The market that stops existing
How platform bundling dissolves markets without a fight: the envelopment mechanism, the firms that survived it anyway, and the line regulators draw.
ARTICLE
Bilal Farhat
10/24/20264 min read
There is a way to lose a market that does not involve losing to anyone. No rival beats your product. No customer rejects it. The market you were in simply stops being a market, the way a room stops being a room when the walls come down and it becomes part of the house.
Economists call it platform envelopment, and the thing I find quietly unsettling about it is how little it asks of the firm doing the enveloping. Eisenmann, Parker and Van Alstyne, who named the mechanism in 2011, were explicit about this.¹ In platform markets, they wrote, entrants generally must offer revolutionary functionality to win share. Envelopment is "a second entry path that does not rely upon Schumpeterian innovation." (Schumpeterian, after the economist Joseph Schumpeter: breakthrough innovation, the genuinely new product that displaces the old.) The enveloper does not need a better product. It needs an existing relationship with your customers, and a bundle that folds your function into something they already use. Then the separate purchase, the small decision that constituted your market, becomes unnecessary. The market does not get crowded, the way markets do when competitors arrive. It ceases to be a separate market.
The difference matters more than it first appears. Against a rival with a better product, the strategy books have real advice. Differentiate. Build barriers. Move again. Against envelopment, that advice goes strangely quiet, because every piece of it is addressed to a customer who is deciding, and the enveloped customer never faces the decision. The function arrives inside the subscription they already pay for. There was no moment at which they chose.
The clearest live example sits inside the software many of us open every morning. In November 2023, Microsoft placed automated meeting notes inside Teams as part of Microsoft 365 Copilot. Meeting transcription had been a standalone market, with independent firms built on exactly that one function. The bundle did not argue with them. It just appeared, inside an enterprise agreement, next to the calendar.
And here is where I have to be honest about what the evidence actually shows, because the story could end with the independents dead and it does not. Otter.ai announced in March 2025 that it had passed 100 million dollars in annual recurring revenue. In coding, GitHub Copilot reached general availability in June 2022, before Cursor even existed,² and Cursor grew anyway, through a 2025 in which Anthropic, OpenAI and Google all shipped coding agents of their own. In search, OpenAI launched ChatGPT search³ and Perplexity's reported revenue more than doubled afterward. Three categories, incumbents shipping the same function into enormous user bases, and three independents growing. The figures are the companies' own and cannot be audited, which is its own story, but the direction is consistent. Bundling does not reliably kill.
So what separates the firms that survive envelopment from the markets that dissolve? The pattern in the cases I studied is almost embarrassingly simple: the survivors kept a direct relationship with their users. Their customers came to them, knew them by name, and kept facing, and making, a choice. The bundle competes for a decision. Where the decision still exists, a better product can still win it. Where the bundle has made the decision disappear, there is nothing left to win.
There is also a Chicago-school objection that deserves its full weight, and I do not think it is defeated. Bundling frequently expands markets rather than destroying them. A function folded into a suite reaches millions of people who would never have bought it separately, and calling that harm requires more than a strategy framework's discomfort. Competition law has wrestled with exactly this for decades... the tying claim in United States v. Microsoft was sent back to be judged under the rule of reason rather than condemned outright,⁴ and in July 2026 the European Commission fined Google 460 million euros for self-preferencing in search.⁵ The line between a bundle that serves users and a bundle that forecloses a market is genuinely hard to draw, and I respect the economists who draw it differently than I would.
From where I sit, the strategy question and the competition question are the same question wearing different clothes. A firm asking "how do I survive the bundle" and a regulator asking "when does the bundle foreclose" are both really asking who still faces a choice, and about what. The framework this series keeps returning to cannot ask that question, because in every market it studied, the choice was simply there, as unremarkable as air.
The full argument, with the cases and their sources, is in the paper. You can read it here.
The smallness of it is the point. No battle, no collapse, no moment to point to. Just a function, appearing one morning inside something you already pay for.
Notes
Thomas Eisenmann, Geoffrey Parker and Marshall Van Alstyne, "Platform Envelopment," Strategic Management Journal, vol. 32, no. 12, 2011.
GitHub, "GitHub Copilot is generally available to all developers," 21 June 2022.
OpenAI, "Introducing ChatGPT search," 31 October 2024.
United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001).
European Commission, "Commission fines Google €890 million for breaches of the Digital Markets Act," 23 July 2026.
About the author
Bilal Farhat is a competition economist and independent scholar. This paper is written in a personal capacity. The views expressed are the author’s own and do not represent those of any organization or firm with which the author is, or has been, affiliated.
Contact: bilal.h.farhat@gmail.com