Why Starting Justin.tv Was A Really Bad Idea, But I'm Glad We Did It Anyway

Editor’s note: The following guest post was written by Justin Kan, founder of Justin.tv.

Right now I’m neck deep in product launch mode, putting the finishing touches on our new mobile video application—Socialcam. Of course, I’ve been here before . . .

Years ago when we launched the Justin.tv show we had no idea what we were doing. This much was obvious to anyone who watched. Outsiders attribute far more strategic thought to the venture than we gave it. Some think that we planned all along to start a live platform, and that the Justin.tv show itself was a way of promoting that platform. While this ended up happening, none of it had crossed our minds at the time.

Emmett Shear and I had been working on Kiko, the first Javascript web calendaring application in the Microsoft Outlook style. We prototyped the application in our final year at Yale, went on to raise money from Y Combinator, then continued working on it for over a year.

Then Google Calendar was released—boom—absorbing most of our nascent user base and capturing most of the early adopter mindshare. But to be perfectly honest, Kiko would have failed regardless. We were too easily distracted and hadn’t really thought through the strategic implications of owning a standalone calendaring property (hint: no one wants a calendar without email). A short time later we were burned out and spending most of our time playing Xbox with the Reddit guys in Davis Square—hardly a startup success story.

Emmett and I started thinking about possible ways to get out of the calendar business. At the same time, I was startup fatigued. We had spent over a year paying ourselves nothing. The seed and angel investment market conditions were the polar opposite of what they are today. It had been a struggle to even raise a paltry $70,000, and we had failed to build a product with real traction. I was starting to think about moving back to Seattle to try something new, maybe in a different industry.

Still, we learned a ton and it was fun to be part of the early Y Combinator startup community (then largely in Boston). We became friends with Matt Brezina and Adam Smith (of Xobni), Trip Adler, Tikhon Bernstam and Jared Friedman (of Scribd), and many others. It’s amazing to see how many of those friendships persist today, and even more amazing how well many of those companies are doing.

Coming back from one particular YC dinner, Emmett and I were discussing strategic ideas for Kiko, and I remember telling Emmett an idea that popped into my head: what if you could hear an audio feed on the web of our discussion? Wouldn’t that be interesting to other like-minded entrepreneurial types? We kept going, and eventually the idea morphed into a video feed. Then it became a live video feed. Then it became a continuous live video feed that followed someone around 24/7. Then it had chat, and a community built around watching this live show, which was now a new form of entertainment. I was hooked.

I couldn’t stop talking about the idea. I mentioned it at YC dinners and to other friends. I even came up with a perfect name for it: Justin.tv. On one trip to DC, I told my Dad and my college friend Michael Seibel what I was thinking. Eventually, in-between drinking sessions, we thought of a brilliant idea for divesting ourselves of Kiko, which is a story for another day. After that, Emmett and I were coming up with other startup ideas (I guess we got excited about staying in the industry after all). One particular favorite was the idea of a web app that would ingest your blog’s RSS feed and then allow you to layout and print physical magazines from it. Excitedly, we drove one afternoon to Paul Graham’s house to pitch it.

We explained the idea to Paul and Robert Morris, who just happened to be at the house visiting. I vaguely recall there also being a “this will kill academic publishing” angle, although I can’t figure out how that sensibly fits in now. Paul didn’t particularly like the idea: he didn’t think people would use it. “Well,” he said, “what else do you have?”

I said the only thing I could think of: “Justin.tv.”

Because it was something I was clearly passionate about, and because creating a new form of entertainment was clearly a big market (if you could invent one!), Paul was actually into it. Robert’s addition to the conversation was “I’ll fund that just to see you make a fool of yourself.” Emmett and I walked out of there with a check for $50,000.

Six months later, we’d recruited two other cofounders (Kyle Vogt, our hardware hacker, who we convinced to drop out of MIT on a temporary leave of absence, and Michael Seibel, my college friend from DC, who became our “producer”). We built a site with a video player and chat and two prototype cameras that captured, encoded and streamed live video over cell data networks, negotiated with a CDN to carry our live video traffic, and raised an additional couple hundred thousand dollars. Our plan? Launch the show and see what happens.

Now, let me just tell you why this was a bad idea:

  • We didn’t have a plan. We loosely figured if the show became popular we could sell sponsorships or advertising, but we didn’t have a plan to scale the number of shows, nor did we understand what our marginal costs on streaming, customer acquisition, or actually selling ads were.
  • We didn’t understand the industry. We didn’t know what kinds of content advertisers would pay for. We didn’t have good insight into what kind of content people wanted to watch, either.
  • We relied on proprietary hardware that we were going to mass-produce ourselves. Smart angels told us to drop the hardware and figure out how to do it with commodity equipment, but we wouldn’t listen because we thought hardware would be easy (or at least, doable). Ironically, months after we were told this we switched to using a laptop.
  • We were trying to build a “hits” based business without any experience making hits. We knew a lot about websites, but little about content creation. Smart VCs (who took our calls because Paul referred us) told us as much: nobody really likes investing in hits based businesses, because it requires the continual generation of new hits to be successful (instead of, say, building a platform like eBay or Google whose success is built on masses of regular users).

How did we get as far as we did?

  • We were passionate. We honestly believed we could create a new form of reality entertainment. Put to the side that we had no experience with creating video (or any kind of content), by God, we were going to make this work.
  • Early stage investing is often about the people, not the idea. Paul has said as much about what he looks for. As two-time YC founders he knew that we worked well together and even if we were working on something totally inane we were going to stick it out with the company and iterate until we found a business model.
  • We sold the shit out of it. Everyone we knew was excited for Justin.tv. Why? Because our excitement was infectious. That’s how we got Kyle to drop out of school. That’s how we got Michael to quit his job and move across the country.

Ultimately, the show failed. But all told, I’m thankful every day that things went the way they did. Why?

  • We built a strong team. The four of us started, and the four of us all still have leadership roles in the company. Along the way we recruited the smartest engineers and best product designers we could find.
  • We were willing to learn, and to pivot. After quickly realizing the initial show wasn’t a sustainable model, we decided to go the platform route, and built the world’s largest live video platform (both on the web and in our mobile apps, which have millions of downloads).
  • It got us started. Some people wait until the stars are aligned before they jump in. Maybe that’s the right move, but plenty of businesses get started with something that seems implausible, stupid, or not-a-real-business but turn into something of value (think Groupon). If we hadn’t started then, would we have later?

Today, I’m more excited about Justin.tv than I’ve been at any time since we launched the initial platform. Why? We’re taking everything we’ve gathered and learned over the past four and half years building the largest live video platform on the Web (17 million monthly unique visitors in Dec according to comScore’s MediaMetrix), and applying it to tackle a new generation of problems in mobile video. Our world class web and mobile engineering team, all of our product development knowledge, our substantial, scaled video infrastructure, and everything we’ve learned about building engineering teams has all been put to work on a new app that we think is going to change everything.

Our new app is called Socialcam, but that’s another story.

Photo by Terry Chay