Part 5. Digital Transformation — Whoever Eliminates Friction Takes the Market

A Genealogy of Payment Habits — A Historical Hintbook for New-Content Business Models Part 5

Friction Is the Enemy of Payment

“I will pay.”

What lies between that decision and the payment itself? Must the customer take out a wallet, enter a PIN, type an account number, or write an address? The shorter this distance, the more often payment happens. The longer it is, the more payments disappear.

The essence of what digital technology has done over the past seventy years is singular: reduce payment friction. Every reduction created a new payment habit, and every new habit gave rise to a new business model.

This chapter traces that process. It also examines what happens when friction is miscalculated.

Credit Cards — The First Habit of “Use Now, Pay Later”

A New York restaurant, 1949.

The familiar story says that businessman Frank McNamara discovered after dinner that he had left his wallet behind, and his wife brought cash to settle the bill. Some witnesses later said the company's publicity staff invented the tale. Whatever its origin, McNamara founded Diners Club in February 1950: one paper card, accepted at twenty-seven restaurants, with an initial two hundred members.

It was the first credit card.

The card's essence, however, was less credit than club—a membership. It signaled the identity of “someone who belongs to this club and can dine with only a signature.” Annual membership cost three dollars. The card was both membership and proof of identity.

Credit cards met resistance as they spread. Banks initially refused them, and consumers resisted the psychology of “going into debt.” They collided with a payment habit thousands of years old: buy with money one already possesses.

What broke down this resistance was simple—the accumulation of convenience. One card could be used in many places, and that convenience gradually dissolved each layer of psychological resistance.

Klarna (2005, Sweden): The modern heir to the credit card: “Buy Now, Pay Later,” or BNPL. Klarna brought the idea to online shopping with exceptionally little friction. No bank-account or card number was needed; an email address and postal code completed the purchase, with a bill thirty days later.

By 2021 it was Europe's largest fintech startup, valued at $45 billion, with more than 250,000 merchants including IKEA, H&M, and Nike.

Structurally, Klarna repeats the history of Diners Club. Nothing is fundamentally new. It repackaged a seventy-year-old habit—use now and pay later—by reducing the friction of online checkout.

Klarna's paradox: As payment friction fell, overspending rose. After interest rates increased in 2022, user delinquency surged. Friction had also served to prevent excessive consumption. Remove too much of it, and consumer protection becomes a problem. That is the essence of the regulatory debate over BNPL.

Amazon 1-Click — The Prototype of Friction Removal

One case demonstrated earlier and more clearly than any other that eliminating friction is itself a business-model innovation.

In 1997, Amazon filed a patent for 1-Click Ordering; the patent was granted in 1999. Once a customer registered card and delivery information, every later purchase could be completed with a single click. Moving to a cart, entering an address, and confirming payment all disappeared.

Amazon proved a simple point: conversion rises when the purchase finishes before the hesitation—“Should I buy this?”—can begin. Every step creates a chance to abandon the transaction; fewer steps mean less abandonment. Amazon used the patent to block competitors legally from adopting similar features until it expired in 2017. Once it expired, one-click checkout became an e-commerce standard.

This case shows that reducing payment friction is not merely a UX improvement. Friction removal directly raises purchase conversion, and conversion directly raises revenue. Amazon defended the connection with a patent because it understood its value precisely.

Stripe — The Seven Lines of Code That Changed Payment Infrastructure

In 2010, brothers Patrick Collison, twenty-one, and John Collison, nineteen, founded a Silicon Valley startup around a simple question: “Why is accepting payment on the internet so difficult?”

At the time, adding payment to a website could require months of direct bank contracts, PayPal partnerships, and PCI security certification. Stripe solved it in seven lines of code. Copy and paste the code, and credit-card payments worked.

The speed at which developers could test business models changed. Payment infrastructure no longer took months; it could take hours.

What Stripe created was not merely a payment service but a dramatic reduction in the cost of business-model experimentation. Once payment became easy, more startups could test revenue models rapidly, shortening the time required to find one that worked.

Paddle (2012, United Kingdom): A less famous comparable service built for SaaS companies. It differs from Stripe in one crucial respect: Paddle acts as the merchant of record. A SaaS company accepting payments itself must handle value-added tax in every country. Paddle becomes the legal seller and manages tax requirements across 135 countries.

The difference seems small, but it removes the friction of global selling in one stroke. Paddle opened a space for developers who had abandoned sales in certain countries because tax processing was too burdensome. Removing that friction opened a new payment space.

The App Store's Thirty Percent — Whoever Manages Friction Takes the Revenue

Apple opened the App Store in 2008. Developers listed their apps and kept seventy percent of the revenue; Apple took thirty percent. This structure dominated the mobile ecosystem for the next fifteen years.

Where did thirty percent come from? It was not an evidence-based figure. Some argue that it drew on the thirty-to-forty-percent margins of physical retailers. Yet physical stores manage inventory and pay rent, costs that digital distribution does not bear. Even so, thirty percent became the “standard.”

Its structural meaning is the price paid to Apple for managing payment friction. Once inside the App Store, a developer need not handle payment infrastructure, fraud prevention, refunds, or taxes. Apple handles them all. Thirty percent is the price of that convenience.

Epic Games v. Apple (2020): Epic, operator of Fortnite, attempted to bypass App Store fees with direct payment. Apple removed Fortnite from the store, and Epic filed an antitrust suit. The court battle reached an initial decision in 2021, although the dispute continued.

The conflict revealed that the App Store's thirty percent is not merely a commission; it is the price of access. Reaching 1.5 billion iPhone users requires paying thirty percent. Whoever manages the friction sets the cost of access.

F-Droid (2010): An open-source Android app marketplace with zero commission. Developers publish apps and keep all revenue. There are no advertisements, and privacy is central. Its scale is small—about four thousand apps—but it is a powerful alternative for users committed to privacy and open source.

F-Droid proves that zero commission is possible, but the tradeoff is payment infrastructure, search exposure, fraud prevention, and related services. The equation holds: managing friction has a cost.

Network Effects — Asymmetric Friction Determines the Winner

The best-known idea in platform business models is the network effect: as the number of users increases, the service becomes more valuable, and that value attracts still more users. Once the virtuous cycle begins, latecomers struggle to catch up.

Some companies have tested this logic in reverse.

Path (2010, United States): A social network that deliberately limited users to fifty friends at launch, raising the limit to 150 in 2012. Based on Dunbar's number—the theory that humans can sustain meaningful relationships with roughly 150 people—it presented itself as a network for “only your true friends.” The limit was designed to suppress network effects.

It failed. Paradoxically, one reason people use social networks is the desire to show their content to more people. The 150-person limit blocked that desire. Path gave users no reason to bring along the people who had moved to Facebook or Instagram.

Vero (2015, Lebanon and United States): A social network promising no algorithm, no advertising, and a chronological feed. In February 2018, creators dissatisfied with changes to Facebook and Instagram algorithms moved there en masse; one million people joined in two days, and the servers went down.

Most returned within a month. The reason was friction. They knew nobody on Vero and had no followers. A creator with one hundred thousand followers on Instagram had to begin at zero. The accumulated network effect was absent, and rebuilding it became the switching cost that sent users back.

The enemy of a new social platform is neither algorithms nor advertising. It is the relationship network already accumulated elsewhere. Migration will not happen unless there is a compelling reason to abandon it.

The lesson of Cyworld (2001, Korea): Cyworld completely dominated Korea's social-network market in the mid-2000s. Users maintained “mini-homepages” and bought background music with a virtual currency called acorns. Roughly half the Korean population used it. It was among the first mass-market services where users paid real money to decorate a virtual space, and annual acorn sales reached tens of billions of won.

It then fell behind in the transition to smartphones and collapsed when Facebook entered. Cyworld was domestic; Facebook connected users to friends, colleagues, and acquaintances overseas. The new value of a “global network” outweighed the locally accumulated relationship network. Cyworld was not alone—Germany's StudiVZ and the Netherlands' Hyves followed the same path—but it became a leading counterexample in social-network competition, where established networks usually win. Facebook offered value in a different dimension.

Cyworld left another legacy. Its package of virtual-space decoration, virtual currency, and monthly background-music subscription influenced later designs at Kakao, Naver Band, and Zepeto. It failed as a platform, but the payment habit-space it created remains in Korea's digital-content market.

Dynamic Pricing — When Payment Habits Are Redesigned in Real Time

Fixed prices were the basis of commerce for a long time: an item bears a price, and the customer buys it or does not. When nineteenth-century department stores introduced posted prices, “commerce without bargaining” became a standard of civilization.

The internet is changing it again.

Humble Bundle (2010, United States): A bundle of indie games sold under “Pay What You Want.” Customers can buy it for any amount. Yet the structure contains an intriguing device: it displays the buyers' average payment in real time and offers bonus games to anyone who pays above the average.

This creates payment pressure through social comparison. The customer no longer decides alone, but in view of what everyone else has paid. Paying below average feels vaguely uncomfortable. It is a nudge, not a command.

Buyers can also direct part of the proceeds to charity. The structure “pay more, and more goes to charity” legitimizes overpayment. Humble Bundle's average payment was still only a few dollars, far below the combined list price of its games. The model's power came not from a high unit price but from combining purchase volume attracted by a low threshold with a nudge that raised the average payment.

Uber surge pricing: Prices rise when demand rises—when it rains or when a performance ends. They change in real time. The system met fierce resistance when introduced: “Why should the same service have a different price?”

Users adapted, learning that “it costs more when it rains.” This became a new payment habit: consumers accepted the move from fixed to dynamic prices.

Wendy's surge-pricing attempt (February 2024): The American fast-food chain announced that it would introduce digital menu boards in 2025 and apply dynamic prices by time of day. The backlash was immediate: “Are you going to charge more at lunch?” Within days, Wendy's clarified that it did not intend surge pricing—that is, peak-time price increases.

Why did consumers accept Uber's dynamic pricing and reject Wendy's?

The difference lies in payment habit-space. The idea that prices rise with demand already existed in transportation through airfares, hotels, and taxis, so Uber landed in a prepared space. Fast food was different: its habit-space was the inexpensive fixed price. Trying to destroy that space triggered resistance. The same technology does not work when it has nowhere to land.

Failure 1 — The Fall of Facebook Credits

In 2009, Facebook announced Facebook Credits, a virtual currency accepted across all games and apps on Facebook, with one credit worth ten cents. Facebook intended to dominate its in-app ecosystem through a currency of its own issue.

Adoption was mandatory. Beginning in 2011, any Facebook app selling virtual currency had to use Facebook Credits. Facebook took thirty percent of the revenue.

Developers strongly resisted. Zynga, creator of FarmVille and CityVille, objected most forcefully because its in-game currencies had to be replaced by Facebook Credits. After roughly a year of conflict, Facebook withdrew the mandatory requirement in 2012 and ended the service in 2013.

Why did it fail? Users had no payment habit-space for putting real money into “a currency issued by Facebook.” And because game developers refused to cooperate, there were few places to spend Credits. A currency gains value when it is widely usable; Facebook tried to impose one before building its ecosystem.

Failure 2 — The Collapse of Paid Apps on Early Google Play

When Android Market—now Google Play—opened in 2008, its paid-app market was far smaller than the iPhone App Store's. Why?

Several factors combined.

First, refunds were too easy. Google's initial policy offered unconditional refunds within twenty-four hours. A customer could buy an app, use it for twenty-three hours, and return it. Developers struggled to sell paid apps under that structure.

Second, Android devices spanned a wide price range, mixing users of premium and inexpensive hardware. The latter group had less habit of paying for apps.

Third, Android allowed direct installation of APK files from outside Google Play, opening the door to piracy.

In December 2010, Google shortened the refund window to fifteen minutes. Conditions improved, but the gap with the iOS ecosystem persisted for years. Developers began to say, “Android users do not pay.” A more accurate explanation is that the Android ecosystem's checkout friction and refund structure failed to form a habit of paying for apps.

The Paradox of Deliberately Adding Friction — Dark Patterns

Removing friction is not always good. Companies also design friction deliberately, a practice known as a dark pattern.

Making the cancellation button hard to find; repeatedly asking, “Do you really want to give up all these benefits?” during cancellation; converting a free trial into a paid plan without notice. All are intentional friction: make payment easy and cancellation difficult.

Amazon Prime's cancellation flow: In June 2023, the US Federal Trade Commission filed a formal lawsuit. Canceling Prime required passing through four or more screens, each containing a leading question such as “Do you really want to quit?” It became a legal issue because the flow deliberately created friction to continue payments consumers did not want.

Dark patterns reduce churn in the short term but destroy brand trust in the long term and invite regulation. The EU's GDPR and US FTC guidelines have strengthened in this direction. If removing friction is business-model innovation, deliberate friction is its distortion; both deeply affect payment habits.

Asia's Early Experiments — Another Route to Friction Removal

Friction removal did not occur only in the West. Faster and more fundamental experiments appeared first in Asia.

Suica (2001, JR East, Japan): A transit card expanded to payments at convenience stores, vending machines, and restaurants. Mobile Suica arrived on phones in 2006, allowing payment with a tap—eight years before Apple Pay.

Suica succeeded because it landed in an existing payment habit-space. Japan already had strong transit-card and vending-machine cultures. For people accustomed to carrying coins, buying from a machine without coins was a natural extension.

WeChat Pay (2013, China): Payment embedded in a messaging app. Launched in 2013, it added hongbao, digital red envelopes containing New Year's gift money, for the 2014 Lunar New Year. It reproduced the Chinese tradition of presenting gift money in a red envelope. Tens of millions were exchanged within days.

WeChat Pay's explosive growth began with hongbao. By inheriting a millennia-old payment habit-space, loading money into a WeChat Pay account felt natural. The service then expanded into offline payment. Show a QR code, and payment is complete—even at street stalls throughout China without cash.

Kakao Pay (2014, Korea): Payment inside KakaoTalk. It added checkout after KakaoTalk had become Korea's near-universal messenger. No new app was required; payment lived inside an app people already used every day.

The three cases share one trait: they did not create entirely new payment habits. They inserted payment into existing behaviors—taking public transit, giving New Year's money, and talking through a messenger. Friction was minimal not only because of technology, but because the habit-space in which it landed already existed.

What the Internet Changed—and What It Could Not

The internet changed these aspects of payment habits:

• Minimized friction: A world where one click completes payment.
• Automated payment cycles: A world where subscriptions continue without interruption.
• Fluid prices: Prices that change in real time.
• Granular payment units: A world where even one-won units are possible.

It did not change these:

• The need for trust: Payers still pay only when they trust.
• The priority of payment habit-space: Even with new technology, payment does not happen without prior experience paying for something similar.
• Perception of value: People still decide whether something is worth paying for.
• Psychological resistance: Resistance to new payment methods exists independently of technology.

Wendy's, Facebook Credits, and the regulatory controversy around Klarna's BNPL all teach the same lesson. Technology can eliminate friction, but technology alone cannot solve the problem when no payment habit-space exists or when the model collides with an established one.

This Chapter's Hints — Checkpoints for New-Content Planners

Checkpoint 1: Where is the payment friction in my content?

After users decide to pay, what steps remain before payment is complete? Do people drop out at each step? Can the friction be reduced, or can a platform solve it on your behalf?

Checkpoint 2: Compare the platform's thirty percent with the friction of direct sales.

The App Store's thirty percent is expensive. Yet directly managing payment infrastructure, taxes, refunds, and fraud prevention also carries significant cost. Outsourcing taxes alone, as Paddle does, can reduce the friction of global sales. Calculate whether the thirty percent is merely a commission or the cost of friction management.

Checkpoint 3: If you are considering dynamic pricing, does your target payment habit-space already contain variable prices?

Uber's surge pricing was accepted because transportation already had variable-price habits through airlines and hotels. Wendy's proposal was rejected because fast food carried a fixed-price habit. The same technology produces different results in different spaces.

The next chapter begins an industry-by-industry investigation, starting with games: where the habit of putting a coin into a machine came from, and what carried that habit into the digital world.

Kim Dongeun WhtDrgon@MEJE.kr 2026