A Pivot that Worked: The Docker Story

Originally published January 22, 2014

For the vast majority of startups, the path to success is not a linear one. Nearly all founders face that moment where their company’s continued existence comes into question. I joined my first startup, Wily Technology, in 1999 right around that moment. Lew Cirne (also the founder of New Relic) hired me to help build a new product to take the company in a different direction. Few people know or remember that the company I founded, Flurry, began by offering a consumer-focused email client for feature phones (remember those?). After a dramatic change in direction, today it’s the leading analytics service and advertising network for mobile applications. DotCloud, the company Trinity funded in 2010, morphed into Docker in 2013.

As we celebrate Docker’s recent success and its closing of a significant new round of financing, I thought it would be worth sharing the story of how dotCloud became Docker and some of the key lessons we learned. Though today we all expect Docker to be a runaway success, there was a time when we weren’t so confident about the company’s future.

Before I get into the story, there are a few points that need highlighting up front. First, this is the story of how the Docker pivot happened, not what ultimately made Docker successful. I’d argue that the way the company pivoted maximized its chance of success, but even a pivot done correctly won’t always change a company’s outcome. Second, as a board member, most of my interaction during this period was with Solomon Hykes, who was CEO of the company at the time. A lot of people inside the company were involved in the pivot, more than I can single out in this post.

First some history…

Solomon Hykes joined Y Combinator in the fall of 2010 and Trinity led the seed round in his new company, which he named dotCloud, in September of that year. On the heels of that small investment, the company raised a Series A in April of 2011, where we were joined by Peter Fenton of Benchmark Capital. At the time, the Platform as a Service (PaaS) market seemed poised for explosive growth and dotCloud had the market’s best product and technology. We were the only “polyglot” PaaS provider in the market, meaning we could run applications built in almost any programming language.

Reality strikes

When a startup fails to live up to the high expectations of its founders and investors, the reasons are rarely simple. Typically, multiple factors conspire to thwart even the best-laid plans. Sometimes markets don’t develop the way you expect. Other times product/market fit is the problem. At the risk of oversimplifying, the most important thing that went wrong at dotCloud was that the PaaS market never really materialized the way we expected. There’s plenty of evidence to suggest that this is the case — in general, horizontal PaaS companies have struggled or been acquired for uninspiring sums. In early 2013, almost two years after the Series A, it was hard to know for sure what was going on in the market. What we did know was that dotCloud’s momentum was waning and the team was getting restless. Despite quarter over quarter revenue growth and a passionate user base, it still felt like we were pushing a boulder up a hill. In the midst of all of this, we got a modest acquisition offer for the company. We didn’t want to take it, but we didn’t feel as though there were many other options. Few companies are able to raise a Series B with the kind of modest growth that dotCloud had seen.

Profiles in ignorance and courage

In the midst of dotCloud’s darkest hour, Solomon approached the board with a radical proposal: the team wanted to open source the container technology they had built for dotCloud. At the time, they weren’t sure what the business model would be, but they had an intuitive sense that Docker’s container technology was something that developers needed. To a board focused on protecting the value of the company’s assets, the idea of open-sourcing the company’s technological crown jewel made no sense. Instead, we encouraged Solomon to double down on making the PaaS business work or find a better M&A offer for the company. Given where the company is today, I feel foolish for not having given further consideration to the merits of this idea. But in fairness to those of us on the board, the team saw something that no one else recognized. And it’s a good thing Solomon decided to go ahead with their plan anyway.

Thanks to Solomon and his team, we have a business called Docker that’s on fire and changing the world of software infrastructure. I wrote a bit about how that’s happening in a previous blog post. What follows are some lessons I’ve learned from this experience.

Raise more than you need

I always say that I’ve never heard an entrepreneur complain about raising too much money, but I frequently hear complaints about the reverse. When fundraising for early stage companies, many entrepreneurs optimize for minimizing dilution and raise almost exactly the amount that’s called for by their financial plan. As we all know, things rarely go as planned in startups. When raising his Series A, Solomon made a decision that in today’s market looks unorthodox: he made room in the round for two deep-pocketed venture firms, Trinity and Benchmark. This meant he got a lot more money, as well as the support of two separate partnerships, but it also meant he took what at the time what must have seemed like an unnecessary amount of dilution. In the end, that money bought time enough to execute on a pivot and dramatically change the fortunes of everyone involved.

Skinny down in times of trouble

After the decision to pivot to Docker, Solomon very quickly restructured the company, whittling it down to a core group of employees dedicated to the new vision. Doing so had two positive effects. First, it further extended the company’s runway, giving us the maximum amount of time possible to see the pivot through. Second, it greatly improved morale among the new core team, reenergizing them for the path ahead. This might not be obvious, but a small team committed to an idea will be happier than a larger team with some doubters. We were fortunate to have a number of people on the team (Sam, Eric, Jerome, Julien, Ben, and others) who helped keep everyone motivated through this difficult time.

Pick your board members well

When picking board members, you never know what the future will hold. One minute your business might be full of promise, and the next minute doomed to fail. As a founder, you need partners, not just investors, on your board — people who will be supportive in times good and bad. We’re fortunate that Solomon picked a great group of board members: Marc Verstaen, Peter Fenton, Jerry Yang, and me. Philosophically, we all believe that the role of the board is to advise and ask hard questions, not to run the company. Even though at the time we didn’t understand fully why Solomon wanted to open source Docker, the decision was his to make and we did everything we could to support him.

Don’t always listen to the experts

Even if you’ve chosen the smartest board members in the world, that doesn’t mean that the CEO should always listen to them. Had Solomon followed the guidance of his board, then Docker would never have been born, and containerization technology would not be getting the amazing adoption we’re seeing today. Instead, Solomon solicited feedback from a wide variety of constituents, including his team and advisors, before making a final decision. In the end, it took a lot of courage on his part, and mutual trust between Solomon and the board, for him to go ahead with the plan. Sometimes the right thing to do is to trust your gut.

The value of vision

Solomon Hykes is an exceptional entrepreneur. None of the success Docker is having today would have been possible without his insight and courage, to say nothing of the technical accomplishments. He knew when to cut bait and when to go with his gut. No combination of good idea and great investors can accomplish anything without exceptional, visionary founders at the helm. And of course, Solomon couldn’t have done it alone.

It takes a village

I’d argue that this was the right way to pull off a pivot and it gave the company the greatest chance of future success. However, just because a company pivots well doesn’t mean that the new business will work. I’m not claiming that Docker was destined to be successful. Solomon has a great team and is fortunate to have Ben Golub as his co-founder and CEO. We have a fantastic new financial partner in Jerry Chen of Greylock Partners. Employees, investors, advisors, and the Docker community itself have made major contributions to the company’s current momentum. We’re proud to have backed this effort and are thrilled by what has been accomplished at Docker.