The Problem with Developing Around a Platform

Some applications and services become so useful and popular that they turn into platforms: people develop other applications and services around them. Some examples of platforms of this kind that many people know are:

• Dropbox: This application/service, created to back up and distribute files, has become the means many applications use to synchronize and store their databases and libraries.
• Google Reader: Few services or applications have attained the almost-standard status of this application. Most RSS clients for Mac and iOS work through integration with it, which is very convenient: whenever you want to change clients, you simply enter your account and all feeds synchronize automatically.
• Twitter: What began as a social network has become a platform around which applications and even entire companies devoted to marketing and public relations have been developed.

Developing around platforms has many advantages; in fact, it could be said that everyone wins:

• The platform wins because it becomes more popular and grows its user base.
• Developers win because they no longer need to build the infrastructure they use through the platform.
• Users win because they can combine services, make their workflow more efficient, and have more ways to use the platform.

Everything seems very favourable, but there is a problem: developing around a platform is a marriage. Developers must stay with the platform through good times and bad, submit to its changes, and worry about constantly updating their product at the pace at which the platform changes. Developers become dependent on the platform; they share its success and failure, and may even be betrayed by it. Most products communicate with platforms through APIs, Application Programming Interfaces, and depend on both the APIs the platforms provide and their goodwill. The day Twitter decides it no longer wants to share its popularity with others, and that complete access to the service will be available only through its official applications or website, it will be the end of many companies. A popular application such as Tweetdeck or Hootsuite cannot tell its users that it will now work with Orkut instead of Twitter; users are there for Twitter.

This problem leaves me with a question: what is the benefit-risk relationship in developing around platforms like these? There appear to be many benefits, but also a very great risk: total dependence on a third party whose motivation is to make its platform successful with, without, or despite developers…