Twitter CEO Evan Williams has just taken the stage for his Chirp keynote. I’ll be liveblogging his talk below. Update: the full talk is now available and embedded below as well.
- Twiterrific helped the Twitter team realize what was possible from third parties, had people telling them that it made them like Twitter.
- Twitter gets 3 Billion requests a day through the API
- According to comScore, that’s about the same traffic as Yahoo (though it’s not an apples-to-apples comparison
- Growth: 1,500% per year
- 175 employees
- Almost a year ago to this day, Ashton Kutcher decided to race CNN to a million followers, then we get a call from Oprah’s people because Ashton was going to come on the show to talk about the race. “We might want some help with the show, to make sure things are going right. Is Biz available? No… but I’m (Ev) available… They said well, you’ll work. This was Wednesday, for Friday. Then they called Thursday and said, we changed our mind, you’re going to be on the show. And then 4chan attacked us…
- “We’re out of chaos now. It feels good. And what this means is more organized and capable company that you can rely on. It also means we can give you more guidance than we’ve been able to before”.
- “The open exchange of information has a positive impact on the world”.
- We did some deals last year with Google, Microsoft, and Yahoo. The prospect came up, should we license the public data stream to these search guys? This wasn’t an easy decision — some board members/investors said “are you crazy?”. And we felt, we’re all about openness, and we believe openness creates more value. But this was perceived that this was the strategic monetizable asset of Twitter. There was concern we were giving away the farm. But we realized the value was to maximize value to end users. Wouldn’t more value be created if we had more guys working on this problem (these guys have more experience with search, etc). Realized these deals were the right thing to do”.
- 4 core priorities
- First is infrastructure: Scalability, quality of service, rapid development are key. This is where we’ve spent a lot of our time. In the last few months we’ve made tremendous progress. We have FlockDB which stores 13 billion edges, we started open sourcing in the last few weeks. Another project is Snowbird, which is moving to Cassandra for tweet storage.
- Last six months reduced 500 errors by two thirds. We’ve reduced Tweet delivery failures by two orders of magnitude. Also shortened Tweet delivery time (used to take a few minutes for people w/ millions of followers).
- Second priority: Friction-free: Twitter is too hard to use.
- We’ve known this (that it was too hard to use) for a long time but we were growing to fast to address it. If you type “I don’t get” into Google, second result is “I don’t get twitter”. (First is “I don’t get drink I get awesome”).
- Goal: Fast, obvious, easy.
- We have a new team that is about getting users from awareness to engagement. We weren’t doing this very well before. They’ve revamped signup process, and has increased retention by 20%. Much better homepage. This is a tough problem because Twitter is different things for different people.
- Mobile. Today around 37% of active users use Twitter on their phones. More than most other web services, but that should be 100%. There should be people using Twitter only on mobile, eventually. Twitter is built for mobile. Location, etc. Mobile is where that will happen. We have deals with 65 carriers, have to get deals where we aren’t paying for every SMS. Now carriers are making deals with us. I think there is a lot of untapped potential with SMS. You can build SMS services, use Twitter’s SMS reach which normally costs a lot of money but we have these deals
- BlackBerry and iPhone. We announced last week, we now have Twitter apps on the BlackBerry. We found that we had to have a core experience on these platforms just like we have to have one on the web. Or else we’re failing the ecosystem because we’re not getting as many people started/engaged. To underscore the problem. We did a user test a couple weeks ago. *shows a video of a young woman searching for Twitter on the App Store*. “What am I looking for, she asks herself”. Mentions Twitterific… still looking… “Is there a better way to find this, I don’t see one called Twitter”. “I’d want a free one…”
- First three days the BlackBerry app, 7-8% of new signups in the last three days come through BlackBerry app. Yesterday BlackBerry app represented 1.7% of tweets.
- The best thing we can do for you guys (devs) is grow user base).
- Next priority is Relevance
- There are about 55 million new tweets being created a day. What does an average user read in a day? It’s a very tiny fraction of what’s available.
- 600 million search queries a day, vast majority through API. Most of it has just been chronological. We want to improve relevance. We have top tweets (show a handful of most noteworthy tweets at the top).
- Location. We started implementing some location features in the API last year. Lat/longs to start with. But we are announcing today we’re launching Points of Interest. Actual places, not just lat/long. You’ll be able to click on Palace of Fine arts, see what everyone there is saying. That’s just the beginning.
- We want to make Twitter not just a tool to present more information, but things people care about.
- Revenue: Revenue is happening this year. Based on three characteristics. It’s not just about us, it’s also about the devs if they choose to participate.
- I want you to imagine the future a couple years, out. Twitter has come a long way, but we are really at the beginning, we’ll look back at this conference in a couple years the way we look back at the home page. A world with hundreds of millions of users, new capabilities in the Platform. That’s the world I want you to pour creative minds into.
Have a new tool called ‘Murder’ Killer improvement in deployment. Uses Bittorrent to transfer lots of files. Went from 40 min. to twelve seconds for deployment transferring of files
Ev’s Takeaways: Twitter is evolving. The goal is to serve users. There is much left to invent.
Q: Can you make acquisitions without alienating developer base?
Ev: I hope so. There is always a complementary relationship and a tension between platform devs. It’s not just acquisitions, it’s also what we build. Hopefully it reaffirms value they’re creating. Clearly a tension, that’s why were trying to share more information so people know where best to build.