flow there, a rare model that resembles Visa's bank-owned structure, but such playbooks are hard to run.

    “if you looked at BATS, the Better Alternative Trading System, that was a company where they got started because they were able to get more high frequency flow”

  • Open-loop networks such as Visa and MasterCard, where many banks own the network and its constituents, are rare and only feasible when the constituents are included from the start, whereas closed-loop networks like American Express control both the merchant and customer relationships.

    “It's a very rare business model which you'll only see that happen if you get the constituents from day one.”

0.25

Visa's simple, capital-light business mainly returns cash to shareholders through buybacks.

5 pointscentrality 2/5
  • Visa has limited reinvestment opportunities and mainly returns capital through buying back its own stock, which Alex thinks is a reasonable outcome given that it doesn't weaken the moat.

    “besides buying back their own stock or something”

  • Cross-border transactions are Visa's most profitable segment, so declines in travel hurt Visa disproportionately even when overall volume grows.

    “That's their most profitable segment.”

  • Visa's operating business is simple enough that it needs only about ten people to keep the lights on, and its networks rarely go down, unlike newer tech companies that have frequent outages.

    “They need probably 10 people to keep the lights on.”

  • Visa's potential growth lies in signing up early-stage non-bank companies that may become large issuers or acquirers later, because the zero-sum fight over bank renewals is less important than the expansion of digital money.

    “That's probably the farming motion that both Visa and MasterCard have to get used to”

  • Visa and MasterCard went public around 2006 mainly to avoid antitrust scrutiny over bank-coordinated pricing, not to seize a growth opportunity, and the IPOs were largely secondary sales of bank stakes rather than new capital.

    “they went public, not because it's like, 'Wow, let's seize this opportunity”

0.21

Early card launches seeded cardholders before recruiting merchants to solve the chicken-and-egg problem.

3 pointscentrality 2/5
  • The unsolicited card drops of the 1950s produced heavy fraud and credit losses because people received credit without applying, which led regulators to outlaw such mailings and require clear disclosure.

    “fraud was out of control. Joe Williams basically got fired”

  • Bank of America's 1958 BankAmericard launch mailed about 60,000 unsolicited cards to Fresno residents and signed up merchants such as Florsheim Shoes, solving the chicken-and-egg problem by giving cardholders a reason to use cards before merchants were recruited.

    “Okay, let's just go mail 60,000 cards to pretty much everybody in Fresno.”

  • Companies often do things that make strategic sense but lack market demand, so a product should be built only if it addresses a real customer need, as Google's social network after Facebook's entry illustrates.

    “you should build something because, A, you should strategically have it, sure, only if it's key. The market actually desires it.”

0.21

Merchants resent interchange as coordinated price-setting, yet litigation settlements remain manageable for Visa.

3 pointscentrality 2/5
  • The consortium's pricing committee set interchange rates that always seemed to rise each year, which merchants resented and regulators disliked, functioning as price coordination that resembled an airline alliance fixing fares.

    “that's called price fixing, you can't do that”

  • Interchange fees are large relative to thin retail margins, so Target's after-tax profit would roughly double if interchange were eliminated, which is why merchants spend heavily on litigation and lobbying against Visa and MasterCard.

    “if they were able to eliminate interchange, their after tax profits would double”

  • Merchant litigation against MasterCard and Visa has been extensive, with one settlement near $10 billion that the market rewarded with a stock increase, suggesting investors see such settlements as manageable.

    “They settled one of their last lawsuits $10 billion and the stock went up.”

0.21

Card networks face geopolitical risk as nations push toward separate national payment networks.

3 pointscentrality 2/5
  • Countries historically controlled money because they printed legal tender, but legal tender now moves over a network controlled by a private entity that can be influenced by governments outside a country's control.

    “the legal tender now happens over a network”

  • After Russia's annexation of Crimea, US sanctions were enforced by Visa rerouting or rejecting cards, showing that a private US-based network can cut off a country's commerce, which makes card networks a geopolitical risk.

    “they could just change that routing and say, 'Nope, doesn't happen.'”

  • Payment networks are becoming the new strategic petroleum reserve: countries that depend on a privately controlled network are vulnerable to geopolitical cutoff, so they may need to build separate national networks, as Russia, France, and the UK are considering.

    “networks, they are the new strategic petroleum reserve”

0.09

New payment products fail unless they first solve a narrow, concrete consumer problem.

4 pointscentrality 1/5
  • Visa Checkout failed despite heavy merchant spending because it did not solve a fundamental consumer need, and PayPal had already solved the problem about twenty years earlier.

    “But again, it just didn't solve a fundamental consumer need”

  • Crypto is unlikely to displace Visa unless it solves a real consumer problem, and its volatility makes deflationary tokens hard to use as spending currency, so decentralization is an advantage only in principle.

    “it's very, very hard for people to want to spend a deflationary currency”

  • Google Checkout launched in 2006 as a broad wallet and checkout product without a narrow problem to solve, and it was a complete disaster despite being built by smart people.

    “They launched Google checkout in 2006, complete disaster.”

  • Network businesses should start with a narrow, well-defined set of constituents who have clear product-market fit and then expand outward, because trying to serve everyone at once prevents the network effect from forming, as with Google Checkout.

    “it's narrow your scope and expand concentrically is the lesson for me”