NB: This is a viewpoint by Adam Goldstein, co-founder of Hipmunk, a flight and hotel metasearch site headquartered in San Francisco.

When we first started Hipmunk in 2010, we had no idea what we were getting ourselves into.

Neither of us had any experience dealing with airlines, hotels, or online travel agencies (OTAs). We were in for a whole bunch of surprises.

As background, I had just graduated from MIT with degrees in Electrical Engineering/Computer Science and Mechanical Engineering. Steve had founded, launched and sold Reddit, which became the largest social news site and one of the largest communities on the internet.

Neither of us had travel experience, and neither of us had experience managing teams larger than five people.

A strange industry

The first thing that surprised me was how incestuous the relationships among intermediaries were.

The notion that an OTA might accept ads from another OTA, or that a metasearch might link to another metasearch, was completely bizarre to me.

Eventually I learned the motivation: consumers will search multiple sites whether you like it or not. But it still seems weird.

The next fun came when integrating with ancient technology systems. Often, our engineers became bewildered with the difficulty—and lack of help—from the companies we were sending business.

The way we fixed the issue was to have our business development (BD) folks do the monotonous follow-up and follow-through, sparing the engineers the distraction of getting "no" for an answer 10 times before finally getting a “yes.”

The next few lessons had nothing to do with the travel business, and everything to do with the startup business.

The most pressing one was hiring: when’s the right time to bring someone else in with experience we didn’t have?

Steve (my co-founder) and I had learned a whole bunch of things on the job—BD, PR, etc.—but at a certain point, we realized we were going to need some experienced help with all those and more.

By the time it became obvious, it was urgent; it would have been much better to hire when we first had the inkling we’d need someone.

Meetings

Another early point of friction was managing the interaction of all our different teams when it came to planning and prioritization.

At first we tried having one-on-one meetings with all the people who had opinions, and then Steve and I would sit down and synthesize it all. But that didn’t work very well—we ended up mis-quoting or mis-prioritizing other people’s ideas.

We then swung in the other direction—getting everyone with an opinion on every planning meeting. That led to gridlock, and meetings took an eternity.

Eventually we settled on a hybrid—small meetings with the relevant teams, distilling opinions, and then having a mid-sized meeting with the relevant people to get it over the finish line.

Staffing

Once we had a real team up and running, and figured out some hiring practices, we eventually ran into people who just weren’t working out the way we wanted.

We tried all the standard feedback tricks, but some of the people still didn’t meet our standards.

The first termination was miserable for both sides. Eventually we learned to keep it short, dignified, and friendly, and things got better quickly.

More recently, we’ve been experimenting with new ways of running releases.

When it was just Steve and I, code was either on our computers or running to 100% of users—nothing in-between.

After one too many buggy releases, we now do staged deployments, with our internal team able to test on a server, then running an A/B test to a subset of our users, and then ramping up to 100% if it’s successful.

Board meetings

When we first started, we had no formal process or schedule for strategic meetings. But then we started raising money, and things changed.

Over time, we added members to our board: one when we raised our Series A round, then two more when we raised our Series B.

Along with these came board meetings—an alien concept at first, but then a refreshing chance to discuss the company's most pressing opportunities and challenges.

The trick was learning to manage them well; originally, the meetings would become muddled in the specifics of product implementation, employee compensation, etc.

I've found the meetings much more helpful when we come in with a message of "here's what's happening, here's where we think we should be going, and here are the major strategic questions we should address first."

NB: This is a viewpoint by Adam Goldstein, co-founder of Hipmunk, a flight and hotel metasearch site headquartered in San Francisco.