Aggregation Theory

Ben Thompson’s account of how value accrues on the internet, to whoever owns the demand-side user relationship and can commoditize supply, and how to tell a real aggregator from a business that merely has scale.

For most of economic history, the way to win was to control supply. Whoever owned the printing press, the shelf space, the spectrum, or the distribution trucks held the power, because supply was scarce and getting a product in front of customers was the expensive, defensible part. The internet inverted that. When distribution is free and supply can be summoned on demand, controlling supply stops being an advantage, and the company that owns the direct relationship with the user captures the value instead. Aggregation Theory is the framework that names this inversion and explains why a handful of internet companies grew to dominate their markets by owning demand rather than supply.

What It Is

Aggregation Theory, developed by Ben Thompson in his Stratechery writing beginning in 2015, describes how value accrues in markets where the internet has removed the historical advantages of controlling distribution. The argument has three moving parts.

First, the internet drives the cost of distribution and transactions toward zero. A pre-internet incumbent’s power often came from owning a scarce distribution channel; once distribution is free, that source of power evaporates.

Second, with distribution free, supply gets commoditized. When any supplier can reach any customer at no cost, no individual supplier holds the scarce position anymore, and the scarce thing becomes the demand: the users’ attention and the relationship with them.

Third, the company that owns the user relationship can therefore aggregate that demand, and the aggregation compounds. Each new user makes the service more attractive to suppliers, which improves the service for users, which attracts more users. Because serving an additional user costs the aggregator almost nothing, it can scale to own the demand side of an entire market, and from that position it dictates terms to a supply base that has nowhere else to go.

The defining test of an aggregator has three conditions, and all three must hold:

• A direct relationship with users. The aggregator owns the user, not a reseller or a channel partner.
• Zero marginal cost of serving each additional user. Growth does not require proportional spending, so scale is nearly free.
• Demand-driven multi-sided networks with decreasing acquisition costs. As the aggregator grows, suppliers are compelled to join to reach the users, and that growing supply pulls in more users. Each new user costs less to acquire than the last, not more.

Google aggregates the demand for finding information and commoditizes the websites that supply it. Amazon aggregates the demand for buying things and commoditizes the merchants and brands that supply the products. Netflix aggregates the demand for watching video and commoditizes the studios that supply it. In each case the aggregator never produced the underlying supply; it owned the user and let supply compete for access.

Why It Matters

Aggregation Theory matters because it identifies, in advance, which businesses can win a winner-take-most position and which cannot. The three conditions are a filter, not a description after the fact.

The framework draws one crucial distinction: between an aggregator and a platform, what Thompson calls the Bill Gates Line. The name comes from Gates’s own test for a platform, that its ecosystem is worth more than the company that built it. A platform like Windows enables a relationship between third parties and earns by empowering them. Developers build on it and keep their own customer relationships. An aggregator does the opposite. It sits between the supplier and the user, owns the user directly, and commoditizes the suppliers rather than empowering them. The distinction decides who holds the power. A platform’s suppliers can grow strong enough to leave; an aggregator’s suppliers depend on it for access to demand they can’t reach any other way.

The three readers use the framework from different seats.

A founder uses it as a strategic test: can this business plausibly own a direct user relationship, serve users at zero marginal cost, and pull suppliers in on improving terms? If the answer is no, if the model requires owning scarce supply, or if each new customer costs as much as the last to acquire, then the winner-take-most outcome the pitch implies isn’t available, and the strategy needs a different basis for durability.

An investor uses it to spot the rare businesses that can capture an aggregation position, which is one of the most valuable outcomes the internet produces and a structural reason a company can clear the defensibility bar. It also sharpens the diligence question: a company claiming aggregator economics but actually operating as a platform, or owning supply rather than demand, is making a weaker bet than the language suggests.

A talent reader uses it to judge the durability behind an equity grant: a genuine aggregator’s position strengthens with scale, while a business that merely has scale today can have it competed away tomorrow.

What the concept gives all three is a way to separate a business that aggregates demand from one that simply has a lot of users. Only the first commoditizes its suppliers and compounds toward a durable position.

How to Recognize It

A real aggregation position shows up as the three conditions holding together and as a specific, observable power over suppliers. A few tests separate it from a business that borrows the language.

• Run the three-condition test honestly. Direct user relationship, zero marginal cost to serve, and acquisition costs that fall as the network grows. If a new customer costs as much to acquire as the last one did, the demand-side flywheel is not turning, and the business is scaling, not aggregating.
• Who owns the user, you or your suppliers? If suppliers reach their own customers through the service and could leave with those relationships intact, the business is a platform, not an aggregator, and the power sits with the suppliers. The aggregator owns the user and rents access to suppliers.
• Can the supply be commoditized? An aggregator’s power over suppliers comes from supply being abundant and substitutable from the user’s point of view. Where supply is genuinely scarce or differentiated (a single must-have studio, a sole-source manufacturer), the aggregator’s bargaining position is correspondingly weaker, because the user came for that specific supplier.
• Does the product improve as more users join? The demand-side network effect is the engine. If the service is no better at a hundred million users than at one million, the compounding that produces a winner-take-most position is absent.

How It Plays Out

Google is the cleanest demonstration. It produces none of the information it serves; the websites of the world supply that. What Google owns is the demand: the user who wants an answer and starts at the search box. Because that user relationship is direct and serving one more query costs almost nothing, Google aggregated nearly all the demand for finding information, and from that position the websites that supply the answers have little choice but to compete for ranking on Google’s terms. Supply was commoditized; demand was owned; the value accrued to the aggregator.

Amazon shows the same shape in commerce. The third-party marketplace turned millions of merchants into interchangeable suppliers competing for the buy box, while Amazon kept the customer relationship, the payment credentials, and the demand. A merchant who leaves loses access to the buyers; the buyers barely notice which merchant filled the order. That is the signature of an aggregation position: the supplier needs the aggregator far more than the aggregator needs any single supplier.

Netflix shows where aggregation runs into a limit. It aggregates video demand, but its suppliers are studios. Past a point, a studio can pull its content and stand up its own direct-to-consumer service, which is exactly what Disney and others did. When a supplier isn’t commoditizable, because users came specifically for that catalog, the aggregator’s power over it weakens, and the supplier can defect to owning its own demand. The framework predicts this. Aggregation is strongest where supply is abundant and substitutable, and it frays where a supplier is itself a destination. The lesson isn’t that Netflix failed to aggregate; it’s that the durability of an aggregation position depends on how commoditizable the supply actually is.

Consequences

Reading a market through aggregation rather than scale changes which businesses a founder will try to build and which an investor will underwrite, and the framework carries real limits of its own.

Benefits. A founder who applies the three-condition test early learns whether the winner-take-most outcome the strategy implies is structurally available or merely hoped for. That tells them to build toward owning the user relationship rather than accumulating supply the internet has already devalued. An investor gets a forward filter for the rare businesses that can capture the most durable position the internet produces, plus a sharper read on the difference between aggregator economics and the platform economics a pitch sometimes mislabels. And all three readers get a checkable question in place of a vague intuition: who owns the demand, and can the supply be commoditized? The intuition that a large internet company is automatically a strong one does not survive that question.

Liabilities. The framework is a lens, not a law, and it invites two errors. The first is overfitting: treating every large platform as an aggregator and every aggregator as permanent, when the supplier relationship and the commoditizability of supply are what decide whether the position actually holds. The second is mistaking the conclusion for a strategy. Knowing that aggregators win doesn’t tell a founder how to become one, and attracting the first users before the demand-side flywheel turns — the cold-start problem — is exactly the hard part the framework names but doesn’t solve. The theory also speaks to one market structure, where distribution is free and supply is commoditizable, and applies poorly where supply is genuinely scarce, regulated, or differentiated, which describes a large part of the economy. It explains where a particular kind of internet power comes from. It does not describe every business, and treating it as a universal theory of competition is its most common misuse.

Related Articles

Sources

• Ben Thompson, “Aggregation Theory” (Stratechery, 2015) — the founding statement of the framework, defining the three conditions and the shift from supply-side to demand-side value capture; the related Stratechery essays “The Bill Gates Line” (2018) and “Defining Aggregators” (2017) supply the platform-versus-aggregator distinction and refine the conditions.
• The pre-internet account of value accruing to whoever controls distribution draws on the industrial-organization tradition in strategy (Porter’s analysis of distribution and supplier power), which Aggregation Theory positions itself against by arguing that free distribution dissolves the source of that power.
• Bill Gates’s formulation that a platform exists when the value of the ecosystem built on it exceeds the value of the company that created it is the origin of the “Bill Gates Line” Thompson uses to separate platforms from aggregators.
• The Google, Amazon, and Netflix cases are drawn from the companies’ publicly documented business models and the well-reported supplier dynamics — search-ranking dependence, marketplace buy-box competition, and studio direct-to-consumer defection — rather than from any single proprietary source.