More on Open Core
Many of the responses to my previous blog post "Open Core or Open Snore?" were in agreement, and some were not. As is often the case, the more interesting ones expressed disagreement. Some took issue with my post by pointing out open core companies that might be termed success stories: SugarCRM, Alfresco, Mindtouch. But then, I never wrote that open core cannot be successful, but rather that any success will be limited by nature of the model. Open core effectively places a cap on community development turning open source efforts into a viral marketing play, when it can be so much more.
One critique that did resonate was how much open source dev models actually impacted the bottom line. A company's or project's success depends on a myriad of factors, including open source strategy and tactics, but mostly on product execution and delivery. Seeing as how some companies will succeed with practically no open source development at all, it's only natural to concede that an open core approach will succed in some markets. However, if I were creating an open source community strategy in a crowded, competitive market, I sure wouldn't want to place an artificial handicap on my community development practices. I'll use 2 case studies to illustrate my point: Red Hat / Fedora and CollabNet / Subversion.
But first, a word about the Open Core success stories thus far - the verdict is still out. However, it would seem that SugarCRM should have been poised by now to become the default CRM platform, vigorously competing with Salesforce and others. Whether their inability to challenge Salesforce rests on product development, community development, and/or some other things, I can't say for sure. However, if I were in their shoes, I would be re-launching their open source product Real Soon Now and using a different strategy to do it.
A few other examples, such as JasperSoft and Alfresco, are worth mentioning. These are examples of markets that are still considered "high margin" businesses, which are often ripe for a disruptive change agent, such as an open source project. In my view, both Jasper and Alfresco have taken an open core approach because they don't need to be any more open than that, at least for now. If more competitors enter their respective markets, however, they may need to revisit that strategy. To summarize, open core may be "open enough" for some markets but not for others.
Some would claim that only a low margin, highly commoditized infrastructure-driven market can account for a successful company like Red Hat, but I don't buy that - mostly because I haven't seen enough companies try it for non-infrastructure projects.
Let's find a couple of projects that I think have hit the right notes, community development wise: Fedora and Subversion. I use these for a couple of reasons:
1. Fedora and Subversion were allowed to develop their own identities separate from the mother companies, Red Hat and CollabNet, respectively. This is what I call trusting your community.
2. As a result of #1, both have developed mature ecosystems around their respective platforms - although it could certainly be better. I've dinged both companies in the past for not recognizing the benefits of increasing their projects' center of gravity.
3. Both projects have found unequivocal success. Fedora may have been out-hyped by Ubuntu in some community respects, but they are the foundation of a profitable company. Also, subversion may not be cool enough for the Ruby kids, but it has definitely encroached on the turf traditionally reserved for commercial products. To the extent that Subversion faces new competition, it's because of its success.
By the way, I'm not dissing Ubuntu - their community strategy has one-upped everyone. The reason Fedora makes such a compelling example is because there's a clear-cut dividing line between Red Hat's pre- and post-Fedora strategy, thus making them a bit easier to study. Also, it's clearer how Fedora feeds into the overall revenue of Red Hat, whereas this aspect of Ubuntu is still taking shape.
Looking at Subversion, can one imagine what would have happened if CollabNet had taken an open core approach? Would the same sort of ecosystem exist today? Probably not. Would CollabNet even exist today? Who knows, but it's pretty clear to me that a substantial chunk for their revenue comes as a result of Subversion's ubiquity.
Note that one of these examples is an operating system, long ago given commodity status, and the other is an infrastructure component of developer stacks, serving a market more coveted for their early adopter tendencies than for direct revenue. Missing here are the "up-stack" applications, and for good reason - I can't think of any which have tried the approach that I advocate or that are mature enough to take into consideration. As I noted in my previous post, most venture-funded outfits have taken the safer approach of open core.
Given the obvious successes of the examples I've noted above, why aren't more companies taking note?