Open Source Economics: Money, metrics, and those that work
It is always interesting to read discussion about what economic impact OSS has, or will have. It goes from the wild enthusiasm of Matt Asay to the more moderate view of Savio Rodriguez. As readers of this blog know, I enjoy writing about this kind of things, and would like to provide a few comments on what is measured, and what really would be important to do.
Beyond Measure by Ian Boys
Savio cites IDC, and calculates that OSS software will reach in 2011 sales of 5.8B$, or 1.8% of the market. On the other hand, Gartner says that in 2008 25% of the software market will be OSS based, either through internal development or through external OSS providers.
Who is right? The answer is of course in testing the hypothesis that is hidden under every measurement effort: that what is measured is a reasonable proxy for the true variable that we want to know. IDC (I have to guess from Forrester press release and Savio’s post) measures OSS software sales, without hardware and services. So, of course, Savio takes RedHat, MySQL, a bunch of others and struggles to reach even 1B$. Is it realistic? Maybe not; while it is true that selling software is one of the possible business models, it is difficult to ignore the economic impact not only of service-based models, but of all those users that internally are using OSS and maintaining it, thus getting an economic benefit and without appearing on anyone’s radar. It reminds me of those reports of just a few years ago, that found that Linux was used in less than a few percent of all servers sold; the problem was that this sampling ignored servers sold without operating systems, or those with non-commercial version of linux installed.
Savio’s comments (and Roberto’s too) are however spot-on for the main problem of the OSS market, or the lack of a commercial channel with credible strength. I enjoyed a lot an evening organized by our local industry association, debating pros and cons of OSS for companies. I heard several nice (and some bad) stories of usage of OSS, and found that ALL the companies that are OSS adopters found about the software and the solutions by themselves. While proprietary solutions (like Microsoft ones) are promoted by tons of companies (I can find more or less 50 in a range of 5 kilometers from my office) there are less than 200 OSS companies in all of Italy. If we want to grow the 1.8% market Savio reports, we should create a channel of companies (maybe through service franchising). At the moment, companies interested in OSS (and there is a very high percentage of those) are forced to go alone, and in many cases spend too much money before giving up. The same companies are quite happy to pay for someone that helps them, but up to now you must be customer of Atos Origin, Engineering, IBM Global Services or Accenture to enjoy such help. Horizontal companies like OpenLogic or SpikeSource are starting to address this problem, but we need smaller entities at the local level, that can provide the necessary handholding and avoid the pitfalls.
Reply