Platforms: The Most Powerful Business Model in the Modern World

We have already discussed many principles and methods for making money, but there is one fact you may not have noticed: the traditional model of opening a shop, making a product, placing it on the market, and selling it directly to customers has become remarkably rare.

Whether you open a store on Taobao, sell products through Douyin, or simply run an ordinary restaurant and want to offer online ordering and delivery, you now need to pass through some kind of platform.

At first, this looked like a technological utopia. Platforms promised to eliminate the middleman. They offered services for free, handed out coupons and subsidies, and said: focus on your products, and we will handle logistics and inventory; focus on serving users, and we will match you with them. Platforms supplied search, advertising, trust ratings, and even social connection. They made information flow with unprecedented efficiency.

You could be forgiven for thinking this was a new commercial civilization in which no business would be difficult to run.

But in recent years, attitudes toward platforms have changed. Merchants complain about rising commissions. Buyers find that search results are filled with ads. Delivery workers feel trapped inside algorithms.

The old middleman bought low and sold high. Today’s platform controls access, ranking, reviews, and traffic. It can even determine who survives. What it earns is not exactly a markup. It is the price of order.

A platform is not an ordinary business. It is the ultimate predator in modern commercial history because it turns the market itself into its product. Through algorithms, it decides who can be seen by whom. That kind of power was never supposed to belong to a commercial player.

Why Platforms Win: Two-Sided Markets and Growth Flywheels #

Platforms are powerful because they are “two-sided markets”[1].

A traditional business is one-sided. If you open a bakery, you mainly deal with people who buy bread. A platform must deal with two or more sides at once: sellers and buyers, drivers and passengers, restaurants, couriers and diners, developers and users, or creators, advertisers and audiences.

This makes a platform extremely difficult to start. A marketplace with no merchants attracts no consumers; a marketplace with no consumers attracts no merchants. You need to mobilize both sides. Yet this is also why a successful platform becomes so hard to compete with once it reaches scale.

How do you solve the launch problem? The most common strategy is to provide a completely free service, or even subsidize new users. In plain language: burn money to acquire customers.

You may remember the early days of the mobile internet economy: a Didi ride cheaper than a taxi, or a shared bicycle unlocked for a few cents. Even today, when food-delivery platforms go to war, consumers may receive aggressive discounts or free drinks.

Venture capitalists will say the spending is justified because of “indirect network effects”[2]. More buyers attract more sellers; more sellers attract more buyers. More drivers make it easier for passengers to get rides; more passengers encourage drivers to come online. This is a process of increasing marginal returns. Each additional user raises the average value of the platform for every other user. New users have more reason to join, and existing users have less reason to leave.

Subsidies are not charity. They solve the chicken-and-egg problem.

Once merchants and consumers arrive, the platform gains a second process of increasing returns: the “data-algorithm feedback loop.” More users create more data; more data improves the algorithm; a better algorithm improves matching; better matching attracts more users.

Making serious money requires leverage and scale. Indirect network effects and the data-algorithm feedback loop are the platform’s most powerful forms of leverage. Platform growth is a self-feeding positive feedback process. Its only limit is the population of a country, or of the world.

Platforms are a new kind of super-profitable organization in human history.

Enshittification: From Subsidies to Extraction #

Once a platform has the most users, the best data, and the strongest algorithms, competitors struggle to challenge it. At that point, you might ask: why not simply become a piece of modern commercial infrastructure, stop subsidizing everyone, charge a fair profit, and serve the public?

Reality is not so simple. Once you have that much power, you will not be satisfied with being infrastructure.

You will begin to harvest.

Canadian technology thinker Cory Doctorow coined the term “enshittification”[3] to describe the process by which platforms decay from serving users into extracting value from them. The American Dialect Society selected it as its 2023 Word of the Year[4]. We might politely call the process “platform decay.”

How did such a new and initially beloved business model become so widely resented? Doctorow wrote a book about the process[5]: first please end users; then sacrifice end users to please business customers; finally sacrifice business customers and claw the value back for the platform itself.

The first stage is establishment. The main strategy is subsidizing end users.

Backed by enormous venture-capital investments, platforms provide high-quality services far below cost, and often completely free, to activate indirect network effects.

When Facebook was competing with MySpace, it had no ads, did not surveil users, and displayed only the information users deliberately subscribed to. Amazon sold books for much less than bookstores and offered unlimited shipping for a $79 annual membership. Early Twitter had a completely open application programming interface. Third-party developers could freely access its data, and many users treated Twitter as their own medium and helped invent features for it.

An early platform resembles a new ruler entering occupied territory: build roads, suspend taxes, distribute subsidies, and smile. All it asks is that you arrive, preferably bringing your social relationships, shopping habits, payment history, content assets, and workflows with you.

The second stage is expansion. The main strategy is subsidizing business customers.

You know the phrase: “If you are not paying, you are not the customer; you are the product.”

Once the platform has locked in enough users, it begins selling those users as an asset to businesses: we have a huge audience; would you like to join us? We will charge a low commission and give you free traffic. Sellers, advertisers, app developers, and livestreamers arrive.

Meanwhile, platforms quietly make concessions at the expense of end users. Facebook offers advertisers precisely targeted ads and allows some unsubscribed media articles into users’ timelines. Apple tolerates commercial surveillance by third-party apps. Twitter sacrifices the speech rights and privacy of local users to accommodate censorship requirements in some countries.

Yet the platform still looks benevolent. It creates opportunities for small businesses and stages for creators. Consumers continue to enjoy low prices. Investors watch share prices rise. Everyone seems happy.

But who pays?

Then comes the third stage: extraction. Platforms squeeze both buyers and sellers. Shareholders are the only group still receiving subsidies.

Consumers see more ads, worse search results, and fewer genuine choices. Merchants face higher commissions, more expensive advertising, and more complex rules. Creators discover that unless they pay for traffic, even their own followers cannot see them. Developers feel as though they work for app stores.

The network effects are already in place. Former competitors have either died or reached a stable equilibrium with the dominant platform. Consumers and businesses now find it difficult to escape.

We once thought the spirit of the internet was freedom and free access. What we received was control and extraction.

Algorithmic Power: Changing Choices Without Force #

Do not misunderstand: platform control is not the same as the coercion of an ancient government. It does not involve violence. The platform cannot force you to do anything, and you always have a choice. But it can quietly influence your choices.

The platform’s weapon is the algorithm. It does not need to send a person to manage you, and it does not care whether you, as a particular individual, accept its authority. It only needs to change group behavior at the level of probabilities.

Consider food-delivery couriers and ride-hailing drivers. They are wrapped in the illusion of flexibility called the “gig economy.” You do not punch a clock. You can start or stop whenever you like, right?

But the algorithm watches every second. It calculates your route, counts down the time available, and adjusts your priority for future orders based on performance. If you are late or receive a low rating, you may be penalized. If you complete many orders, you receive a reward: access to more orders, and more valuable ones.

One extreme practice attributed to companies such as Uber is the use of acceptance behavior to estimate how desperately a driver needs money. Once the algorithm decides that you will accept any request, it can assign you the least profitable work. Is that not pressure applied precisely to the vulnerable?

Platforms do not go easy on consumers either. First comes price discrimination, often described in China as using big data to exploit loyal customers. Existing users frequently claim to see higher prices than new users. Some travelers believe repeated flight searches produce higher fares because the algorithm detects a stronger willingness to pay.

Another method is compressing choice. For the same search, the platform places what it wants you to see in the first row and buries what it does not want you to see. Defaults, hidden cancellation buttons, and coupon design can all steer behavior while leaving you convinced that you simply followed your own preferences.

Platforms are even more direct with merchants. Commissions rise openly. If you want consumers to see you, you must also pay advertising fees, service fees, and promotional fees. You must buy traffic. Your visibility has become a paid product.

At the same time, the platform does not need to pay retirement benefits or insurance for delivery workers, and it does not absorb merchant losses. It says it only provides a matching service and sells information. Yet it quietly claims a large portion of the value produced by everyone else’s labor.

Profitable businesses collect economic rent. What rent could be more powerful than the platform’s?

The Infrastructure Paradox: Regulation, Antitrust, and Interoperability #

We can no longer live without platforms.

You have to admit that platforms create enormous value. They let small merchants sell niche goods into remote regions, help independent developers reach global users, and provide a huge amount of employment. Search and maps are still free for most users. Platforms are infrastructure for the modern economy.

But a platform is also infrastructure that competes with the people it serves, while exercising significant control over them. This is no longer a pure market economy, and perhaps no longer capitalism. It resembles claiming a mountain and charging everyone who passes. Some call it “technofeudalism”[6].

What can be done? Law and government regulation matter, especially antitrust enforcement. We once defined monopoly as gaining control of a market and then raising prices. With platforms, we must rethink price: a consumer who pays no money may still bear real costs. The European Union’s Digital Markets Act (DMA) calls large digital platforms “gatekeepers” and aims to limit the abuse of their power over access[7]. China also regulates platforms through the Personal Information Protection Law and rules governing internet-platform pricing behavior.

Doctorow and other scholars[8] strongly favor “interoperability.” The basic idea resembles keeping your phone number when you change carriers. To prevent platforms from locking users inside their castles, the law should ensure that a departing user can take data, social relationships, purchase history, and content assets along, and migrate relationships with users and followers to a competitor. In plain language: tear down the walls and promote competition between platforms.

Conclusion: Platforms Should Be Tools, Not Masters #

Some ride-hailing drivers keep several phones open and accept jobs from several platforms at once. That is exactly right. A platform has never been loyal to you, so it has no right to demand your loyalty.

Algorithms are not laws of nature. They are institutions written by people. A human being is not a data interface, much less a slave to an algorithm.

More people are becoming wary of platforms. That is good news. It means people are not so easy to manipulate, and that free markets still have hope.

Platforms are the most fearsome money-making machines in today’s world. We must ensure that they remain useful tools, rather than masters that are exceptionally good at doing the arithmetic.

Notes #

[1] Rochet, Jean-Charles, and Jean Tirole. “Platform Competition in Two-Sided Markets.” Journal of the European Economic Association 1, no. 4 (2003): 990–1029.

[2] Parker, Geoffrey G., and Marshall W. Van Alstyne. “Two-Sided Network Effects: A Theory of Information Product Design.” Management Science 51, no. 10 (2005): 1494–1504.

[3] Doctorow, Cory. “My McLuhan Lecture on Enshittification.” Medium, 2024.

[4] American Dialect Society. “2023 Word of the Year Is ‘Enshittification.’” January 5, 2024.

[5] Doctorow, Cory. Enshittification: Why Everything Suddenly Got Worse and What to Do About It. MCD, 2025.

[6] Varoufakis, Yanis. Technofeudalism: What Killed Capitalism. London: Bodley Head, 2023.

[7] European Parliament and Council of the European Union. Regulation (EU) 2022/1925 on Contestable and Fair Markets in the Digital Sector, 2022.

[8] For example, Wu, Tim. The Age of Extraction: How Tech Platforms Conquered the Economy and Threaten Our Future Prosperity. New York: Knopf, 2025.