Chris Dixon makes the point that to assuage developers, platforms should aim to achieve predictability.
To mitigate subsumption risk, the platform should give developers predictability around the platform's feature roadmap. Platforms can do this explicitly by divulging their product roadmap but more often do it implicitly by demonstrating predictable patterns of feature development. Developers and investors are willing to invest in the iOS platform because - although Apple will take 30% of the revenue - it is highly unlikely that Apple will, say, create games to compete with Angry Birds or news to compete with The New York Times. Similarly, Facebook has thus far stuck to "utility" features and not competed with game makers, dating apps, etc.
I would argue that only one thing is entirely predictable about closed platforms: they will do whatever is in the business's interest. When building your business on someone else's platform, you should always assume that if it is in the platform owner's interest, they will shut you down tomorrow with no warning.
So, you need to do two things at the very least:
- Ensure your goals and the platform owner's goals are always in harmonious agreement.
- Have a backup plan for what happens if your access to the platform is revoked tomorrow.
People seem to keep learning this lesson over and over again... with Facebook, with Twitter, and these days with Apple...
If you read this far, you should follow me on twitter here.