Given all the talk about the talent war going on in Silicon Valley, O’Reilly had to ask what Facebook’s magic was for attracting great people.
Zuckerberg’s response: engineers want to have the most impact possible. Small startups let engineers have an internal voice, but they don’t reach many people. Longstanding successes like Microsoft and Google reach tens of millions of users, but are so large that individual engineers tend to get drowned out by the noise and bureaucracy.
Facebook is experiencing a golden era–small enough that any single engineer can be heard, but large enough to reach 500 million people. That’s a ratio of more than 1 million people for every engineer.
The challenge will be staying small and nimble enough to avoid becoming the next dinosaur.
Zuckerberg discussed a wide range of other topics as well, including privacy concerns and yesterday’s messaging announcement.
Our coverage of the talk is below:
8:06 PM ET: John Battelle starts by highlighting the messaging announcement yesterday, and asks Zuckerberg to explain why they did it. Zuck repeats a story from the press conference yesterday: high school kids don’t use e-mail that much because it’s “too slow.” What they meant is it’s too formal. SMS and IM are simpler. That’s what the new Facebook messaging system’s about.
8:10: We didn’t start with e-mail and add to it, we subtracted from it. No subject lines. They just get the message they want wherever they want.
8:11: O’ Reilly asks about the social graph and the e-mail namespace. Zuckerberg notes that this can help solve the classic e-mail problem of spam. Filters have gotten good enough that you don’t get the real blatant pitches for Viagra, but you do get a lot of random messages. The real way to solve spam and filtering is have a list of people you want to hear from. We have your friends, and their friends. We know who’s in your network and is likely to be sending you something.
8:13: O’Reilly notes this takes a lot of manual effort. What about using algorithms like Google? Zuckerberg believes that manual curation offers better results. Look at photos: 10 years ago, everybody would have said facial recognition is the way to recognise people in photos. Turns out, letting people tag them is much easier.
8:15: Battelle notes some people don’t like people being in groups. He suggests that Facebook doesn’t ask for permission, it asks for forgiveness. This happened with Groups, for instance. Is that intentional?
8:16: The friend graph is intentional. When you’re my friend, you can do certain things.
Battelle: but not export my contact data, unlike Google lets you do.
Zuckerberg: e-mail is different. There are some pieces of content that are in the middle of permissions. With photos that I’m tagged in, do you have right to share it? Same with e-mail addresses of your contacts–does somebody else have a right to see that? We’re not sure we’re right on this. The feedback we get from people is “we want control of what’s on our site.”
8:18: O’Reilly asks, so why not give people the option to download those e-mail addresses? Zuckerberg says the answer isn’t completely obvious.
8:20: Battelle: how do you handle inquiries from the U.S. Congress, which is “putting Internet privacy in the crosshairs.” Can you go and say “leave us alone, we’re figuring this out?”
8:21: Zuckerberg says that people are on the side of giving the user control, and that’s where they’re coming down. These are the kinds of debates Facebook has internally all the time, and it’s exciting to try and solve them.
8:23: Battelle suggests Facebook should be ready to drive its ad system off its domain, like what Google did with AdSense. Zuckerberg says they’re not ready yet–there’s still a lot left to do. Rather than pushing out any of these systems aggressively, he thinks that most industries will get rethought to be built around social and people. Humans are wired to be interested in people, so we built that system. More than 50% people on Facebook use it every day. More than 500 million users, 250 million use it daily.
8:26: Zuckerberg continues, we’ve had teams of two or three people building systems like Photos or Groups that compete with entire startups. They start simple–Photos wasn’t high res, you couldn’t organise photos in order. But what it had was the social component–you could share immediately with our friend. So what happened is that our photo product became the most-used on the Web. Bottom line: if you make something social, it remakes the whole space.
8:27: So in 2007, we pivoted and decided to build a platform instead. There have been four game companies, Zynga, Playdom, Playfish, and Crowdstar. Zynga’s market cap is more than EA. That’s a fundamental shift in the business. We should help play a role in reforming all those industries. We’ll get some value proportionate to what we put in.
8:30: O’Reilly notes that there are two possible models. You can build functionality into Facebook, or take the platform and enable a bunch of startups to build their own social services in a particular area. How does Facebook decide between pulling in and pushing out? Zuckerberg responds that they’re different things and it makes sense for people to do both.
8:32: O’Reilly asks to think about music. You could help Google build a social music platform, or you could build your own. Zuckerberg says that the platform model is preferable. Anything that doesn’t have to be built by us, they’d rather not build. Facebook is still a small company with a few hundred engineers, and will pursue partnerships. If Apple or Google wants to build a music service, they build it themselves. We believe it should be more decentralized, let the entrepreneur build their own music company, then integrate social into that platform. We think that will blow everything else out of the water.
8:35: Battelle: some people don’t trust you because they feel you have two much power. Do you feel that pressure when dealing with entrepreneurs? Is there really only one social graph, or is there room for multiple ones, LinkedIn, Twitter, and so on? And you’re still not integrated with Ping from Apple.
8:36: Zuckerberg: It’s a different way of thinking. Companies that are building games assume that every one of their users is socially enabled. We’re only now entering the zone where a lot of the best entrepreneurs will consider working with us, assuming that most of their users are on Facebook.
8:39: What about big companies like Apple? Would you work with them, or would that be too much of a threat? Zuckerberg says that Facebook’s built all this infrastructure out, and if Facebook’s going to support tens of millions of users, they need to understand how the partner is going to use that information and give something back.
8:41: Battelle jokes that Steve Jobs said that Zuckerberg’s demands for Ping-Facebook integration were “crazy,” and “only Steve gets to make those demands.” Zuck’s opinion: everything’s going to be remade to be social. “Get on the bus.”
8:43: O’Reilly asks about culture. Zuckerberg responds you can pick a few behaviours that you want to be different from other companies. Facebook has five values they write down. Two they’re focusing on now are “move fast” and “be bold, take risks.” Tech companies tend to get slower as they get larger, and they get replaced. We want to make this company operate as quickly as possible.
8:45: O’Reilly: is that part of your magic, how you’re sucking in people from other companies? Zuckerberg says that people want to make the most impact. We track the ratio of users to engineers, and we’ve been over 1 million for a while now. So on one end you have startups that are small but have few users. Google and Microsoft have the users, but they’re very large. Facebook’s in the golden spot right now–big enough to reach a lot of users, but small enough for engineers to have major impact.
Now he’s taking questions from the audience.
Q: What’s it like to be young and to be so far ahead of other CEOs and entrepreneurs?
A: I’ve made so many mistakes, any mistake you think you can make I’ve made, or will in the next few years. The Facebook story is a great example of how if you’re building a product that people love, you can make a lot of mistakes.
Q: Is education an opportunity for Facebook?
A: We have a 13-year-old limit, which cuts out a lot of time people are developing. A lot of problems in education are around technology and people sharing things, there’s a big opportunity there. It’s like enterprise software versus consumer. Enterprise software has a feature checklist rather than what’s really good to use. That’s why consumer products like IM are taking off in the enterprise–they’re enjoyable to use. That’ll be true in education as well.
Q: What about social commerce in addition to advertising? Selling stuff through Facebook, is that interesting?
A: Social apps are more engaging. That creates opportunities for people to build great businesses, whether advertising, commerce, or subscriptions. It all starts with building a product that people love.
Q: How much commission will you take?
A: You want to pick a number?
John Battelle quotes Ari Emanuel from last night: “fair is where you end up.”