The Real Reason Open Source Startups Fail

Editor’s note: Alex Freedland is a co-founder and chairman of Mirantis and an OpenStack Foundation board member.

The recent news around Nebula shutting its doors has stirred speculation that OpenStack startups are struggling because of the state of the OpenStack market. There is even a piece claiming that the OpenStack dream is on “life support.”

This couldn’t be further from the truth. The reality is that winning in open source requires a playbook that is drastically different from one that most VCs investing in technology today are used to.

In summer 2011, Mirantis was a 150-person IT services company doing consulting around open-source infrastructure technologies.

We ran across OpenStack in a few engagements, saw the community momentum behind it and decided to go all in. In the search for early business, we called Randy Bias of Cloudscaling, an OpenStack services startup that had just raised a Series A to build an OpenStack distribution.

“We raised money and it’s all about building product for us now,” he said. “We have this deal to do an OpenStack deployment for Internap – do you have people to subcontract for us?”

This was the first pure OpenStack project we ever did, and this focus on services is what laid the groundwork for us to grow into the OpenStack product company that Randy originally wanted Cloudscaling to be.

For Randy, unfortunately, raising a VC round was the beginning of a very different journey. Following the guidance of his investors, Randy dismantled his thriving multi-million dollar cloud consulting business to build Cloudscaling OpenStack, a product that was too early and too incomplete to satisfy the services-hungry OpenStack market.

Three years later, Randy watched Red Hat pay almost $100 million for eNovance, an OpenStack consultancy barely bigger than Cloudscaling of 2011. Kudos to Randy for being able to land his ship in the calm waters of EMC Corporation in his sale of the company last year.

Where did Cloudscaling and Nebula miss the mark? They used the wrong playbook. A typical IP-based company playbook goes something like this:

  • A product wizard creates a differentiated technology;
  • The founders and the sales team sell a handful of deals to show proof points;
  • The company is sold for several hundred million to someone with existing distribution channels; or
  • A number of major growth rounds are raised to build out distribution channels in order to  IPO

This playbook works for innovation markets. The VCs are buying into partial ownership of the differentiated technology and monetize on the leverage when the technology is in demand.

Open source ecosystem markets behave differently and therefore require a very different playbook. There, the differentiation is not in the technology you build; it is in the process and expertise that you slowly amass over an extended period of time. All of the successful entrants (Red Hat, Cloudera or Hortonworks, etc) have followed the same playbook. It works like this:

  • Start selling services engagements around the ecosystem and build customer intelligence;
  • Invest heavily to build influence in the respective upstream communities by contributing directly to the code base you’re betting your business on (In OpenStack, it’s easy to see who is and who is not contributing on Stackalytics.);
  • Start offering training;
  • Add commercial packaging of the technology;
  • Optional: Build value-added components above and around the technology.

Building an open-source company requires operational excellence and staying power, not just genius hackers with a vision of taking down VMware. Startups and VCs that run the company by the innovation playbook fail, while those that focus on expertise first — and thus the open source playbook — do well. This is what we are witnessing in the OpenStack space.

OpenStack’s imminent death makes a great headline, but it lacks understanding of the current state of the ecosystem. The truth is that companies are seeing their revenues and customer engagements steadily grow. Red Hat, Mirantis, HP, IBM, Huawei and Cisco are successfully selling their own versions of OpenStack.