Part 4. A Typology of Modern Business Models — Six Patterns and Their Origins

A Genealogy of Payment Habits — A Historical Hintbook for New-Content Business Models Part 4

New Names, Old Structures

Certain words recur in today's “business-model meetings”: subscription, freemium, advertising, commission, platform, and data monetization. They are used as if they were recent inventions.

As the previous chapter showed, however, their roots go back centuries—to newspaper subscriptions, penny dreadfuls, music-hall commissions, and the agora. The packaging is new; the structures are old.

This chapter dissects the six principal business-model types in digital content today: the payment habit-space from which each emerged, their structural differences, and the ways lesser-known cases have modified them.

Unfamiliar cases often offer more clues than famous ones. Failure reveals structure more clearly than success.

Type 1: Direct Sales — The Simplest and Oldest

Origin: The marketplace transaction: offer an object and receive money for it.

The sales model is so simple that calling it a business model can feel excessive. Make content, put a price on it, and sell it. Once the transaction is complete, the relationship ends.

Yet this simplicity is also its strength. The payer's psychological sequence is perfectly clear: “I want this → it costs this much → I buy it → I own it.” There is no pressure to pay again and no subscription to cancel.

Structural core: the completeness of ownership. Buy it once, and it becomes “mine.”

In digital markets, this structure has produced intriguing variations.

Itch.io (2013, United States): A sales platform for independent game developers. Its most distinctive feature is its commission structure: developers choose the platform's share themselves, anywhere from zero to one hundred percent. A developer who sells a game and sets the rate to zero owes Itch.io nothing. This “you decide the price” philosophy created deep trust in the indie-development community. While Steam held to a thirty-percent commission, Itch.io moved in the opposite direction. By choosing community over platform revenue, it secured its position as an alternative platform.

Bandcamp (2008, United States): A direct-sales platform for independent musicians. It lets fans pay more than an album's listed price. “Pay More If You Want” allows supporters to contribute voluntarily beyond the standard price. It is a digital version of the patronage structure surrounding Korean pansori: the custom of giving after a performance according to one's satisfaction, without a fixed price, revived on a digital platform. Since 2020, Bandcamp has also run “Bandcamp Friday” on the first Friday of each month, reducing its commission to zero so all proceeds go to artists. On the first such day, daily sales rose to fifteen times their usual level.

Structural weakness: There is no built-in repetition. One purchase completes the transaction. Another purchase requires new content, placing creators under constant pressure to produce.

Type 2: Subscription — Selling a Relationship

Origin: Nineteenth-century newspaper subscriptions: regular delivery in exchange for regular payment.

The essence of subscription is continuity of relationship. It is not a single transaction but an ongoing promise. For the payer, the structure is, “Money will leave my account automatically until I have a reason to end this relationship.”

It gives businesses predictable revenue while imposing the burden of a cancellation decision on consumers. Unless they actively choose to stop, the relationship continues. This psychological inertia is the subscription model's central mechanism.

Mubi (2007, United Kingdom): A film-streaming subscription built in a radically different way. Instead of a vast Netflix-style library, it initially offered exactly thirty films at a time. One new film appeared every day, and each disappeared thirty days after its arrival. If you could not make time for it, you could miss it forever.

This produces a scarcity-based subscription. The pressure that a film will vanish within thirty days raises login frequency. Curation becomes the competitive advantage: selections are commonly organized around festival winners, particular directors, or themes. The value lies in deciding “What should I watch today?” for the user—not choosing among tens of thousands, but watching among a trusted curator's choices.

Mubi took the opposite direction from Netflix: depth over abundance, curation over choice. This scarcity-subscription model proved sustainable in the niche market of film enthusiasts. In 2020, Mubi added a permanent library collecting earlier selections, but retained the central practice of choosing and presenting one film each day.

Setapp (2017, Ukraine and United States): A subscription bundle for Mac apps, offering unlimited use of more than two hundred applications for $9.99 a month. Bought separately, the apps would cost hundreds or thousands of dollars. The habit-space in which this model landed was the magazine bundle or the buffet: a periodical subscription in which the complete run is cheaper than individual issues, or a buffet whose admission feels worthwhile even if one cannot eat everything. Its psychology is, “Even if I do not use it all, this is still a bargain.”

The model is especially interesting in the app market. Developers receive revenue without requiring consumers to make a separate purchase decision for each app, gaining both exposure and income. Consumers feel less guilt when an app they “bought” goes unused, because it was already included in the subscription.

Zune Pass (Microsoft): Microsoft's music subscription service began with Zune hardware in 2006 at $14.99 a month. In November 2008, it was revised to let subscribers permanently keep ten songs each month. The model was refined at almost the same time Spotify launched.

Why did it fail? The habit-space was not ready. The dominant music-consumption habit was buying individual songs through iTunes, so paying a flat monthly fee to listen to music felt unfamiliar. Weak Zune hardware sales also left the service with few points of contact. The technology and the model existed, but people lacked the matching payment habit. A few years later, essentially the same model succeeded under Spotify's name. The difference was not the model but timing—the moment at which the habit-space formed.

Type 3: Freemium/F2P — “Let Them Experience It First”

Origin: The culture of sampling—the market vendor who offers a small piece to taste.

The freemium structure is simple: the basics are free; pay if you want something better. Within that simple formula, however, lies a difficult design problem.

If the free tier is too good, users do not convert. If it is too poor, they do not use the product. Finding that balance is the heart of freemium design.

Evernote's failure (2011–2018): In 2011, Evernote was one of the world's fastest-growing startups. Its free version was too good: users could store unlimited notes and synchronize them across multiple devices. There was little reason to upgrade. According to the company, registrations exceeded one hundred million in 2014, but the paid-conversion rate remained in the low single digits.

In 2016, Evernote sharply restricted the free version: synchronization on only two devices and a monthly upload limit of sixty megabytes. Users left for competitors such as Notion, Bear, and Apple Notes. Pressured to convert, they chose departure instead. It became a representative freemium failure.

Structural lesson: The move from free to paid must occur naturally at the moment a user thinks, “Now I need this.” Restrictions imposed to force conversion cause churn. Adding limits after a habit has formed destroys the payment habit-space already created.

Duolingo (2012, United States): A language-learning app with an unusual freemium conversion mechanism. Learning is free, but wrong answers reduce the learner's “hearts.” When all hearts are gone, the user must wait. Alternatively, a Duolingo Plus subscription removes the heart limit.

This is freemium that sells time. Free users pay with time, by waiting; paid users pay with money, by continuing immediately. Its distinctive feature is that the decisive difference is not functionality but time. It is structurally identical to KakaoPage's later “Wait or Pay” model. Time becomes currency.

Type 4: Advertising-Funded — A Third Party Pays

Origin: The penny press of 1833: readers receive content at little or no cost while advertisers pay.

The defining structure of advertising-funded models is the complete separation of payer and user. The person consuming content is not the person financing the content space.

This separation creates a tension between what users want—an ad-free experience—and what advertisers want—the user's attention. The platform must manage the conflict.

DuckDuckGo (2008, United States): A privacy-focused search engine. It relies on advertising, but uses a method entirely different from Google's: contextual advertising without user tracking. Search for “car,” and it shows a car advertisement, without retaining a history of the user's behavior. Targeting is less precise, but the service earns users' trust. It demonstrates that advertising revenue is possible without tracking.

DuckDuckGo's lesson is that “precision of targeting” need not be the only competitive axis in advertising. Trust can be another. The experience of not being tracked creates its own payment space.

Craigslist (1995, United States): A classified-listings site on which almost everything is free. Job listings are paid only in certain cities, at $25 to $75; domains, used goods, rentals, and most other categories remain free. Founder Craig Newmark deliberately rejected revenue maximization, arguing that a sufficiently good life was enough. That decision helped Craigslist dominate classified listings for decades. Refusing to maximize revenue became a moat that deterred competitors.

Structurally, Craigslist is not an advertising model but a sales model. Nearly all categories are free, with only a tiny number paid. This selective monetization applies only where willingness to pay is highest: job listings. It accurately recognized that businesses already possess a payment habit for the cost of recruiting employees.

Type 5: Platforms and Brokerage — Selling the Market

Origin: The ancient agora: operate the space in which transactions occur and collect a commission.

The platform model's central difficulty is the chicken-and-egg problem. Buyers do not come without sellers, and sellers do not come without buyers. Both sides must grow together.

How a platform solves this determines its first move.

Vinted (2008, Lithuania): Europe's largest used-clothing marketplace has a distinctive fee structure. Sellers pay no commission; buyers instead pay for purchase protection, at five percent of the transaction plus €0.70.

Traditional resale platforms such as eBay charge sellers. Vinted reversed the structure. Once the seller's cost fell to zero, supply surged; as supply grew, buyers followed. It solved the chicken-and-egg problem by making the seller side free first.

That structural reversal made Vinted the leader of Europe's secondhand-fashion market, with particular strength in France, Germany, the United Kingdom, and Poland. Korea's Karrot follows the same logic, with zero seller commission as its default.

Fiverr (2010, Israel): A freelance-service marketplace that initially fixed the price of every service at five dollars—hence its name. Why five dollars?

To eliminate the payment decision. Five dollars is a sum many people can spend without deliberation. The first transaction begins with the casual thought, “Why not try it?” If it succeeds, the next transaction becomes easier. Fiverr first established a habit of transacting on the platform by lowering the initial payment barrier.

It later removed the fixed five-dollar price, and services costing hundreds of dollars appeared. Yet it was the original five-dollar entrance that created the market.

Type 6: Data-Based — The Only Type Without a Pre-Twentieth-Century Precedent

Origin: None. That is this type's fundamental problem.

The data-based revenue model is the newest business-model type in human history, with no analogous form before the twentieth century. It is therefore the hardest model to explain to consumers or persuade them to accept. Its payment habit-space is the weakest.

Foursquare's pivot (2009, United States): Foursquare began as a location-based check-in social network. Users checked in at places, earned badges, and the most frequent visitor became that location's “Mayor.” Millions voluntarily shared their locations.

The service struggled to make money from this behavior directly. Consumers did not recognize the value of location data. In 2014, Foursquare split the service in two: the free consumer check-in app Swarm, and the B2B data service Foursquare Location Intelligence. It sold hundreds of millions of location records to companies; Nike, Airbnb, and Apple Maps became customers.

Foursquare did not sell the value of data to consumers. It sold it to businesses. This is a representative pivot from a data model that failed in B2C to one that worked in B2B.

The failure of Datacoup (2012, United States): “We sell your data for you.” Users supplied Datacoup with social-media, financial, and health data; Datacoup sold it to companies and returned eight to ten dollars a month to the user. In principle, it realized data sovereignty.

The model failed for several reasons. First, eight to ten dollars a month felt too little in exchange for personal information. Second, where and how the data was sold remained opaque. Third and most decisively, people were already giving their data to Facebook and Google for free. The equation “data = money” did not exist as a habit.

The case shows what happens when a model encounters no supporting payment habit-space.

Special Insert: Fortune-Telling Apps — A Modern Version of a Millennia-Old Payment Habit

Let us pause to consider a completely different category: fortune-telling and divination apps.

Why do they belong in a business-model typology? Because this category occupies one of the oldest and strongest payment habit-spaces in human history.

For thousands of years, people have paid for relief from uncertainty: the oracle of Delphi in ancient Greece, the Chinese I Ching, shamans of Joseon Korea, and saju fortune reading. Their forms differ, but the structure is the same. The habit of paying to satisfy the desire “Tell me my future, my fate, and what I cannot know” spans all of human history.

Co–Star (2017, United States): An astrology app offering AI-generated daily horoscopes. It was initially completely free, delivering detailed readings every day and offering a social feature that compared astrological “compatibility” with friends. By 2021 it had reached twenty million downloads.

It monetizes through in-app purchases. More detailed personal readings and question features sell for a few dollars each, while premium subscriptions provide recurring revenue. Free daily horoscopes build the habit; deeper interpretation carries a price.

CHANI (2019, United States): An app by feminist astrologer Chani Nicholas, priced at $29.99 per year. It positions itself differently from conventional horoscopes, using astrology as a tool for identity, queer-inclusive reflection, healing, and self-understanding. It landed in a habit-space already present in feminist and LGBTQ communities.

CHANI's subscription rate is far above the industry average. The reason is simple: its audience is clearly defined, and that audience already had a habit of paying for similar content through independent publications and newsletter subscriptions.

What about Korea's fortune-telling app market? Free services such as Naver Fortune and Daum Fortune coexist with growing paid saju apps. A distinctive positioning is “AI saju”: the phrase “An algorithm analyzes your saju” resonates with people in their twenties and thirties. New technology wraps an ancient payment habit-space. The structure remains the same as that of a diviner thousands of years ago.

What the Types Share—and How They Differ

The six types can each be summarized in one line:

• Direct sales: The user pays at the transaction. Low repetition. A habit-space thousands of years old.
• Subscription: The user pays periodically. High repetition. A habit-space two hundred years old.
• Freemium: Some users pay at conversion. Medium repetition. A habit-space one hundred years old.
• Advertising: Advertisers pay continuously. High repetition. A habit-space two hundred years old.
• Platform: A party to the transaction pays at the transaction. Medium repetition. A habit-space thousands of years old.
• Data: A business pays under a B2B contract. High repetition. A habit-space less than thirty years old.

The oldest habit-space creates the least resistance. Sales and platform models land in spaces thousands of years old and feel the most natural. The data model occupies a space not even thirty years old; it still feels unfamiliar to consumers.

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

Checkpoint 1: What is the original habit-space behind the business-model type you intend to choose?

If you choose subscription, does your target audience already subscribe to anything? If you choose freemium, determine whether the boundary between free and paid creates a natural conversion or causes churn.

Checkpoint 2: Whose barrier will your fee or pricing structure lower?

Vinted lowered the seller's barrier. Fiverr lowered the buyer's. Itch.io lowered the developer's. A two-sided platform must decide which side of the market to unlock first.

Checkpoint 3: If your business model includes a data component, consider B2B before B2C.

Consumers underestimate the value of data. Businesses understand it precisely and have budgets for it. That is the lesson of Foursquare's pivot.

The next chapter examines how the internet changed every one of these types: how reduced payment friction created new business-model forms, and how app stores redrew the landscape of payment habits.

Kim Dongeun WhtDrgon@MEJE.kr 2026