While most open source companies have zagged away from open source, Chef just zigged—a move that promises to be profitable. Here’s why.
If this year’s open source theme thus far has been all about closing off code to turbocharge monetization, Chef’s decision to dump its semi-proprietary (“Open Core”) licensing model for 100% open source is a nice rebuttal. While it’s too soon to tell whether the effort will succeed, there are sound reasons for suspecting more open equals more good.
The problem with half open
Despite Red Hat’s unparalleled success as a 100% open source company, the vast majority of open source companies have done some variation on the Open Core theme. That is, they might make most of their code available under an open source license, but they hold back some code (related to security or advanced operations or other things) so as to induce a would-be customer into becoming a paying customer.
For a variety of vendors, including MongoDB and Cloudera, this model has mostly worked. While Red Hat is far and away the most successful open source company, a number of others now generate in excess of $100 million in annual revenue using Open Core models.
SEE: Open source vs. proprietary software: A look at the pros and cons (Tech Pro Research)
Behind the scenes, the Open Core model introduces all sorts of problems. I spent nearly 15 years working for a variety of Open Core vendors, and one common theme at each of them was the inordinate bother it was to draw a bright line between what should be open source and what should be closed. Engineering, in turn, would have to manage different code bases (the “enterprise” version and the “community” version), constantly juggling which features would be in which. On the sales side, this model also forced salespeople to market against their own products, telling customers, on one hand, that their open source project was amazing, while on the other, in order to push the enterprise build, that it was “not amazing enough.”
The reason—get paid for writing code—is clear, but the gymnastics required to maintain it are unnecessarily burdensome and, arguably, not nearly as effective at compelling a purchase as believed.
This is why Chef’s decision to go 100% open source is so interesting.
Opening up open
Despite the friction inherent in the Open Core model, it persists and remains the most popular licensing model across the open source industry. Companies haven’t been willing to take the risk that maybe, just maybe, the value of their code isn’t really code at all.
In talking about the change, Chef’s co-founder and former CTO Adam Jacob wrote:
With a Free Software Product model, we say the software we produce is a valuable component of a complete Product. The software is fully open, creating the potential for collaboration, integration, and network effects across the entire stack. The value of the product is that software, plus the security, testing, build pipelines, sales, marketing, support, documentation, operational knowledge and content produced by Chef Software. In other words, it’s the value of being the best possible producers of an enterprise product, one that solves real problems for real customers. There is no more tension in the strategy — open source software produces better outcomes for everyone.
See the shift? Software is important, but it’s just part of the overall value for which a customer pays. In reality, as we’ve seen with the rise of Amazon Web Services, it’s perhaps the least valuable part of the equation, with the much more complex act of running the software (“operations”) the real product that customers hope to buy.
Chef isn’t going that far. Not yet, anyway. Rather than make its software only available as cloud services, Chef is opening up its code while offering a certified distribution of it, Red Hat Enterprise Linux-style. In this way, it’s doubly open: You can get the source code, or you can get the cloud service, or both.
SEE: Software licensing policy (Tech Pro Research)
As just one indication that this is a smart path forward, MongoDB has been on a tear lately with its cloud service. While the company engages in licensing gymnastics, its Atlas database-as-a-service offering generated over $100 million last year, is growing at 400%, and now accounts for 34% of the company’s overall revenues (more than triple the percentage of this time last year).
This is the future of open source: Running in a cloud. Dumping Open Core and going all in on open source, while making that software valuable through a cloud service, will generate more revenue, more community, and more peace of mind for vendors and customers alike. As Jacob noted in his blog, it unites customers and vendors at the top of the funnel, engaging them in a true community, one that should be much more lucrative than the discordant Open Core model of yesteryear.