Ben Golub is CEO/CoFounder of Docker/dotCloud, an open source engine that makes it easy to pack, ship, and run any application as a lightweight container.
I’ve seen more startups fail because they failed to pick a direction than because they picked the wrong direction. Leaders need to both articulate a strong point of view and then align all activities (roadmap, hiring, target users/customers, positioning, messaging, funding) etc. around that point of view. In the best case, the direction is right and you suceed wildly; in the second best case, you makethe direction right because of the weight that goes behind it; in the worst case, you learn quickly that the direction is wrong and can change course.
We have executed a major pivot as a company as we shifted from being primarily a public PaaS company to being an open source company behind the Docker.io project. We’re now going after a much bigger and much more strategic set of problems and markets, and are experiencing a massive increase in momentum and strategic interest.
We’ve invested in making our project really easy to understand and to get started using. One particularly useful tactic was introducing an online tutorial that takes about 15 minutes to complete, doesn’t require any download, and provides developers with an intuitive understanding of Docker. Three weeks after launching, over 7,500 people had completed the tutorial, and we saw massive growth in almost all other metrics.
We are an open-source company, and that has proven to be a great way to find and attract talent. Three of our best, recent hires have been from members of the community who were doing great work in the community. Being open source let’s us see who is interested, who is talented, and who is motivated.
We have taken being open to an extreme: our code, our processes, our roadmap, our tests–eveything is out in the open for the public to see. Developers at dotCloud have to use the same processes as anyone in the broader community to contribute to our code base. This has helped in innumerable ways–everyone knows what we are doing, everyone is aligned, and we can avoid the temptation to do stupid things.
We discussed Docker at a presentation at PyCon, just intending it to be a technical preview of a bit of code that we were making available to the open source community. Reaction was huge. A few weeks later, someone leaked the existence of our website to hacker news. The take off since then has been amazing: over 90,000 downloads in less than 6 months, with over 6,000 github stars, thousands of Dockerized applications, over 150 contributors, etc.
One great thing about start-ups is that you are free to experiment–to throw things against the walls and see what sticks. But, you’ve got to budget time to clean up the wall. Every project, no matter how small, enacts some ongoing cost. If you don’t take the time to prune failed experiments, the cost of maintaining them will drag you down.
When we redesigned our website, we forced ourselves to have just two main calls to action: on the main page 1)Learn more, and 2)Get started.
Making this change has dramatically reduced user confusion, abandonment rates, etc. Of course, we have links in the footer to all of the common other things that a motivated or directed user would want to find.
Many years ago, I was part of a group of students trying to get a business school started in Tashkent, Uzbekistan. This was just a few years after the Soviet Union had collapsed, and I think we thought that if we could teach people how to read a balance sheet, freedom, democracy, and free markets would blossom. Of course, after a few months on the ground, reality set in–we taught some courses, but it takes a lot to overcome 70+years of history and practice. Still, everybody should have at least one failed startup!