ITGumbo: spicing IT up

For proprietary software companies this task is drastically easier as the nature of their development and distribution model isn't built upon open [even free] exchange of the product in question. Since closed source products can either [legally] be purchased or downloaded using a trial-based system, the distribution channels are already narrowly constructed. Open source, on the other hand, enables products to be downloaded and used as-is for as long as the user desires. While customers who end up paying for some form of extended product feature set, support or indemnification are of course accounted for, they can be a minority when compared to the total number of users.

Some counter that if an open source product is implemented within a production environment, odds are that the implementing party will require some sort of support. Frankly, I agree with this assertion fully. However, there are a whole host of open source products which continue to see real-world use without a dime of support having been purchased. This is especially the case with developer-centric tools and applications. The Eclipse Platform is one example, the Spring Framework is another. Additionally, there is a sizeable number of individuals and companies who integrate open source products into custom built architectures where they maintain responsibility for providing support for the use of them. It's quite impractical to wave off such uses as insignificant by not taking steps to document them in some form.

I am of the perspective that there is a gold mine of information about the non-paying user which needs to be harvested in order to better understand and visualize the concept of an open source community. Commercial open source companies in particular should be taking steps to get their hands on this type of data for the purposes of identifying, analyzing and integrating it into their competitive strategies. The possibility of combining more detailed information about individual non-paying users with other passive measurements (bug fix ratio, mailing list activity, source code commit frequency, etc.) and customer data makes for several interesting and highly useful scenarios.

For example, when I'm briefed by an open source company one of the topics I inquire about is the recent growth trends exhibited by their user base. I ask this because it remains critical to attaining a better picture of how well their entire operation (business, product development and community) is growing. More is needed apart from how many new customer wins they've had the last couple quarters (even though this remains gravely important, too). At the end of the day, signs of stagnant community are non-positive indicators for the growth prospects of the company in general. The ability to produce objective proof that the community branch, of the tree that is an open source business model, is robust and vigorous goes quite a ways towards vouching for long term health.

Along these same lines, the more detailed data which can be produced the more convincing the argument for a given value proposition. Consequently, I am of the perspective that more efforts should be invested in figuring out how to best collect variations of this data. A free-form, product registration-esque system that enables users to enter as much or as little information about themselves and the intended use of a product would do well towards that end. The registration shouldn't necessarily proceed/protect the downloading of a product, but should be an entirely optional entity. Users would, of course, be given a guarantee of anonymous protection of their product registration data. In addition, in the same way that open source companies have started implementing incentive driven programs to encourage participation of various forms, the same could be done to provide motivation for sharing bits and pieces of data.

Perhaps the first reaction to such a proposal is to ponder the seemingly inevitable qualms related to privacy issues. However, if the process of asking for the data were done in a non-invasive manner, protected by a privacy protection statute while carrying with it the potential for some type of compensation, there is very little reason to question whether users would consider providing at least some data. And some data is better than none at all, especially when those small parts are coagulated with other portions to construct a better portrayal of an open source community.

As is to be expected, there would be a number of issues that beg to be worked out, and no method would be a one size fits all. However, considering the current approaches to measuring community, a concept similar to a product registration system is worth bearing in mind.

About the blogger: Alex Fletcher is lead industry analyst at Entiva Group Incorporated, a research and analyst firm which concentrates exclusively on the open source software industry. Alex's main focus is working to help clients of all sizes formulate strategy and policy surrounding their use of open source software within the enterprise. He has prior experience as a consultant, software engineer and start-up founder. Alex can be reached at alex dot fletcher -at- entivagroup dot com.