When Does Shipping on Someone Else's Runtime Become a Single Point of Failure?
The 30% take rate gets all the speeches. It's the part you can model in a spreadsheet. The part that actually ends studios is subtler: you built your whole business on a platform whose rules you don't write, and one policy change can move your revenue line more than any feature you ship. Here is how to know when "distribution" became "dependency."
Q: What is the real risk — the 30% or the rules?
The rules. The fee is a known, negotiable, modelable cost. The risk is the platform changing discovery, payout terms, content policy, or simply deprioritizing your category — none of which you voted on. A runtime that decides what your players see, how you get paid, and whether your game is allowed isn't a distributor; it's a landlord. And landlords can renovate the building around you while you're still paying rent.
Q: How do you measure concentration risk?
One number: what share of your revenue or players comes from a single platform? If it's above roughly 60%, you don't have a distribution strategy — you have a single point of failure wearing a logo. The danger compounds because concentration feels like success: the platform rewards you, you lean in, and three years later you can't leave without losing most of the business. Measure it quarterly and treat a rising number as a warning light, not a trophy.
Q: Can you de-risk without leaving?
Partly, and that's usually enough. Own the relationship the platform doesn't: your own user accounts, your own billing where allowed, your own communication channel — email, an owned community. Keep your game logic portable so a second runtime is a config change, not a rewrite. And diversify the top: never let one catalog exceed your concentration ceiling. You're not trying to leave — you're trying to be able to.
Q: What does a platform owe you, and what it doesn't?
It owes you what the contract says, and almost nothing beyond it. No guarantee of discovery, no guarantee of continued category support, no obligation to warn you before a rule changes. The teams that survive platform shifts are the ones who read the terms as a lease, not a partnership, and built their exit before they needed it. Dependency isn't evil — unexamined dependency is.
| Lever | What it buys you | What it doesn't |
|---|---|---|
| Owned accounts | a direct relationship | platform discovery |
| Portable logic | exit is a config | an instant zero-downtime move |
| Diversified top | no single point | maximum reach on one |
| Billing ownership | payment control | the platform's warm base |
