A recent article on TechCrunch provides some worrying insight into the question of the worldview of VC investors in open source businesses.
The authors have impeccable credentials. Max Schireson was CEO of MongoDB and Dharmesh Thakker was managing director of Intel Capital, a hugely influential investor in open source startups. Both are now executives at VC Battery Partners. Yet the article is deeply flawed, with serious factual errors and an outdated, blinkered view of monetization. If this is how people who could be expected to be experienced and enlightened see open source, the prospects are grim for the rest of the field.
Getting a detail wrong happens all the time in journalism -- and usually gets corrected as soon as it’s pointed out. But when supposed experts get fundamentals profoundly wrong, alarm bells ring. The most serious factual error, reflecting an impossibly bad understanding of the Apache Software Foundation, is the assertion that Apache tried selling support services in the past and failed. The authors state:
The Apache Software Foundation, whose open source server software burst onto the tech scene in the 1990s, tried a similar strategy — essentially, selling customers support as a sort of insurance policy on their deployments, which some users considered risky because they were based on open-source technology.
As a U.S. nonprofit charity that’s almost entirely staffed by volunteers, Apache could not have offered any commercial service, as its VP of Brand, Shane Curcuru, points out in the article’s comments. But even if it could, it wouldn’t, because Apache is fundamentally a neutral meeting place for developers (whose motivations are extrinsic to Apache) to collaborate freely. Apache doesn’t even let projects fund-raise for their direct operational needs, let alone offer commercial service. Any VC who believes Apache could, did, or would offer commercial services has completely missed the point of its existence.
Then there’s an amazingly poor grasp of the role of open source licensing on display in another bold factual error that refers to the Open Source Initiative’s list of approved licenses:
Some licenses allow anyone to create derivative works that build on the original product, while others reserve that right only for the owners of the original product.
That’s simply wrong. According to the Open Source Definition rule 3, “The license must allow modifications and derived works, and must allow them to be distributed under the same terms as the license of the original software.” There are no OSI-approved open source licenses that reserve the right to create derivatives to the copyright owner.
The authors don’t even understand the business model of market leader Red Hat. They say:
So while Red Hat maintained a vibrant open-source community around Fedora, updating it with new features and releases every six months, the company reserved enterprise-grade features and optimization on standard Intel hardware for people using the paid enterprise product.
But as Red Hat employee Harish Pillay pointed out, “RHEL has no Enterprise-only features. What is in RHEL is in Fedora.” The authors may believe -- based on biases arising from their own careers -- that the only viable business model for a company monetizing open source software is to create artificial scarcity to ape the models of legacy corporations, but Red Hat clearly proves you don’t have to do that.
Those biases seem to arise from an outdated view of the market for open source software. Students of history know that pioneers of new markets are able to command profit margins approaching 100 percent as long as they can behave as monopolists. As their markets becomes subject to fair competition, margins fall. Expecting 90 percent margins is probably not realistic, yet the authors clearly do:
But don’t try to build your business on services revenue. It carries a lower profit margin than other revenue streams -- perhaps 20 to 30 percent, as opposed to 90 percent for product-related revenue -- and it’s not generally repeatable.
In plenty of markets a business with 25 percent margins is considered healthy, and the increasing maturity -- and commoditization -- of the software market suggests that the margins these VCs are seeking are based on historic exceptions rather than future norms and are aimed at generating one or two high-yield investments in the face of the overwhelming statistics of failure of businesses that let the VCs in.
More significant, the way open source projects arise and the way their existence is monetized by their creators has changed significantly in recent years. A collaborative co-development community adds innovation, expands and multiplies markets, and reduces anticompetitive behaviors. Recognizing this, companies like Twitter, Facebook, and Rackspace have chosen to monetize their software by giving permission in advance to all comers to use and improve their software. Yet the authors are still stuck in a world characterized by products like MySQL and MongoDB, where co-developers are discouraged and “community” is synonymous with “customer.” Their discussion of one of these communities betrays their outlook:
OpenStack recently has been criticized for moving in multiple directions, driven by the various hardware players who all have competing business and strategic agendas and, now, products that may not work well with each other.
The errors highlighted here are not merely mistakes; rather, they reveal a worldview. People who believe that Apache is a competitor, OSI approves licenses that permit monopolization, Red Hat is a business that’s succeeded through artificial scarcity, and open source communities with diverse agendas are "broken" are not the people you want in your new open source business.
They will try to persuade you to secure software patents so that they have an asset to trade when you fail; they will eject you from your own company when you try to hold true to software freedom principles; and they will treat your business as a failure if all it does is earn a decent living for you and your employees. You may want to grow your open source-based business another way.