Michael C. Vincent: Mark, quick question before we get into this — what platform did you use most this week? Just off the top.
Mark Delaney: Uh — WhatsApp, probably. Family group chat, work stuff, all of it goes through there.
Michael C. Vincent: And when did you last consider leaving it?
Mark Delaney: I mean... never, honestly. Not really. Which is weird because I don't even think it's the best one.
Michael C. Vincent: That's the whole episode right there. WhatsApp's value is attributed to user-count scale — not to its encryption innovation, not to any feature advantage. Telegram launched with stronger privacy claims and never displaced it. Same mechanism on paper. Different outcome entirely. And that gap — between what people say creates a moat and what actually does — that's where competitive analysis tends to quietly collapse.
Mark Delaney: Oh, that's — wait, so the thing that kept you there wasn't the product. It was everyone else being there. That's the demand-side feedback loop doing actual work, not a feature team.
Michael C. Vincent: Now you're naming it. A network effect, properly defined, is when the value of a product to each user increases as more users adopt it. The self-reinforcing loop — more users, more value, more adoption — that's a very specific mechanism. Facebook gets called the canonical example of it. But CRV's investor analysis flags that most startups claiming that label are describing something structurally weaker and hoping no one checks.
Mark Delaney: So what does a real one actually look like — not the pitch, the actual thing.
Michael C. Vincent: The phone. 1880. Pick it up — there is no one to call. Every telephone on earth is useless until there are two of them. That's it. That's the whole idea. The hardware never changed. The value came entirely from the other subscribers.
Mark Delaney: Right — and that's demand-side, not supply-side. Like, the phone company didn't get cheaper to run with each new customer. It got more valuable to every existing customer. Those are two totally different things.
Michael C. Vincent: That distinction is the formal frame — demand-side economies of scale, sometimes called network externalities. Your benefit rises with compatible users. Nothing to do with what it costs to build the thing.
Mark Delaney: Okay but — wait, so is there a point where that just... tips? Like where the flywheel starts spinning on its own without anyone pushing it?
Michael C. Vincent: Critical mass. That's the threshold. Cross it, and the value is self-sustaining — new users arrive because the network is already valuable, not because someone convinced them. Before it, you're pushing uphill. After it, the hill is behind you.
Mark Delaney: And that's — uh, I mean, that's where the moat gets real, right? Because once you're past that threshold, a competitor isn't just racing you on features. They're starting at zero users against your million. The network itself is the advantage.
Michael C. Vincent: The largest network attracts more users, which raises its value, which makes displacement harder regardless of what the challenger builds. The fax machine is the same story — a fax to no one is a paperweight.
Mark Delaney: Which is why I never left WhatsApp. Not a feature decision. Just — everyone I need is already there, and that fact alone does more work than any product update ever could.
Michael C. Vincent: But that's where the clean version quietly falls apart — because not every network effect is the same animal. There's a structural split that changes everything about how you read a market. Direct effects, indirect effects. People use those terms interchangeably and they shouldn't.
Mark Delaney: Wait — like, mechanically different? Not just, uh, stronger or weaker?
Michael C. Vincent: Structurally different. A direct effect — same-side, one group — is Facebook. Each new friend you add increases the platform's value for every existing member. It's one population feeding itself. An indirect effect is Visa. Two completely separate groups: cardholders on one side, merchants on the other. More merchants makes the card more useful to you. More cardholders makes accepting the card more valuable to a merchant. Neither side benefits from more of itself — they benefit from growth on the other side.
Mark Delaney: Oh — so Airbnb is the same thing. More hosts, guests get more choices. More guests, hosts get more bookings. The two groups are kind of... pulling each other up.
Michael C. Vincent: Exactly so. Now — here's the part worth pausing on. If indirect effects are winner-take-all the way direct effects supposedly are, explain Mastercard.
Mark Delaney: Ha — yeah, right. Visa and Mastercard have both been sitting there for sixty years. Same basic network, same merchant-cardholder structure. Neither one killed the other.
Michael C. Vincent: Sixty years, successive waves of technology, and the duopoly holds. Which means — and this is the part the winner-take-all framing doesn't survive — market structure, multi-homing costs, platform design, those determine the outcome. Not the network effect alone. You reach into your wallet at a checkout counter. The merchant takes Visa. Also Mastercard. You could hand over either card. Both networks are right there, simultaneously, and the world didn't end.
Mark Delaney: That's — I mean, that's weird when you sit with it. Because everything we said earlier about the moat being the network — it's still true for both of them at once. That's not supposed to be possible if the logic runs the way people say it does.
Michael C. Vincent: And that's before we get to the question of whether Facebook's displacement of MySpace was even the same mechanism — because that one, it turns out, wasn't really about scale at all.
Mark Delaney: Wait — MySpace had the bigger network. Like, more users, more established. And still lost. That breaks the whole thing we just said about the moat.
Michael C. Vincent: It does. And a 2011 PACIS paper looked at exactly this — concluded Facebook didn't outgrow MySpace in raw user count. It changed the topology of the network itself.
Mark Delaney: Topology — meaning, like, the shape of how people connected? Not just how many?
Michael C. Vincent: The structure of who reaches whom. Facebook opened its platform to third-party developers. That shifted which users could be connected and how — not the node count, the architecture underneath it. MySpace had more nodes. Facebook had better wiring.
Mark Delaney: And the average person switching didn't — uh, I mean, they weren't thinking about developer APIs. They followed their friends. That was the whole decision.
Michael C. Vincent: Social topology was the operative mechanism. The user didn't choose network-effect logic. They chose their actual social graph — real names, real photos — over a platform full of band pages and bots.
Mark Delaney: Which is — okay, so that's a design win disguised as a scale win. And the MIT Sloan research is basically saying that's the pattern — winner-take-all pronouncements ignore how much variation exists in how strong these effects actually are across platform types.
Michael C. Vincent: And most claimed data network effects? Same problem. A genuine closed loop: usage generates data, data improves the product, improved product drives more usage. Most startups describing a data moat have a large dataset — no feedback loop. That's a data scale effect. Not the same animal.
Mark Delaney: So the label travels without the mechanism. You call it a network effect, investors hear moat, but the loop was never actually closed. That's — man, that's the part that should make people nervous.
Michael C. Vincent: That's the one that keeps sitting with me. WhatsApp and Telegram — same value proposition on paper, stronger privacy claims on Telegram's side, launched two years later into a world where WhatsApp already had 450 million users by 2014. And Telegram never displaced it. The mechanism was identical. The outcome wasn't. And I don't think the research fully resolves why.
Mark Delaney: I mean — because if the moat is the network, and Telegram had the same network mechanics available to it, then... uh, what was the actual barrier? Was it just that WhatsApp got there first and the critical mass was already locked in? Or was it something else we don't have a clean label for yet?
Michael C. Vincent: That's where I'd be careful landing on any answer. A company that mistakes a data scale effect for a network effect misreads its own defensibility — and misreads the true barrier facing anyone coming for it. Telegram probably looked at WhatsApp and saw a user-count problem to solve. Maybe the actual barrier was something closer to what Facebook had over MySpace — not the number of nodes, but the shape of the connections. Which is a very different thing to compete against. You see, you can build a better product. You cannot easily rebuild someone's actual contact list.
Mark Delaney: Yeah — and I don't think most investors or founders can tell the difference until someone actually tests the moat. Like, it all sounds the same in a pitch. Network effect, self-reinforcing loop, critical mass. But whether it's real or just a story about the data you happened to collect — that only comes out when a Telegram shows up.
Michael C. Vincent: And sometimes not even then. The moat holds and nobody learns anything. Well — I think that's the honest place to stop. The gap between what the label promises and what the mechanism delivers is real, and most of the time it stays invisible right up until it doesn't.
