The industry trades are rife with discussion of Open Source business models. The debate rages on about whether they are valid and sustainable or whether there are self-cannibalizing and doomed to failure. But to me, the vast majority of this debate misses the key point by focusing on only two valid open source business models, to-wit support and dual licensing schemes. The most significant impact of Open Source is rarely discussed. I can’t really give the IT press too hard a time for focusing on the support and dual licensing business models, as those models are the only two the analysts want to talk about. The analysts’ obsession reflects the practical reality that in both situations there are traditional business metrics to assess and more importantly, there are firms willing pay consulting fees to advise them on those models (and to promote them as well!). But to look only at those two business models is to miss the entire forest for the sake of two trees. The giant redwood of the support business model is RedHat, who has proven soundly that given enough scale, support can constitute a viable business. Therein also lies the problem: scale. If you want to run your firm on a pure support model, you better have access to a significant installed user base. The dependence on the size of the user base puts companies who rely on the support model in a bind. They must at the every least maintain market share to survive, and if they want to grow, they had best be innovating. RedHat, for example, has created market share by giving something away, but as time passes and other players enter the market RedHat will have to invest to stay ahead of the competition and compete on quality. Experience says that keeping ahead of the pack will at some point become increasingly expensive, while conversely support prices are likely to come down as competition increases. Dual licensing models abound – this is more a thicket than one single tree. The model is premised on the idea of a firm giving away one version of the product while holding back another version under a proprietary license. The proprietary version is typically more full-featured or robust and is often labeled the “enterprise” version. The vendor will frequently round out the offering by adding on installation, support and patch management services. Some even offer the services aspect to the users of the open source version of the product on a per incident basis, which enables them to create their own support-based revenue stream built round the open source product. The problem companies face with the dual licensing scheme is two fold. First, you have the potential to become a victim of your own success. If your open source version takes off, it will attract a community of developers who will begin to innovate and expand the product. They will look at the features in the proprietary version and say “hey that’s cool, let’s figure out how to do that.” They will cannibalize your innovation and they will do it at even less than a fraction of the company’s costs. Therefore, by definition you only want it to be so successful, which creates an inherent conflict of interest: You are competing with your own product. Second, courtesy of the nature of open source licenses, if the community develops something really useful on their own, you are prohibited from importing it into your proprietary product, else you may lose the proprietary rights. The innovation drain, in other words, flows only one way. So, to strain my analogy even further: Given these rather serious shortcomings with the two trees discussed, where’s the potential in the forest? The biggest potential in any forest lies in the totality of its parts; not simply the trees, but the undergrowth and the creatures that flourish in the shadows. While it may initially sound like I am drifting towards some egalitiarian hippie manifesto, bear with me through an example and the point will become clearer. Take as one example the Mambo Open Source Project. Mambo is a Content Management System. It is a complex application with a very large installed user base. Mambo is run by a non-profit Foundation and all the people working on the project are volunteers. There is no support model, no dual licensing model. This is pure open source. Due to the complexity of the application and the size of the user base, Mambo creates opportunities for business to spring up in its shade. Small developers offer templates for the system, web design studios sell integration and design services, programmers sell applications which extend the usefulness of Mambo, writers and trainers sell documentation and training services, and so on. To borrow some nomenclature from Gartner, Mambo has created an Open Source eco-system. Or put another way, Open Source is a catalyst in the classic chemical sense, that is, “a thing that initiates or accelerates a reaction without itself being consumed or destroyed.” Looking at the long view, this is where the value is: Open Source creates wealth by serving as a catalyst for the birth and growth of services and related products. Support and dual licensing models are only two small parts of the long term impacts of this movement. Originally published in the Bangkok Post, 19 January 2006.