How to Start Up an Open Source Company

24 Oct 2014

Evolveum is a successful open source company now. We develop open source Identity and Access Management (IAM) software. We have legally established Evolveum in 2011 but the origins of Evolveum date back to mid-2000s. In 2014 we are getting out of the startup stage into a sustainable stage. But it was a long way to get there. I would like to share our experiences and insights in a hope that this will help other is their attempts to establish an open source business.

The basic rules of the early game are these: It all starts with an idea. Of course. You need to figure out something that is not yet there and people need it. That is the easy part. Then you have to prototype it. Even the brightest idea is almost worthless until it is implemented. You have to spend your own time and money to implement the prototype. You cannot expect anyone to commit the time or money to your project until you have a working prototype. So prepare for that - psychologically and financially. You will have to commit your savings, secure a loan or sacrifice your evenings and nights for approx. 6 months. If you spend less than that then the prototype is unlikely to be good enough to impress others. And you cannot have a really successful project without the support of others (I will get to that). Make sure that the prototype has proper open source license from day one, that it is published very early and that it follows open source best practice. Any attempt to twist and bend these rules is likely to backfire when you are at the most vulnerable.

Then comes the interesting part. Now you have two options. The choice that you make now will determine the future of your company for good. Therefore be very careful here. The options are:

Fast growth: Find an investor. This is actually very easy to do if you have a good idea, good prototype and you are a good leader. Investors are hungry for such start-up companies. You have to impress the investor. This is the reason why the prototype has to be good. If you started alone find an angel investor. If you already have a small team find a venture capitalist. They will give you money to grow the company. Quite a lot of money actually. But there is a catch. Or better to say a whole bunch of them. Firstly, the investor will take majority of shares in your company in exchange for the money. You will have to give away the control over the company. Secondly, you will need to bind yourself to the company for several years (this may be as long as 10 years in total sometimes). Which means you cannot leave without losing almost everything. And thirdly and most importantly: you must be able to declare that your company can grow at least ten times as big in less than three years. Which means that your idea must be super-bright ingenious thingy that really everyone desperately needs and it also needs to be cheap to produce, easy to sell and well-timed - which is obviously quite unlikely. Or you must be inflating the bubble. Be prepared that a good part of the investment will be burned to fuel marketing, not technology. You will most likely start paying attention to business issues and there will be no time left to play with the technology any more. Also be prepared that your company is likely to be sold to some mega-corporation if it happens to be successful - with you still inside the company and business handcuffs still on your hands. You will get your money in the end, but you will have almost no control over the company or the product.

Self-funded growth: Find more people like you. Show them the prototype and persuade them to work together. Let these people become your partners. They will get company shares in exchange of their work and/or money that they invest in the company. The financiers have a very fitting description for this kind of investment: FFF which means Friends, Family and Fools. This is the reason for the prototype to be good. You have to persuade people like you to sacrifice an arm and a leg to your project. They have to really believe in it. Use these FFF money to make a product out of your early prototype. This will take at least 1-2 years and there will be almost no income. Therefore prepare the money for this. Once you are past that state the crucial part comes: use your product to generate income. No, not sell the support or subscription or whatever. This is not going to work at this stage. Nobody will pay enough money for the product until it is well known and proven in practice. You have to use the product yourself. You have to eat your own dogfood. You have to capitalize on the benefits that the product brings, not on the product itself. Does your product provide some service? Set up a SaaS and provide the service for money. Sell your professional services and mix in your product as additional benefit. Sell a solution that contains your product. Does your product improve something (performance, efficiency)? Team up with the company that does this "something" and agree on sharing the revenue or savings generated by your product. And so on. You have to bring your own skin to the game. Use the early income to sustain product development. Do not expect any profit yet. Also spend some money and time on marketing. But most of the money still need to go to the technology. If the product works well then it will eventually attract attention. And then, only then, you will get enough money from subscriptions to fund the development and make profit. Be prepared that it can take 3-6 years to get to this stage. And a couple more years to repay your initial investment. This is a slow and patient business. In the end you will retain your control (or significant influence) over the product and company. But it is unlikely to ever make you a billionaire. Yet, it can make a decent living for you and your partners.

Theoretically there is also a middle way. But that depends on a reasonable investor. An investor that cares much more about the technology than he cares about money, valuations and market trends. And it this is extremely rare breed. You can also try crowdfunding. But this seems to work well only for popular consumer products that are not very common in the open source world. Therefore it looks like your practical options are either bubble or struggle.

And here is a couple of extra tips: Do not start with all-engineer team. You need at least one business person in the team. Someone that can sell your product or services. Someone that can actually operate the company. You also need one visionary in the team. Whatever approach you choose it is likely that your company reaches full potential in 8-12 years. Not any earlier. If you design your project just for the needs of today you are very likely to end up with an obsolete product before you can even capitalize on it. You also need a person that has his feet stable on the ground. The product needs to start working almost from the day one otherwise you will not be able to gain the momentum. Balancing the vision and the reality is the tough task. Also be prepared to rework parts of your system all the time. No design is ever perfect. Ignoring the refactoring needs and just sprint for the features will inevitably lead to development dead-end. You cannot afford that. That ruins all your investment.The software is never done. Software development never really ends. If it does end then the product itself is essentially dead. Plan for continuous and sustainable development pace during the entire lifetime of your company. Do not copy any existing product. Especially not other open source product. It is pointless. The existing product will always have a huge head start and you cannot realistically ever make that up unless the original team makes some huge mistake. If you need to do something similar than other project already does then team up with them. Or make a fork and start from that. If you really start on a green field you have to use a very unique approach to justify your very existence.

I really wish that someone explained this to me five years ago. We have chosen to follow the self-funded way of course. But we had to explore many business dead-ends to get there. It was not easy. But here we are, alive and well. Good times are ahead. And I hope that this description helps other teams that are just starting their companies. I wish them to have a much smoother start than we had.

(Reposted from https://www.evolveum.com/start-open-source-company/)