Network Effects

Network effects — the property that a product’s value to each user rises as more users join — is one of the most heavily-trafficked moat concepts in Kyle’s corpus, and several sources he reads treat it with notable skepticism about overuse. Metcalfe’s Law is the recurring lens: in The Idea Factory the Picturephone’s failure is framed explicitly through Metcalfe, where a minuscule network had vanishingly small per-user value, and in The Founders Bill Harris invokes Metcalfe’s Law to argue that “true networks are a naturally monopolistic business.” The Power Law — Venture Capital & the Making of the New Future applies the same law to eBay and Yahoo, and casts VCs themselves as relentless network-cultivators (the Saxenian “weak ties” thread) — while puncturing WeWork, whose “alleged network effects were weak, at best” (adding tenants on Park Avenue doesn’t improve the Fifth Avenue experience).

The skeptical thread is sharpest in The Mysteries of an Economic Engine - Research, drawing on Bill Gurley’s “All Revenue is Not Created Equal”: network effects are “perhaps the strongest moat of all,” but “discussed way more than they exist” — many supposed network effects are merely economies of scale, far weaker; the key variable is the decay rate of value of the incremental user; and genuine strong-form network-effect companies (Microsoft, eBay, Skype, Google AdWords, Facebook in their prime) are rare, and when they exist are leading 10X-revenue candidates. The concept also appears in its growth-tactics form in Product Led Growth — The End User Era (viral products like Calendly “targeting consumers inside an enterprise”), in Working in Public as open-source’s “cornucopia of the commons,” in the Todd McKinnon — Creating and Defining a New Market Category interview (Okta’s standard-setting ambition resting on network effects and positive returns to scale), and in Oh Snap, It’s Here (why a generationally-bounded network is invisible to those outside it).

Context: A network effect exists when each additional user increases the value of a product for existing users — direct (telephones, social networks) or indirect (marketplaces, app-store platforms). Metcalfe’s Law, the canonical formalization, holds that a network’s value scales roughly with the square of the number of connected users.

Where this appears

• The Idea Factory — the Picturephone’s failure framed via Metcalfe’s Law; a small network has vanishingly small per-user value.
• The Founders — Bill Harris invokes Metcalfe; “true networks are a naturally monopolistic business.”
• The Power Law — Venture Capital & the Making of the New Future — drives eBay/Yahoo; VCs as network-cultivators; WeWork’s network effects exposed as “weak, at best.”
• The Mysteries of an Economic Engine - Research — the skeptical core (via Gurley): the strongest moat, but “discussed way more than [it] exist[s]”; decay rate of the incremental user is the key.
• Product Led Growth — The End User Era — viral/network products like Calendly as “consumers inside an enterprise.”
• Working in Public — open-source’s “cornucopia of the commons.”
• Todd McKinnon — Creating and Defining a New Market Category — Okta’s standard-setting ambition rests on network effects and positive returns to scale.
• Oh Snap, It’s Here — why a generationally-bounded network is invisible from outside it.
• Martin Casado and Bill Gurley on X (2026) — the term’s lineage from Rohlfs (1974) to Rochet & Tirole (2000), plus Arthur’s increasing returns.

Where the term came from

In October 2026 Martin Casado posted a short intellectual history of the term (archived here), arguing that “network effect” now means several different things because the business press picked it up at different points along the way:

• 1974 — Jeffrey Rohlfs (Bell Labs): a simple model of communications networks built on subscribers and their interactions.
• 1985 — Michael Katz & Carl Shapiro generalize it, which in Casado’s view loses the idea of a network and becomes “largely tautological and useless.”
• 1994 — Stan Liebowitz & Stephen Margolis criticize Katz and Shapiro and start classifying network effects properly. Casado: “this is a great paper, I recommend reading it.”
• 2000 — Jean-Charles Rochet & Jean Tirole split out Two-Sided Marketplaces as their own category.
• The result, in his words: “the business jargon tower of babel.”

Bill Gurley quote-tweeted it agreeing with Tren Griffin’s reply: “You can’t leave out Brian Arthur.” Griffin’s lineage adds W. Brian Arthur’s 1989 work on increasing returns and positive feedback in technology adoption, plus Robert Metcalfe (George Gilder named Metcalfe’s Law in 1993) and Andrew Odlyzko’s critiques. Gurley linked Arthur’s “Increasing Returns and the New World of Business” (Harvard Business Review, 1996). Casado’s reply sharpens the history: Arthur’s positive-feedback work came before both his use of “network effects” and Katz’s paper, and Arthur later adopted the term more narrowly, as one cause of those feedback loops.

This fits Gurley’s own warning above that network effects are “discussed way more than they exist.” Part of why they seem to be everywhere is that the term has no single agreed definition.