In 2002, Caterina Fake (her real name) was developing a video game with her husband, Stewart Butterfield, when they decided to drop photos into instant messaging--akin to adding pictures to conversations, they thought. The idea blossomed into Flickr, which launched in February 2004. Initially it didn't proliferate. But once they added "tagging" it exploded--it was well-timed to make the most of the growing popularity of digital cameras and the ever-expanding blogosphere. Bloggers used Flickr as a photo repository and others gravitated to it so they could share pictures. The site that began as an afterthought to Fake's and Butterfield's multiplayer game matured into the fifth most popular site on the Internet, generating revenue by charging a fee to heavy users who wanted a place to store their multitudes of photos. A year after launch, the formerly debt-ridden couple sold Flickr to Yahoo for $40 million. Fake's latest venture is Hunch, a service that combines artificial intelligence with Web 2.0.
Penenberg: Fake is real, right?
Fake: Yes. I can't tell you how many times I've booked an air ticket only to get to the airport and find out they killed my ticket because it goes into the system and the program tosses a ticket that says "fake" on it. Twice I've gone to the counter for a KLM flight through Northwest and have been rejected. They say, "You don't have a ticket." I give them a confirmation and after some investigation I learn my ticket has been cancelled because the system deleted it. For a while I couldn't join Facebook because of my last name. During the registration process I was asked for my real name and when I wrote "Fake" it rejected me. Finally a friend working for Facebook took care of me.
Penenberg: There's both a good and bad side to virality. Products with viral hooks that are so strong they coerce people to sign up--in order to achieve a huge initial viral rush--are obviously bad. Not only do they alienate users, they don't lead to a sustainable business. On the good side, you have organic growth, which comes as a natural byproduct of something that spreads simply because people like it--eBay, Hot or Not, and Flickr. I can't think of an antonym for it.
Fake: How about brute force growth?
Penenberg: That's good. Maybe we should trademark the term.
Fake: You could also advertise your way to growth.
Penenberg: Of course, but that's not what you did with Flickr. Tell me how it all came about.
Fake: My background isn't in social software, it's in online community, social networks, personal publishing, blogging, self-expression on the Web. I got on the Internet in the 1980s, and the magic moment for me arose from my being a literature geek, especially Dante and Shakespeare. This was back in the day when you could have fruitful conversations in chat rooms and everyone who was online was at a university. They were all power users; everyone knew how to operate their machinery. One day I had a fascinating conversation with a Borges scholar in Denmark. My initial experiences in the things that made me love the Internet were primarily social--connecting to other people, and people that you wouldn't otherwise be able to find, like Dante and Borges scholars. By the 1990s I was blogging--this was before there was blogging software, and they didn't even call it blogging. It was an online diary or notebook. At the time a personal Web log was seen as weird or creepy. It seemed like a fringe activity--kind of weird and strangely self-absorbed and very odd. And people didn't have photographs of themselves. There was no such thing as a profile online, it didn't exist.
Penenberg: When and why did the shift happen?
Fake: I think it happened with Friendster. While there were earlier social networks like Six Degrees, which was circa 1998, Friendster hit at a time when digital cameras were becoming ubiquitous and so was broadband. Suddenly it was possible to post photos, and that helped people create online identities.
Penenberg: Then it comes down to ease-of-use and accessible technology fostering a push into the online realm?
Fake: The act of self-presentation by regular people was now possible. Earlier you actually had to know how to use the tools, to be able to write HTML, to understand FTP to upload stuff to the Internet, and to have a server somewhere. With Friendster, people discovered that they love to present themselves in the most favorable light. This tells us something about them. They put pictures of themselves, friends, and family up, and described who they were and the bands they listened to. In essence, Friendster was an identity establishment exercise. Friendster was basically making available to everybody that experience I had with the Danish Borges scholar. That sense of "wow," that manna-from-heaven feeling when you find something or someone or some person you lost touch with, like your old classmate. Or you discover somebody else who feels the same way you do, say, if you're the lonely gay teenager in Indiana. All those people looking for connection, that perennial human desire. It's just insatiable.
Penenberg: Humans have a built-in desire to connect. That's what language is all about. What's the point of being able to speak if you can't speak to anybody? Online is just an extension of this need we have. Before you started Flickr, you and your husband Stewart were working on a game.
Fake: We got married on June 2, 2002, in Victoria, B.C., and started the game two weeks later. I was a freelance Web designer, and Inner Hole Research, where I was working, had just shut down. I had been involved in this startup that had been spun out of there and which was doomed from the start. It was social software, a collaborative storytelling tool where you could post images and characters. You moved these characters around and created captions that looked like a cartoon filmstrip to tell stories. It was really cool--but it was doomed from the start. The technology was not up to par; we were building it in Flash but it wasn't sophisticated enough to handle this kind of collaboration. The premise of it, though, the collaborative storytelling, was a great idea.
Penenberg: It might have been hard to come up with a revenue model...
Fake: That, too.
Penenberg: What made you move from that to a video game?
Fake: Game Never Ending came out of an earlier game called NeoPets, a children's game that adults also played. It was a series of mini-games where you accrue points and can acquire objects--houses and all kinds of stuff. These things are taken for granted now, but at the time nobody was doing it. Game Never Ending emerged from this because we saw that it was an incredible context for social interactions. The growth of NeoPets was hampered because children need a protected space. But a game built for adults, where communication could come more freely, would mean the social interactions would be much more fruitful. My background had nothing to do with games, but the social aspects of this became immediately apparent to me. I also liked to play games on the Internet and socialize online. We created this as a social game, as a context for social interactions.
Penenberg: But you weren't able to attract venture capital funding.
Fake: No, we weren't. We went everywhere, but in 2002 it was nearly impossible to raise money and nobody understood what the hell we were doing. They would ask, "Is it shrink-wrapped? Can I buy it at Wal-Mart?" They didn't get the concept of an online multiplayer game.
Penenberg: Can you boil down the essence of the game to a few sentences?
Fake: Starting with a large map, you could explore the terrain and find objects and create things. It wasn't a first person shooter. It was more like World of Warcraft. It was social, fun. To give you an example, you could find a wombat whistle and blow it, and all these baby wombats would cuddle up against you. You could find elements of things and put them together. You could find venture capital, hype the Web, and put it all together to create a new economy and then go into a crowded room and deploy your new economy--and then everybody would go broke and you would get rich. You could cheer people up, too. There was a kind of magic sparkle powder, so if you went into a crowded room and sprinkled it on everyone, their mood would lift. There were all sorts of metrics...
Penenberg: Sounds like a stoner's dream. How did Flickr rise out of this?
Fake: We were unable to raise any money so we collected about $250,000 from friends and family and bootstrapped the whole thing. No one had gotten paid in six months except the employee with three kids. We were out of cash, frankly, and what happened was the backend development was about six months behind front-end development. We were finding ourselves about to go out of business. It was pretty clear that's where we were headed. We were broke.
Penenberg: I sense a eureka moment coming any second now.
Fake: Stewart had this idea that we could take what we had learned and develop it. A big part of the game was instant messaging, so we could take the instant messaging part of it and develop a client that we could drop photographs into. In essence it would be like adding photos to a conversation.
Penenberg: It would be as innovative a concept as inserting hyperlinks into online chat room conversations, a concept that came some years earlier and which beautifully illustrated the interconnectivity of the Web.
Fake: We built an IM client in Flash that you could drag and drop photos into a window. We thought it was pretty neat. Nobody could do that at the time. But it wasn't a successful product. It wasn't destined to be a standalone product. It was a feature. Another problem was that people were on different IM platforms--some were on AIM, others on MSN.
Penenberg: Which meant that users were balkanized. A person on AOL's messaging platform couldn't communicate with those on a competing one.
Fake: So we started it in December 2003, launched the IM client in 2004, and it just fizzled. It was kind of sucky; as a business it just didn't work. It was simply a cool feature for AIM. What was interesting, however, was that we had a permissions structure for photographs to be shown and shared. We had a profile page, which I don't think has changed much since the outset.
While there were photo share sites like Ophoto, Shutterfly, and SnapFish, they took their metaphor from prior technology. They were designed to be reminiscent of pre-digital photo albums. But Flickr was designed the way it was because it arose from digital technology--in the same way jellyfish are designed the way they are because they live in water. In contrast, Shutterfly and the others foisted themselves on the Web. They didn't adapt to the technology.
Penenberg: But it took tagging for Flickr to truly take off, right? Because without it you wouldn't have been able to organize photos online.
Fake: Yes. Joshua Schachter, the founder of Delicious, created the concept of tagging sometime in 2004. It grew out of his using bookmarks to organize Web sites. But tagging, I think, is best used in photographs. The problem it confronts is that the photograph has become delaminated from its delivery mechanism. In other words, you have a hard drive crammed with digital files, but dot jpeg-named files are meaningless unless you have a context in which they can propagate and other people can add information. By being able to crowd-source information on photos it becomes a democratic way of organizing information, where anyone can add works and comment on them.
The decision to make all the photos public versus private was motivated by the fact that conversations are where metadata happens.
Penenberg: What does that mean?
Fake: Say you search for "Caterina smiling." If our search is working properly, this picture or another photo of me smiling would come up. That comes not only from the tagging, but from what we call the conversation around the photo. The tags, title, description, comments--all of that stuff.
Penenberg: How did Flickr grow virally?
Fake: Stewart and I got into this argument in the early days of Flickr, before we could offer paid accounts. We had not set up any merchant accounts, not even a PayPal merchant account--nothing. He said we should let people join for a certain amount of time. I said, since these are early adopters, one of these people is going to be the mayor of Flickr--you never know. Georgina (George) Oates, an early employee of Flickr, and I greeted everyone who came into the instant messaging conversations or posted on the site. We said, "Oh you know, I notice you're into Norwegian Metal and I noticed you live in Pittsburgh--here's another person from Pittsburgh. Communities take on the kind of character of a party at the outset and you need to be a good host.
Penenberg: There's nothing you built in to be viral?
Fake: That's not true. We did that, and it was a huge part of building an online community. If you come to a party and nobody offers you a drink, you leave to go to a happening party somewhere else. You need to care, to work being the host. You don't do that forever, but Stewart said we should let everyone in. I thought we should make it so that if people invited five friends, we'd give them three months free. We didn't charge at the time but we didn't hide the fact that we intended to charge at some point. If it was always free, no one would invite their five friends. But if it's free for a limited time or because we hadn't set up a PayPal system, it's valuable enough for each user to invite some people.
Penenberg: In essence, you incentivized the invitation process. Do you remember how quickly Flickr grew?
Fake: For the first few months Flickr was growing 75% to 100% month over month, mostly because it was small enough. That eventually leveled off to 30% to 50% month over month growth. My "wow!" moment in terms of watching the growth was when I found this enormous group of people from the UK that I didn't know. That was two or three weeks in. They were all connected to one other. Let me show you this graph from April 2005--it's all the people who were connected.
The Flickrverse, a graphic that depicts the relationships between Flickr's first 2,367 people (created by GustavoG).
Penenberg: What was making Flickr grow like this?
Fake: It was a confluence of stuff: bandwidth, tagging, and a very social medium--photographs.
Penenberg: It also grew in tandem with the exploding blogosphere. As the number of blogs grew, the number of bloggers needing a place to store photos grew with it. This is the quintessential "organic growth" that took place because people liked the product. And the more people used it, the more others got exposed to Flickr. You're exposing people who click on the blog to Flickr, and there's tagging on Flickr--so with all the tragedies, people could take pictures and check photos. If I had a family album, I could put it on Flickr and my family could check it out. You have several viral loops going on, but they're soft loops. Were you shocked at the growth? You had a revenue scheme, right?
Fake: Yeah, there were paid accounts based on storage limits. You could store 200 images for free. It's important to have free accounts because it's social, but for a fee you could store more, crop photos, and get stats.
Penenberg: You must have been racking up serious bandwidth costs.
Fake: Not really.
Penenberg: Was scaling an issue?
Fake: Yes and no. We didn't have a Friendster situation where we suffered a lot of downtime. Flickr was a well-engineered application--it was very stable. We had to decide if we were going to attract venture capital or go for acquisition.
Penenberg: Were you profitable when Yahoo bought you?
Fake: We were probably at about break-even.
Penenberg: Do you feel like you sold out early?
Fake: Well, looking at it now, it looks like Flickr could be a billion-dollar company, but who knew that down the road? At the time Blogger had just sold to Google. In the context of 2004, Flickr was remarkable, but who was to know?
Penenberg: Are you sorry?
Fake: Well, you could sit around and cry but if you throw a rock in Silicon Valley you find somebody that could have been a contender. My life is not so bad; in fact, my life is pretty good. I think it's a waste of time to sit around and think of what could have been.
Penenberg: Certainly, besides a nice chunk of change, Flickr has brought you a level of fame.
Fake: I design Web sites, its not like I'm Angelina Jolie. I think you need to tune that stuff out, otherwise you're never going to be able to build stuff, be an entrepreneur, take risks, fail. That's part of entrepreneurialism, being able to launch a stupid IM client. The hardest part about Flickr being successful is wanting to do it again. I think there's a benefit to being one of six people that no one knew. No VCs would return our calls and we were broke and bootstrapping it and operating under the radar so we could focus on the most important things: the product, the users, what we were building. There's all this noise, the tech-crunch, which you have to tune out if you want to build good product. None of that stuff is additive; it all takes away from building a product. You try a lot of things and you don't know what the hell you're doing. If you're actually inventing something you shouldn't know what you're doing.
Penenberg: Being an entrepreneur is never easy, is it?
Fake: Its tough. Flickr was a product that was the right product at the right time done in the right way, and those three things are really hard to nail.
[Photo by Joi Ito]