Aggregation Theory
A reader's summary of Ben Thompson's framework, first published on Stratechery in 2015 — its history, author, main ideas, critiques, and afterlife. Editorial synthesis, not a substitute for the essays.
The essay at a glance
Aggregation Theory is Ben Thompson's answer to a simple question: why did the internet produce a small number of extraordinarily powerful companies — Google, Amazon, Facebook, Netflix — that don't own factories, fleets, or, in most cases, the content that runs through them? His answer is that the internet collapsed the cost of distribution and transactions to zero, moving power away from whoever controlled the supply chain and toward whoever controlled the demand: the aggregator that owns the user relationship. It is less a single essay than a framework refined over a decade of near-daily writing, and it has become one of the most widely used vocabularies in technology strategy and, increasingly, in technology regulation.
The author
Ben Thompson founded Stratechery in 2013 after working at Apple, Microsoft and a Taiwanese electronics company, and built it into a subscription publication that analyzes technology strategy on a near-daily cadence — one of the first successful paid newsletters in the genre it helped create. Aggregation Theory, laid out across “Aggregation Theory” (2015) and refined in “The Aggregators” and “Defining Aggregators” the same year, is the piece of that body of work most often cited outside it — the closest thing Stratechery has to a unified theory.
History and context
Thompson framed the shift as a move from the “Age of the Supply Chain” to the “Age of Aggregators.” In the supply-chain era — retail, media, telecom before broadband — the constraint was physical distribution, so power accrued to whoever controlled it: the department store, the cable operator, the newspaper with the printing press. The internet made reaching one more customer, and transacting with them, costless at the margin. Once distribution stopped being the constraint, the theory argues, the winners would be whoever best aggregated user attention and commoditized everyone competing to reach it — a reversal that traditional strategy frameworks built for the supply-chain era did not anticipate.
Main ideas
Distribution went to zero
Industrial-era power sat with whoever controlled distribution — the supply chain, the shelf space, the cable line. The internet drove the marginal cost of reaching one more customer to zero, so control of distribution stopped being scarce and stopped being where the money was.
Platforms vs. aggregators
A platform (Windows, iOS) needs third parties to build on top of it before it has value. An aggregator (Google, Netflix, Amazon) needs no permission from its suppliers — it aggregates end users directly, and suppliers show up afterward, competing for access to the demand it already owns.
Commoditizing the supply side
An aggregator's leverage comes from modularizing what it aggregates: websites become interchangeable search results, shows become interchangeable tiles, sellers become interchangeable listings. The aggregator, not any single supplier, becomes irreplaceable to the user.
The feedback loop
Better user experience attracts more users; more users attract more suppliers competing for access to them; more suppliers means more selection and lower prices, which improves the user experience further. The loop needs no direct connection between users — unlike a classic network effect, it runs even where users never interact with each other.
Super-aggregators
A 2017 follow-up named the strongest case: platforms like Facebook and Google that aggregate both users and suppliers at zero marginal cost on both sides, with no direct relationship or contract required with either — content arrives seeking distribution, users arrive seeking content, and the aggregator mediates without owning or paying for the supply.
Owning the relationship is the asset
The scarce resource shifted from owning the product or the factory to owning the customer relationship and the data that comes with it. Whoever the user opens first — the search box, the feed, the marketplace app — captures the value regardless of who made the underlying good.
Critique
- Not every two-sided market fits. Uber and Airbnb are often described with the same vocabulary, but their supply side — drivers, hosts, physical capacity — has real marginal costs and real constraints the theory was built to abstract away. The fit is looser wherever supply isn't infinitely elastic content or listings.
- Descriptive, not prescriptive, for regulation. The theory names the mechanism well but doesn't by itself settle what regulators should do about it — critics note that “this company is an aggregator” is an observation, not an antitrust remedy, and policy debates have had to supply the rest.
- Buzzword dilution. A decade of adoption in pitch decks and strategy memos has, as with most successful frameworks, loosened the term from its precise original meaning — plenty of companies are now called “aggregators” that merely resell or list, without the zero-marginal-cost feedback loop that defines the category.
- The theory's own disruption. Thompson has written at length about the possibility that AI chatbots answering questions directly, rather than linking out to ranked results, could disintermediate the aggregator layer itself — the same zero-marginal-cost logic that let Google commoditize websites now threatens to let a model commoditize Google.
Impact
Aggregation Theory became a standard reference point for how technology strategists talk about internet-era competition, cited across venture capital, corporate strategy and journalism as the default explanation for why a handful of platforms sit at the center of so much online activity. Its vocabulary shows up, sometimes unattributed, in regulatory language too: the European Union's Digital Markets Act designates “gatekeepers” on essentially the same logic of demand control and supplier dependency that the theory describes. For the operational counterpart — what aggregation looks like from inside the factories being modularized rather than from the platform aggregating them — see the factory as institutional fact.
Notable engagements
- “Super-Aggregators” (2017) — Thompson's own follow-up, narrowing the category to platforms that aggregate both sides of a market at zero marginal cost.
- EU Digital Markets Act — its “gatekeeper” designations echo the theory's demand-control logic, whether or not the drafters used its vocabulary.
- Antitrust commentary on Google, Amazon and Meta — repeatedly analyzed through this lens across a decade of Stratechery coverage of the major platform antitrust cases.
- Thompson's AI-era writing — an ongoing argument about whether foundation-model chatbots become the next aggregator layer or the thing that finally breaks the current one.