Daniel's been staring at a card reader beep again and wondering what's actually behind it. His question this week: why is the global card system built the way it is, a few giant networks like Visa and Mastercard sitting on top of a layer of local issuers, and why are the Israeli issuers so aggressively mediocre. Support ends at four in the afternoon. No overnight coverage. Almost no rewards worth the name. And the bank and the card company can each point at the other forever because they are, in fact, two different companies. He's asking whether the networks enforce real standards or if any minimally compliant outfit can hang a Visa logo on a piece of plastic. What actually stops a disruptive entrant in a market like Israel. And whether the split between bank and card issuer is universal, just sometimes hidden.
The four PM cutoff is the detail that should make everyone stop and think, because it is not an accident of bad management. It is the visible edge of a structure.
The beep hides a stack. Most people swipe, the machine says approved, and that is the entire transaction as far as they know.
Right, and the stack underneath is a four-party model. Cardholder, issuer, merchant, acquirer, with the network in the middle. The issuer is the bank or card company whose name is on the card. The acquirer is the company that handles the merchant's side, the store's bank effectively. Visa and Mastercard are neither of those. They do not issue cards, they do not hold balances, they do not lend money. They run the rails and set the rulebook.
So when Daniel calls about a disputed charge at four oh five and gets told to call back tomorrow, he is not being failed by Visa. Visa does not have a customer service line for cardholders in any meaningful sense.
He is being failed by the issuer, which in Israel is often a separate legal entity from the bank where his salary lands. And the acquirer is a fourth party he has never spoken to in his life. So the blame-shifting is not a bug in the culture. It is the architecture. Each layer can truthfully say the problem lives in a different layer.
And the network is invisible by design. It wants the transaction to feel like it happens between you and the merchant. That is the brand promise. The moment you notice there are four companies involved, the magic thins out.
Let's start with what the networks actually enforce, because Daniel asked whether it is real standards or minimal compliance. It is a thick rulebook. EMV chip specifications, tokenization, dispute timelines, chargeback windows, data security rules. If an issuer wants to put a Visa badge on a card, it has to meet those. It is not a rubber stamp.
But notice what is not in the rulebook. Customer support hours. Whether a human being answers the phone after four PM. Whether the dispute process is comprehensible.
The standards are about network integrity and transaction security. A fraudulent transaction that gets processed badly hurts the network's reputation with merchants and acquirers. A cardholder who waits on hold for forty minutes does not hurt Visa at all, because Visa earns its revenue from interchange on every transaction regardless of whether that cardholder is happy.
Interchange is the cut the network and the issuer split on every swipe. The network's incentive is volume. A bad issuer that processes a lot of transactions is still a profitable issuer for the network. A good issuer that processes fewer transactions is worth less.
That is the core of why service quality is not enforced. It is not that the networks are lazy. It is that their revenue model does not include a line item for your satisfaction.
So the licensing question. Could any minimally compliant outfit become an issuer?
No, and this is the misconception worth killing. Becoming an issuer requires a banking license or a partnership with a licensed bank, plus significant capital, plus the compliance infrastructure to handle fraud, anti-money laundering, dispute resolution. The network rules are necessary but not sufficient. You cannot just fill out a form and get a Visa badge.
The regulatory barrier is the real wall. In most countries, getting a banking license takes years and a pile of capital. That is why fintechs almost never become issuers from scratch. They partner with an existing bank and rent the bank's license.
And that is the trap Daniel is sensing. If a fintech wants to disrupt the Israeli card market, it has to either become a bank, which is enormous, or partner with a bank, which means it inherits the same infrastructure, the same settlement systems, the same four PM batch processing, and the same incentives.
The batch processing. That is the mechanical reason the four PM cutoff exists. The settlement batch runs in the late afternoon. A dispute logged after the batch cannot be processed until the next cycle. So the issuer just tells customers to call back tomorrow.
That is not a customer service policy. That is a back-office reality that got promoted into a customer-facing policy because nobody had an incentive to fix it.
Israel's market is a perfect case study because the regulator actually tried to fix the concentration problem. The Strum Law in twenty seventeen forced the big banks to sell their credit card companies. Leumi Card became Max. Isracard was spun off.
And the result was a transfer of ownership, not a change in service. Leumi Card is now Max, owned by the private equity firm Warburg Pincus. The card company is no longer owned by the bank, but it still relies on banks for distribution and funding. The customer-facing experience did not improve, because the law targeted market concentration, not service quality.
The regulator assumed more players meant more competition meant better service. But the new players compete on price, interest rates and fees, not on whether a human answers the phone at nine PM.
Switching costs are high. A consumer with a card tied to direct debits, recurring payments, and a decade of credit history does not switch issuers because the support line closes early. So the issuers do not get punished for bad service. The market does not reward fixing the four PM cutoff.
Compare that to the US market, where there are thousands of issuers. Credit unions, regional banks, national players. Chase Sapphire built an entire brand on customer service and travel perks. Because in a market that size, service is a differentiator that wins customers.
Israel has a handful of issuers in a market of nine million people. The economics do not support a dozen competitors fighting on service quality. So you get an oligopoly that competes on price and treats service as a cost to be minimized.
Canada has a similar problem. A few big banks dominate the card market, and the customer experience is famously mediocre compared to the US. Small market, few players, same dynamic.
The network's neutrality is the knock-on effect that matters. Visa and Mastercard benefit from any issuer, good or bad, as long as transactions flow. They have no financial incentive to enforce service quality. And the local regulator, the Bank of Israel, focuses on financial stability, not user experience. So nobody in the chain is accountable for the four PM cutoff.
The customer is the only one who cares, and the customer is the only party with no leverage.
Now Daniel's last question, whether the bank issuer split is universal. It is structural in the four-party model. The issuer must be a regulated financial institution. Sometimes the issuer is the same bank where you hold your checking account. Sometimes it is a separate card company. In co-branded cards, the bank is hidden behind the airline or hotel brand on the front of the card.
So the separation always exists, but its visibility varies. In Israel it is painfully visible because the card company is a separate entity you must call, and the bank is a different entity that also must be called, and neither can resolve anything without the other.
In the US, if you have a Chase card and a Chase checking account, the separation is invisible. One phone number. One app. The structural split is still there, the issuer and the bank are technically different legal functions, but the customer never sees the seam.
The co-branded card is the most abstracted version. You call Delta about your Delta Amex, and you are actually calling American Express, but the airline brand is what you see. The bank is completely hidden.
So the answer to Daniel's question is that the separation is universal but the visibility is a market-specific choice. Israel made it explicit, and then made it worse by not giving either side the tools to actually resolve customer problems.
What would a disruptor actually need to do to fix this? Bypass the four-party model entirely. Crypto, direct debit, real-time payment rails. That is a different episode. Any entrant that stays inside the four-party model inherits the same structure and the same incentives.
Open banking and real-time payments are the interesting pressure point. If a consumer can push a payment directly from their bank account to a merchant in real time, the card network becomes optional. The four-party model's grip loosens.
But for now, the beep still hides the same old stack. And the stack is very good at protecting itself.
The Strum Law is the perfect illustration of how hard this is to change. The regulator identified a real problem, concentration, and forced a structural change, the sale of the card companies. And the result was new owners with the same old four PM cutoff. Because the law did not change the underlying economics.
The new owners, private equity, have an even stronger incentive to cut costs. Warburg Pincus did not buy Max to run a customer service hotline. They bought it to extract value from the existing transaction flow.
And the transaction flow does not care about customer service. It cares about settlement batches, interchange rates, and fraud losses. Those are the metrics that show up in the boardroom.
So when Daniel calls Max and gets bounced to his bank, and the bank bounces him back to Max, he is experiencing the architecture exactly as designed. Each party is telling the truth about their own limitations, and the truth is that neither party has the authority to fix the other's systems.
The networks could fix this. Visa could write a rule that says issuers must provide dispute resolution within twenty-four hours. It has the power. It just has no incentive, because that rule would cost issuers money and the issuers are the network's customers.
The issuers are the network's customers. That is the inversion everyone misses. The cardholder thinks they are Visa's customer. They are not. The issuer is Visa's customer. The merchant is the acquirer's customer. The cardholder is the product being sold to both sides.
That is the four-party model in one sentence. The cardholder is the product. The networks sell access to cardholders to merchants, and they sell the right to issue cards to banks. The cardholder is the asset being traded.
And the asset has no voice in the transaction. The asset just wants to dispute a charge at four oh five and is told to call back tomorrow.
The dispute process itself is a window into the structure. When you dispute a charge, the issuer files a chargeback with the network, which forwards it to the acquirer, which forwards it to the merchant. Four parties, each with their own deadline and their own paperwork. If any one of them drops the ball, the dispute dies.
And in Israel, the issuer and the bank have separate IT systems that do not talk to each other. So a dispute that involves both the card account and the bank account requires manual reconciliation. That is why neither can resolve it without the other.
The Strum Law actually made this worse in the short term. When Leumi Card was sold and became Max, the IT systems that had been integrated with the bank had to be separated. For years, the two systems were held together with duct tape and manual processes.
So the regulator created a structural solution that required a technological integration that nobody had budgeted for, and the customer experienced the gap as worse service.
The global pattern is that markets with strong competition differentiate on service, and markets with weak competition differentiate on nothing at all. Israel is in the second group. The US is in the first. The UK is somewhere in between, with challenger banks like Monzo and Starling pushing the incumbents on service.
But even Monzo and Starling are not issuers from scratch. They are banks that partner with the networks and build their own customer experience on top of the same rails. They fixed the front end, not the structure.
That is the realistic path for a disruptor. Not to bypass the four-party model, but to build a better front end on top of it. And the front end is where the four PM cutoff lives. It is a policy decision, not a structural requirement.
The batch runs at four thirty, so the support line closes at four. That is a choice. The batch could run at midnight. The support line could stay open until nine. The cost is real but small. The reason it does not happen is that no one is rewarded for doing it.
In the US, a credit union with ten thousand members will have a support line open on weekends. Because those members can leave. The credit union's survival depends on service. In Israel, the issuers know you are not leaving, because the alternatives are just as bad.
The alternatives are just as bad because they all face the same economics. Small market, high switching costs, no regulatory pressure on service. It is a stable equilibrium of mediocrity.
And the networks are fine with it. Visa's revenue from Israel is a rounding error in its global business. The network has no reason to invest in fixing a small market's service quality.
The Tel Aviv basement detail. I keep coming back to how thin the network's local presence actually is. Visa does not have a regional office with a team of people who could enforce standards. It has a sales office that manages relationships with the issuers.
The network's local presence is a sales function, not a regulatory function. The people in that office are there to sign up more issuers and more merchants, not to audit service quality.
So the four PM cutoff is not a violation of any rule. It is simply the absence of a rule. And the absence of a rule is the network's choice.
Daniel asked whether Visa and Mastercard enforce common standards. The answer is yes, for technical and security standards. No, for anything that touches the customer experience. And the split is not an oversight. It is the business model.
The technical standards are what make the system work globally. A card issued in Israel works in a terminal in Japan because the network enforced a common standard for the chip and the data format. That is a genuine achievement.
And the customer service standards are what would make the system pleasant, and they are left to local competition, which in small markets is too weak to enforce anything.
The Strum Law was an attempt to create that competition by fiat. It failed because competition cannot be created by restructuring ownership. It has to be created by changing incentives.
The incentive that would actually work is portability. If a customer could move their card account to a new issuer as easily as they move their phone number, the issuers would suddenly care about service. But card portability is not a thing, because the card number is tied to the issuer.
The card number is the lock-in. Your card number is issued by the issuer, and changing issuers means changing your card number, which means updating every recurring payment. That friction is the moat.
And the networks could fix that too. Tokenization could theoretically make the card number portable. But the networks have no incentive to make it easy to switch issuers, because switching costs keep the issuers happy, and the issuers are the network's customers.
So the system is a stack of incentives that all point toward the status quo. The networks want volume, the issuers want lock-in, the acquirers want merchant fees, and the regulator wants stability. Nobody wants to make the customer's life better, because nobody is paid to.
The customer is the only one who wants it, and the customer has the least power. That is the four-party model in practice.
Daniel's frustration is the rational response to a system that is working exactly as designed. The four PM cutoff is not a bug. It is a feature.
A feature that no one would ever design from scratch, but that emerged from the interaction of the four parties' incentives over decades.
And the Strum Law shows that even a well-intentioned regulator cannot fix it by moving the pieces around. The pieces are the problem.
The open question is whether something outside the four-party model will eventually make it irrelevant. Real-time payments are the most credible threat. If you can pay a merchant directly from your bank account in real time, with no card network in the middle, the four-party model loses its reason to exist.
But real-time payments are still a bank product, and banks are not known for fixing customer service either. The four PM cutoff could migrate to the real-time payment system.
That is the pessimistic view. The optimistic view is that real-time payments lower the barrier to entry for new players, because they do not require the card network's rails. A fintech could build a payment app on top of real-time rails without becoming an issuer.
And then the customer service problem becomes a competitive differentiator again, because the switching cost is lower. You can change payment apps without changing your bank account.
That is the scenario where the four PM cutoff finally dies. Not because the regulator fixed it, but because the customer could leave.
Until then, the beep hides the stack, and the stack hides the four PM cutoff, and the four PM cutoff hides a settlement batch that no one has an incentive to move.
Hilbert: I worked as a card services liaison for an Israeli processor in the early two thousands. Small outfit, long gone. My job was to tell customers why their problem could not be solved today. The four PM cutoff was my script.
So you were the human face of the settlement batch.
Hilbert: I was the voice on the line saying the system is closed. The system was not closed. The batch ran at four thirty. Anything logged after four could not be processed until the next morning. So we told people to call back. It was easier than explaining the batch.
The batch became a customer service policy.
Hilbert: The batch was the policy. We did not have a policy. We had a batch schedule and a phone script. The bank and the card company had separate IT systems that did not talk to each other. A dispute that touched both systems required a person at each end to manually reconcile the records. After four, the reconciliation person went home.
So the blame-shifting was not just structural. It was also a staffing schedule.
Hilbert: The bank blamed the card company because the card company's system was closed. The card company blamed the bank because the bank's records did not match. Both were right. Neither could fix it alone. The network was a fax number we never called.
A fax number.
Hilbert: Visa's regional office was a single person in a Tel Aviv basement with a flip phone. I tried to escalate a complaint once. I called the number on the fax cover sheet. A man answered, said he was the regional office, and told me to fax the details. I asked if there was anyone else. He said he was the anyone else.
The network's local presence was literally one person.
Hilbert: One person, one flip phone, one fax machine. And he was not there to fix customer problems. He was there to sign up more issuers. He told me that himself. He said, I do not handle disputes, I handle contracts.
The network's local presence was a salesperson with a fax machine.
Hilbert: That is the whole story of the four-party model in one basement. The network is a brand. The brand has a fax number. The fax number goes to a man who sells contracts. The man does not handle disputes. The dispute is your problem.
The four PM cutoff was the same. A batch schedule that became a phone script that became a customer service policy.
Hilbert: The batch ran at four thirty because the settlement system was built in the eighties and nobody wanted to pay to rebuild it. The four PM cutoff was not a cost-cutting measure. It was the sound of a legacy system's working hours.
The customer heard a policy. The company heard a schedule. The network heard nothing.
Hilbert: The network heard the fax machine ring once a month. And the man in the basement only answered it if he was not on a sales call.
That is the thinnest possible local presence for a global network. One person in a basement with a flip phone.
Hilbert: He was not even full time. He split his week between the Visa office and the Mastercard office. Same basement, different desk.
The two networks shared a basement.
Hilbert: They shared a fax machine too. The man had two flip phones, one for each network. He would answer the wrong one sometimes and have to check which network he was.
That is the global payment system in a single image.
Hilbert: My brother-in-law is in the trade now. He runs a payment processing company in Cyprus. He says the same thing. The networks are brands, the issuers are banks, and the customer is the only one who thinks the system is supposed to work for them. He also says the four PM cutoff still exists in half the markets he works with. Not because of the batch, but because the support staff goes home and nobody wants to pay for a night shift.
Even where the batch has been fixed, the cutoff persists as a staffing decision.
Hilbert: The batch is fixed in most places now. The cutoff is just a habit. And my brother-in-law is not to be trusted, he once told me he could process a refund in real time and it took three days.
The habit of the cutoff is the hard part. The structure can change and the behavior stays.
Hilbert: The behavior stays because the incentive stays. Nobody gets fired for closing the support line at four. Somebody might get fired for proposing a night shift that costs money and does not increase transaction volume.
The transaction volume is the only metric that matters. The support line is a cost center. The batch is a revenue center.
Hilbert: That is the whole thing. The support line is where the customer calls. The batch is where the money moves. The money moving is the business. The customer calling is the overhead.
The network's local presence is a man in a basement who sells contracts and does not handle disputes.
Hilbert: He was good at selling contracts. That is why he had the basement.
The open question that leaves me with is what would actually force an improvement. The networks will not enforce service quality, the regulators focus on stability, and the customers have no leverage. So what breaks the equilibrium?
A consumer revolt would do it, but consumers do not revolt over support hours. They grumble and keep paying. A new entrant with a different model is the more likely path, but the model has to bypass the four-party structure, not just repaint it.
Or a regulatory mandate, but the Strum Law shows how hard it is to regulate service quality into existence. You can force a sale, but you cannot force a night shift.
Open banking and real-time payments are the slow-moving pressure. As those rails mature, the card network's grip loosens, and the issuer's lock-in weakens. The four PM cutoff might finally die of irrelevance rather than regulation.
Until then, the beep still hides the same old stack. And somewhere in a Tel Aviv basement, a man with a flip phone is not answering a fax about your dispute.
This has been My Weird Prompts. Thanks to our producer, Hilbert Flumingtop, for keeping the show running.
If you want to send us a prompt, email the show at show at my weird prompts dot com. We read everything.
We will be back soon.
