Call something a platform and its valuation multiple doubles. That's why every pitch deck claims to be one. But a platform isn't a label — it's a structure: more participants on one side make the product more valuable to the other side, and leaving gets harder the longer you stay. Get both and you compound. Get neither and you're a commodity middleman one price war from irrelevance.
The durable platforms engineered both deliberately. Apple's App Store pairs a cross-side loop (more apps attract more users, more users attract more developers) with deep switching costs (your apps, purchases, and identity all live there). That combination, not the app catalog, is the moat.
A platform without switching costs is just an aggregator waiting to be undercut.
Are you actually a platform?
You have a real platform — not just a marketplace — when both of these are true:
- ◆Each side genuinely attracts the other: more of one makes the product better for the other (a cross-side network effect).
- ◆Leaving gets costlier over time — data, identity, integrations, or workflows that compound the longer a participant stays.
- ◆You can grow the loop faster than a well-funded rival can copy the surface features.
The method
Solve the cold-start on one side first
A two-sided market is worthless until one side shows up. Subsidize, seed, or hand-build the harder side (usually supply) until the loop can spin on its own. Airbnb photographed listings by hand; the loop came later.
Engineer the cross-side loop, then protect it
Make each new participant on one side visibly increase value for the other — and instrument it so you can see the loop turning. Once it spins, your job is to defend it, not to tax it so hard early that you choke the side that's still fragile.
Build switching costs that compound, not cages
The good kind of lock-in is value that accrues by staying: history, reputation, integrations, data. WeChat became un-leavable because your payments, identity, and social graph all lived there. Punitive lock-in (hard to leave, no reason to stay) just invites a challenger.
Own the rails, monetize gently
Platforms make money by sitting in the flow — a small, fair toll on transactions you make possible (Google's ad auction, Apple's cut). Take too large a cut too early and you hand a wedge to anyone willing to undercut you. The auction, not a rate card, keeps the toll honest.
Defend the loop as you scale
Maturity invites two failure modes: over-taxing the participants who made you, and letting a rival open a parallel loop in a niche you ignored. Keep reinvesting in the side that's hardest to attract, and watch the edges where a focused competitor could start their own flywheel.
- Adding more participants doesn't make the product better for anyone — it just adds inventory.
- A participant could leave tomorrow and lose nothing they value.
- Your "network effect" is really just scale or a brand, not each side pulling the other.
- You're taxing the ecosystem hard before the loop can stand on its own.
- Your moat is a feature set a funded rival could clone in a quarter.
Done wrong vs. done right
WeWork as a "platform"
WeWork sold itself as a tech platform, but adding a member didn't make the space more valuable to other members, and leaving cost nothing. Underneath was a real-estate sublease business — no cross-side loop, no switching cost. The $47B story collapsed to ~$8B the moment a filing forced the structure into the open.
Apple's App Store ecosystem
More users pull more developers and vice versa, while your apps, purchases, and identity make leaving genuinely costly. Apple monetizes the flow with a cut rather than a wall — a loop that compounds instead of a cage that resents you.
Both called themselves platforms. Only one had the structure underneath — a loop that gets stronger and stickier with every participant.
Will your platform actually lock in?
Two structural questions decide whether you compound like a platform or compete like a commodity: do the sides pull each other, and does leaving get harder?
Question 1 of 4
Do more users on one side make the product more valuable to the other side?
Where this plays out
Apple's App Store ecosystem
Cross-side loop plus deep switching costs — the toll booth that compounds.
Done rightGoogle's ad auction
Monetizing the flow with an auction, not a rate card.
Done rightWeChat's super-app
Payments, identity, and social all in one — un-leavable by design.
Done rightAirbnb's trust platform
Solved the cold-start by hand, then let the loop spin.
Done rightGoogle buys Android
Owning the rails so no one can stand between you and the user.
Done wrongWeWork's collapse
A sublease business dressed as a platform — no loop, no lock-in.