I ran across Gordon Haff's Open Source: What Makes for Success?, which is a synopsis of several perspectives that Kevin Efrusy of Accel Partners has on topics related to the success of open source. And I thought it might be useful to add some commentary to the points brought forward therein.
"Low Surface Area." I agree that the open source model has, thus far, lent itself quite well to software projects which don't feature too many 'touch points.' However, I would caution to add that we are actually only in the early stages of an evolution process which features open source at its core. Therefore the fact that open source ERP, as Kevin thoughtfully highlighted as yet to achieve mainstream success, is more a symptom of the current phase of the process than it is an inherent characteristic of the model. Furthermore, I consider the argument for low surface area to apply better to the initial approach to open source software development than it is target market determination. In other words, open source works best when smaller feature sets are introduced consistently over a longer period of time.
Positioning against incumbents. I love to joke that the entire software industry has been in the process of setting itself up for the disruptive effects of open source for years. With corpulent margins guaranteed by closed software that serves to reduce choice and breed vendor lock-in, the proprietary model basically internalized its own structural weakness such that it is now being reduced to watching a diametrically opposed force eat its lunch. To compete and to slow its ebbing tide, proprietary companies are being forced to accept the validity of the very concepts that it effectively proscribed in the past. So while open source can indeed thrive in well-known categories by introducing innovation and disruption, its entrenched competitors have all but opened the floodgates.
Success metrics. I'm in agreement wholeheartedly that downloads, alone, just don't cut it...they're only a part of the big picture of progressive success. The numbers game actually matters more than some might assume. More interesting is Kevin's mention of the following:
"You want to know about people actually using the product. Kevin suggested that checking for updates is a reasonable heartbeat; the communication with the customer needs to provide something valuable in exchange for any information collected. (And, of course, companies have to be very sensitive to privacy concerns that can spiral out of control in the Web 2.0 echo chamber.)"
In a blog entry about reaching (quantifying) the non-paying customer, I proposed the development of stronger methods and strategies of gathering more useful data about who exactly is using open source software. The point isn't to trample on anonymity and/or privacy but to replace the blind science of relying solely on download numbers to portray the makeup of an open source ecosystem.
A changing game. The game (of building a successful open source business) is indeed transforming before our collective eyes, yet the fundamental principles/rules remain static. Among other things, open process must remain a focus, smart growth over rapid growth, and meeting the needs of customers/users will invariably mean the opportunity to gain more.
As the open source marketplace formalizes in structure its success factors are bound to do the same. At this point, in the relative beginning of its life cycle, those which exist serve as a statement of progress made, a reflection of the current state and indicators of the near future. One thing is sure, it will be amusing to look back at blog entries such as this in three to five years and compare past perceptions to eventual reality.