The Accidental Launch

We accidentally got 10,000+ users in 24 hours, and funding from Y Combinator just a few days later. This post tells that story.

We were determined to take part in Y Combinator, so we spent weeks crafting our entry and polishing Rapportive. At the start of March, we were finally ready. We held our breath and clicked "Submit". We looked at each other, relaxed, and slowly started to breathe again. A few hours passed uneventfully. We were in no way prepared for what happened next.

Somehow, the press had found us. TheNextWeb ran the first piece. ReadWriteWeb picked it up after that. Then Lifehacker. Then WebWorkerDaily. We had headlines like: "Stop What You Are Doing & Install This Plug-In." Our twitter account was aflame with thousands of mentions in just a few hours. We had accidentally launched.

We saw our user count grow from 5 to over 10,000 in 24 hours. I had a case of beers in my drawer in case we ever needed to celebrate anything. We drank all of them.

I stayed awake for two days straight: the emails didn't slow down, the tweets kept pouring in, and new Skype chats would appear as soon as I'd finish old ones. But we were determined to quickly respond to every single last email, tweet, and chat, so we soldiered on.

The next day, investors from across the world started contacting us with offers of funding. These weren't just any old investors; these were some of the best angels and venture capitalists in the world.

We didn't have time to wait for the normal Y Combinator interview, which would have happened a month later. I contacted Harj, Venture Partner at YC, and they offered to do the interview over Skype. (I vaguely knew Harj from our university days — it's a surprisingly small world.)

A few days later, Martin, Sam and I were huddled around around a laptop talking to pg, Jessica and Harj. They weren't quite as huddled, so we spent most of it talking to pg's legs. We talked for half an hour, but I felt like it passed by in an instant. A few minutes later, we had our answer: Y Combinator would fund us!

We celebrated in the traditional British manner. When we were next coherent, we booked a fundraising trip to the Valley.

Lessons Learnt

We did several things that worked well during this phase:

  • Offer surprisingly great service. Most companies deliver terrible service, and users have come to expect it. Surprise them. Make it abundantly clear how users can contact you. Monitor all your channels. Respond to people as soon as you physically can. Thank everybody and go the extra mile. I personally find that it really helps to smile, even when the user is thousands of miles away and on the other end of a tweet. We use a shared Gmail account for email support, and CoTweet for twitter. Our YC batchmates rave about Olark.
  • Use a feedback forum. Make the forum really easy to find. Include links to it from your product. Make the links especially visible when the product isn't working properly. If your forum provides single sign-on (so users don't have to create new accounts) then use it! We use UserVoice and have fallen irrevocably in love with it.
  • Release early. We didn't choose to release early: it was a complete accident! But in hindsight it turned out to be very useful. Our feedback forum rapidly filled up. We quickly learnt peoples' likes and dislikes, and prioritised building what people want. If you don't release early, then you might build the wrong thing and you won't find out until much later. Even if you build the right thing, somebody else might build it first and steal your thunder. So get out there.
  • Be ready to scale. You never know when traffic will hit. Now I realise that "be ready to scale" may sound like classically bad advice, but cloud computing has changed the economics. You can be ready by simply choosing the right hosting provider. If we were on a cheap VPS, we would have crumbled to pieces like Cobb's limbo in Inception. As we were on Heroku, we could simply increase the number of dynos. I still vividly remember when our traffic hit. I was away from my desk, so I reached for my iPhone and dialed us up to 20 dynos using Nezumi. A few seconds later, we had scaled.
  • Build for the press. It turns out that Rapportive works exceedingly well for technology bloggers, because they spend so much time corresponding with people who have significant online presences. It is not worth building functionality only for the press (unless, of course, they are your target market), but it is worth being aware of this effect.
  • Build early. This advice is specifically for companies applying to Y Combinator: start as early as you can, as the deadline will come soon. The most impressive thing you can do is make something that people want.

One of our favourite books is Founders at Work, a collection of interviews with founders about their early days. We're now collecting stories of our own, which we will post in a series, Rapportites at Work. This post is the first of the series.

Update: It turns out @plc tipped @zee, which sparked off all the press — thanks Pete :D

Jobs: We are looking for a talented developer to join our team.  Could it be you?