5 Things I Learned Going from Open Source to Commercial Open Source

I started a London-based software agency back in 2001. We spent over 2 decades working with global enterprises, startups and everything in between. We’ve worked with engineering teams of all shapes and sizes, and we’ve seen how hard it can be to just ship software.

We spent a lot of time understanding why shipping is so hard, seeing common patterns, and figuring out how we could make things better for our clients.

Flagsmith was born out of these learnings. We built it as an internal feature flagging tool, and in 2018 we open sourced the project and launched a paid SaaS product. 

After growing revenue to enough to support ourselves we decided to go full time and bring the commercial open source feature flag and remote config product to market. That was a year ago, and since then we have grown revenue 10x. We wanted to share our 5 key learnings from the last 12 months.

1. Flexibility of deployment expands your market. Keep your stack as simple as possible. It makes it easier for people to contribute, and reduces deployment friction for Enterprise customers. We’re running a REST API based on Django, ReactJS and a SQL database. That’s it. We have seen this be extremely valuable to enterprise customers who are running Kubernetes, Helm, Openshift, Docker across all the cloud providers.

2. Remove impediments to growth before pouring on the gas. Once we decided to go full time building out Flagsmith as a commercial open source feature flagging platform, we spent the first few months setting ourselves up for growth. These things elevated us from a project to a company. Three key things we did were;

• Rebranded from Bullet Train to Flagsmith. This established our brand, and had a big impact on our ability to grow through organic search (it turns out that Bullet Train returns, well, more trains than feature flags).
• Unified to a single GitHub repository, and started developing in the open. This helped us grow engagement with our open source community.
• We established our open source philosophy and commercial model. This led to a dual license structure and more clarity about which features should be open.

3. Identify your target customer and focus relentlessly on them. Entering the feature flag space we were aware that LaunchDarkly is a massive success, but recognized that there is still a lot of the market that is underserved. By clearly defining our niche and getting lots of feedback from those engineering teams, we were able to add features they need, become a reputable LaunchDarkly alternative and deliver value to them. We don’t think that winner-take-all is either real in B2B software or best for customers. So run your race and focus on your target customer.

4. Providing first class support is a clear differentiator, and a great way of doing customer development. Jumping on a call to help someone with their code is a great experience for them, and the perfect opportunity for you to learn how they are using your product. We also set up a customer advisory board and listened to our customers’ wants and needs.

5. Don't be afraid to talk to and learn from others. Despite our revenue growth over the last year, we recognize that isn’t an overnight success. We have prioritized learning from others and turned that into something that everyone can learn from. The Open Source community is incredibly giving. We started The Craft of Open Source podcast and have learned a tremendous amount from the founders we have spoken with. It has now been listened to over 20,000 times and the learnings have helped us tremendously along the way.

These learnings have been really valuable for us, but we’d love to hear from other projects and founders who are pursuing the commercial open source route.