Last time: “Killer Apps,” in which I claimed that it was possible to engineer an application that had good characteristics for success within its chosen market, rather than just having to count on “built it and they will come.”
Now: What are the ways that revenue can be produced from a Killer App on an open-source platform?
[This is an excerpt from a Lisp conference talk I gave in 2002.]
If the platform and killer app are free, how do you make money?
Service: Redhat provides technical support and training on open source Linux. They also provide the “service” of a commercial packaged version of Linux. But IBM is the real winner in servicing and commercializing Sun’s J2EE. In 1993, as Lou Gerstner took over as CEO, IBM lost more than $8 Billion. By last year, they had an annual profit of nearly the same amount. The big difference is that they moved from selling a melange of platforms with 27% of revenue from services, to now deriving 41% of revenue from services. Much of that is on Sun’s J2EE platform. Even in the new new economy, they have been rewarded by having their market capitalization jump from less than $30 Billion to nearly $190 Billion over this same nine year period.
Multi-Tier products: How do you extract revenue from different markets that are willing to pay different amounts? Some people won’t pay anything, and some business will pay a fortune. The answer is to have different products. The usual software practice is to have “personal” and “enterprise” versions of the same product. But think about huge food companies that have one brand for buy-in-bulk hypermarkets, and an organic brand for people willing to pay more. Again, IBM provides an example. Sun provides an open source reference implementation of a J2EE server, and there are other free versions. Macromedia’s commercial version is $900. But IBM’s version costs $35K per CPU. There are a bunch of other commercial versions at various prices. They all conform to the same platform spec.
But the different tiers don’t have to be of functionally equivalent products. Maybe the free killer app for a P2P platform does file sharing or instant messaging, but one commercial product from the same family does distance learning while another one trades in derivatives.
Secret sauce: Maybe the apps are more or less free, but some can make use of an additional piece, such as an identity certificate, of which we are the sole provider. We sell identity. Of course, identity certificates aren’t valuable without a context, so there still has to be an application. Other examples might involve access to part or all of a shared database.
That’s it. Then we had our fishbowl discussion.
Start of nine part series.
