Talking about it is like nailing a jelly on the wall
By Rajiv Jayaraman
11 Dec 2008
Yahoo plans to launch a mail app Platform. The number of iPhone apps developed for the iPhone mobile app platform crosses 10,000. Mozilla served their 1 billionth add-on download for their browser extension platform. Just in the last one month, we have witnessed some amazing developments in the Internet and mobile economies. More and more companies are converting their products into ‘platforms’ for third-party developers to build applications on and more and more app developers are using the opportunity to build fascinating apps on the ‘platforms’ to reach millions of users.
All this sounds great. But do we really understand what ‘Internet/mobile platforms’ are? To me, trying to talk about them is like nailing a jelly on the wall. The term platform conjures up myriads of, often incompatible, ideas in our heads. The problem arises primarily from the fact that this term is heavily overloaded. It is somewhat akin to saying ‘prego’ in Italian, a word that is used to mean many different things in many different contexts. Another factor that adds to the confusion is the fact that the whole realm of Web computing is evolving at such a rapid pace that it is often difficult to keep up with the latest in the Internet industry.
Marc Andreessen’s blog on Internet platforms is a must read for anyone who is even remotely interested in the understanding the different kinds of Internet/mobile platforms (http://blog.pmarca.com/2007/09/the-three-kinds.html). In this post, I will use his framework to further explore the implications of using different platform models.
Let’s start by first defining what a platform is. A ‘platform’ is a system that is programmable and hence, can be customised by third-party developers for mutual economic benefit. Why are companies vying hard to establish themselves as Internet platforms? To build an ecosystem around their technology and feed a positive feedback loop. Why is this important? The company that has the biggest ecosystem finally wins in the ‘winner-take-all’ dynamic that is common on the Internet these days.
Let’s now look at a few Internet platform models that exist today. The models differ basically on the location where the developer’s code is executed and the complexity of programmability.
The Google APIs model
This is the most common form of Internet platform.
Highlights:
• Developer’s code resides and gets executed in their server
• Easiest model to program from the developer’s point of view
• Developer’s app is not part of the platform user experience
Suppose I want to build a dashboard, like the one on Google Analytics. I can use the Google Visualization Web Services APIs to create graphs and charts for my application. In this case, my code resides on my server. I simply call the APIs and expect data in return.
Other examples: E-bay, Flickr
Benefits for platform companies
• Hook developers to your APIs and make them part of your sub-system.
• Leverage the innovations that are built on top of the APIs
Benefits for app developers
• Avoid re-inventing the wheel
• Access to tried-and-tested technical tools
The Firefox ‘plug-in’ model
In this model, third-party app developers enrich the user experience of the platform users by extending the functionality of the platform. Firefox and Facebook are shining examples of this model.
Highlights
• The third-party developer’s code resides and gets executed outside the platform.
• The app, however, becomes part of platform user experience.
• Scalability is an issue that the app developer has to deal with.
Benefits for the platform company
• Third-party apps enrich user experience
• Cater to the needs of a wide audience at a very low cost and in a very short time.
Benefits for app developers
• App developers have the opportunity to build a viable business. They instantly get access to millions of users through the platform. The cost of scalability and availability, however, are borne by the developer.
• VCs have set-up funds for such app developing companies. Kleiner Perkins’ iFund is a fund created for iPhone apps. Facebook, along with Accel Partners and Founders Fund, have created a fund for Facebook app developers.
The Salesforce model
In this model, the app developer builds an app, uploads the app to the platform and runs it online. For instance, Salesforce lets its users upload apps to its platform to extend and customise Salesforce features. Other prominent examples in this model are Google Apps, Second Life, and Amazon EC2.
Highlights
• In this model, the third party application actually resides and runs inside the platform.
• The platform company assumes the risk of executing arbitrary third-party application code.
Benefits for the platform company
• Utilise spare capacity to allow developers to build applications on your system.
• Leverage the meta-information that is available to you through the apps that run on your system for ads.
• Assume the cost of scalability and availability, so that user experience is not compromised
Benefits for the app developer
• App developers needn’t assume the cost of scalability and availability.
In my view, platform economics is here to stay. In a flat world, product companies need to stay ahead of the innovation game and one way of doing that is by crowd-sourcing. CTOs, business managers and developers need to take this model seriously. Product companies need to think about how to collaborate with third- party app developers to keep their offering fresh and innovative. App developers can now tap into funds that are available for developing apps for platforms. There has been no better opportunity to reach millions of users without the cost of acquiring customers.
Rajiv Jayaraman is the founder and CEO of KNOLSKAPE Solutions, a Singapore-based educational software company. He has an INSEAD MBA and has worked in the grid computing team at Oracle, US. He is currently leveraging the power of Internet computing to develop products for the education market.
