Stewart Butterfield on Flickr

by Richard Koman

We collect images with cameraphones and so forth, but we have no good mechanism for advancing them out into the world. Here's a mechanism for batching them into a locked-and-loaded tool for firing them into the world.
-- Ben Cerveny, Ludicorp's Itinerant Philosopher

At the O'Reilly Emerging Tech Conference in 2004, a startup called Flickr introduced a funny little social networking app that let you upload digital photos into chatroom and IM conversations. While the original launch met with rave reviews from attendees, the Flickr team kept adding features and evolving the service. By July 2004, they had achieved a critical mass of features, and Flickr was becoming the hottest thing on the net. In January 2005 alone, Flickr has been profiled in Wired, The Wall Street Journal, Forbes, and the Washington Post.

As of this writing, Flickr boasts 270,000 users, four million photos, 30 percent monthly growth in users, and 50 percent monthly growth in photos. (As a datapoint, when I interviewed Stewart about ten days ago, those numbers were 240,000 users and 3.5 million photos.) And these numbers don't even begin to tell the story. Flickr is a phenomenon, a fundamentally different way of using digital photography and the Internet. Flickr is simply the manifestation of the perfect storm of camera phones, consumer broadband, blogs, RSS, and folksonomy tags.

Several of these factors come together in a simple post by Cory Doctorow on the Boing Boing site.

To quickly deconstruct, we're looking at someone discovering an extraordinary photo by subscribing to an RSS feed defined by a metadata tag and posting it to his blog. While Cory copied the photo over to Boing Boing's site, it could just have easily been served by Flickr via their integration with major blogging software.

Flickr is part of something else too, something radical: the massive sharing of what we used to think of as private data. Photos, bookmarks, and journals used to be considered personal. The social networking revolution--which encompasses everything from Flickr and to blogs and wikis to P2P itself--encourages us to share everything.

The sharing imperative includes not only stuff but also ideas, such as how we think about things (tags) and how we program (APIs). From this openness spring galaxies of supporting applications, revolving around a core web service like Flickr.

O'Reilly Emerging Technology Conference.

At this year's ETech conference, Flickr CEO Stewart Butterfield will discuss two aspects of the Flickr phenom: building a business on an open API and the power of user-created metadata. I talked to him by phone as he travelled by taxi from the Vancouver airport to his home.

Richard Koman: Stewart, Flickr is a bona fide phenomenon. What's it all about?

Stewart Butterfield: There's a core difference between Flickr and what people have traditionally thought of online photo sites like Ofoto, Shutterfly, Snapfish, and Yahoo Photos. Their business is photofinishing; they give away sharing as a loss leader to get people into the funnel, to get them to print. That's good enough for a lot of users--they take some photos at a wedding, and they want to upload them to a web page and share that URL with friends.

However, there's a sizable and growing minority of people for whom that's not good enough anymore; they just have thousands of digital photos, and they've had digital cameras for several years. Not only that, but they live pretty connected lifestyles; for most of the day, they have access to an Internet connection, and at least one if not two or three photo capture devices. For instance, I have a new digital camera I got this Christmas, an older digital camera, and two camera phones. I'm closer to the edge case, but I'm not all that unusual.

So the proliferation of capture devices, the always-on lifestyle, and the fact that people are now more familiar with computers and the Internet, very simply leads people to be more comfortable with interacting with each other online. It's not weird to publish a stream of your photos and have people tune into that.

Koman: Publishing a stream of your photos is not an obvious thing to do with your photos, and it creates really unforeseen social behavior around photos in Flickr. In what ways is Flickr more of social networking software than an online photo service?

Butterfield: Some people really are primarily using it for sharing photos with friends and family. They have contacts in the system and select certain photos to be only available to friends or family, only occasionally making a photo public. The majority, though, are making almost all of their photos public. Of those 3.5 million photos, 82 percent are public.

People are finding all these different ways to interact with each other through photography. There are a lot of Flickr groups (shared spaces that users can join) that exist just for adding photos to a themed pool of photos, and some of those can be strictly delineated. An example of this is a group called squared circle. We have a feature whereby pretty much any collection of photos can be turned into a slideshow--the results of a search, anything tagged with a certain word, in addition to photosets that people deliberately create--any collection that Flickr can be caused to produce can be turned into a slideshow. So the rules of the "squared circle" group are that the photo has to be cropped to a square shape, the image has to be circular, and the edges of the circle have to touch the edge of the square. And the only reason to do that is that it creates a really cool slideshow, so now there are more than 600 members and more than 3,000 photos, and it is a beautiful thing to watch. There's a lot of emergent behavior like that.

Squaredcircle Photos

There's also behavior that emerges out of the global tagging system. People who would be reluctant to provide metadata most of the time do so on Flickr because there's a payoff for them. A) other people see their work--work is probably the wrong word because I don't think most people see it as work in a serious artistic sense-- but people see what they're up to, see what they're creating. And B) because they derive some pleasure from building value in the global collection.

So, partly for those reasons and partly because using the social network you can give people permission to add metadata to your photos, 71 percent of the photos have some kind of human-added metadata that was added in Flickr. That's extremely high, even compared to software like Adobe Photoshop Album, which is designed from the bottom up to facilitate the adding of metadata. When you're doing it for yourself, it's like a chore, its drudgery. When you're doing it as part of a community, in a collaborative way, it can still be a little bit of work, but the payoff is so much larger.

Squared Circle/Flickr Collaborative Poster Project
Squared Circle and Flickr Collaborative Poster Project

Koman: About the tagging, Flickr and are usually mentioned in the same breath of having pioneered this concept of "folksonomies." Adam Mathes calls this a shift in who creates the metadata from the author to the user. Lots of and Flickr users describe an object--a web page or a photo--in the way that makes sense to them, with no interest in how the author might have classified the page.

Information experts like Lou Rosenfeld complain that folksonomy tags, while interesting, don't support searching or browsing particularly well, don't support relationships, don't offer true classification, and can't even cluster synonyms. So why does this uncontrolled mass-tagging by users makes sense?

Butterfield: First, going back to the comparison for a second, we were definitely directly inspired by There's an interesting difference, though: because people are adding URLs to, and many people can add the same URL, you end up with multiple ways of tagging the same thing--different people's vocabulary for the same item. On the other hand, each photo a user uploads to Flickr is unique and belongs to them; however, more than one person can tag it. (For people who have large personal networks in Flickr, they definitely find that they upload a photo and go to sleep, and in the morning there are tags all over it.)

The complaint that it's uncontrolled and it's not going to be captured in a consistent way to me is really irrelevant. Because tags are first and foremost for people to organize their own photos--and if they weren't, it wouldn't work. It's a happy accident that the whole global collection emerges. And let's say it's only 50 percent accurate and complete and let's say right now we have 10,000 photos tagged "Italy;" it might actually be 20,000 photos that should have been tagged "Italy," but who cares? No one is going to look at all 10,000 photos, let alone 20,000 photos. And in six months, it will be 50,000 photos instead of 100,000 photos.

Now there are some things that will annoy people, like singular versus plural forms, homonyms and homographs, and alternative spellings. But a lot of this is easy to deal with when you have enough data. In general, disambiguation is a harder problem.

We've done step one of relatedness of tags, which is just cluster analysis of how people tag, and then we suggest related tags. So if you tag "Italy," it will suggest "Rome," "architecture," "travel," "food," "Europe," etc. And it works astoundingly well. People usually think there's a human editor, but it's just cluster analysis. And there is a lot more of that coming--some of which we're hoping to show at the 2005 [ETech] conference.

Koman: Do you see Flickr and its open API as representing a next generation of web services? What things can developers learn from what has happened with Flickr?

Butterfield: On the strictly practical side, I think we had one person inquire about using the SOAP version of the API. I don't know if any apps were actually built. There is at least one application built on XML-RPC. But all the others--I don't even know how many there are--are built on the REST API. It's just so easy to develop that way; I think it's foolish to do anything else.

It's really valuable for any new product or service to reach the hyper-geek audience, who are particularly influential. And for them, the open API is a sign of good faith, a sign that your photos and your data are not going to be locked up in Flickr--even though we don't currently offer a feature to download your photos to your own computer (we will), you could develop one.

It makes a difference for us as a business that other businesses are interested in working with us because they can tell up front how much work it's going to be. Basically third-party apps fall into one of two categories, useful or cool, and some things are both. Useful would include uploaders for a bunch of different platforms, a screensaver that pulls in your contacts' most recent photos, and an application called 1001 for OS X that grabs the most recent photos from your contacts or specified tags, and it pulls from them like an RSS reader. And then there's a bunch of applications that are just cool, like one that takes photos tagged with different colors and arranges them into the shape of a rainbow.

It makes a difference for us as a business that other businesses are interested in working with us because they can tell up front how much work it's going to be. They can have their engineers look at the API and say, "This is what I want to do, how long do you think it's going to take?"

On the flip side, there were a bunch of people we were looking to work with for printing and the people who didn't have good, clean APIs, we just couldn't work with them. We didn't want to put in the cycles in dealing with their integration engineers on a one-off project and have them develop the functionality we needed and then start implementing. We wanted to get a document that specs it out and start going. We actually chose a print partner based on the quality of their API.

Koman: In the write-up for your web services session at ETech, you say, "Capturing the creative energy of the hive can be scary. It requires giving up some control, and eliminating lock-in as a strategy." Tell me some more about that.

Butterfield: Ofoto is a pretty good example. I don't want to pick on them too much, but they create a pretty artificial kind of lock-in. When you upload your pictures to them, you might upload a three- or four-megapixel image, but all you can get back from them is a 600-pixel image; if you want to get the original back, you have to buy it on a CD. There's no way to get it out because if you got it out, then your friends and family could get it out and print it out at home, and they're in competition with Lexmark and HP as well as the other online photo services. So that's one aspect of it.

There's also a tendency to want to capture all the value that's being generated or will potentially be generated by new business. What I mean by that is, we don't explicitly allow commercial uses of the API yet, but we definitely plan to. And we know that there are people working on products based on our API that we want to do, but outside developers will get to it first. What letting go in that context means is letting go of all the control you have over users by being the one who owns the database, because other developers can generate businesses and products that hook into you, and that takes some value away.

Ultimately, any user of any app built on the API has to be a user of Flickr as well. It also just introduces for a company at our stage a lot of headaches and risks, in that people are, excuse me, but shitty coders; they're not responsible in their use of the API or they write buggy code that hammers us. And that's happened a couple times. People can hammer you in other ways: they can scrape you, but they can hammer you a lot harder if you're exposing methods to the things that are most database- or CPU-intensive.

It also exposes you to a bunch of risks. For instance, this wasn't built on the API, but Technorati introduced a tag search feature (which pulls blogs, Flickr pictures, and pages sharing the same tags onto a tag-centric page). There're definitely people who feel alright about making their photos public and having it available to other people in the context of Flickr but who feel uncomfortable about them showing up in other places like that. There are some users who would like the option to include robot text files for Google to not index their photos because they don't want them showing up in image search.

But the API just compounds that; it exposes us to all kinds of gray areas, things we might have run into eventually but we run into them a lot quicker this way. Even things like that rainbow app, using pictures without people's permission ... the consensus is no one's profiting off of it and its fair use, but it's another thing we have to deal with that Ofoto doesn't. To some degree, its inherent to Flickr whether we have an API or not, but its driven faster because of the API.

Koman: It strikes me that Flickr could be viewed as a huge stock photo bank if only you could have a conversation about rights.

Butterfield: We actually just heard a story from a user, the first one I know of, where someone bought a photo through Flickr. A photo editor at the Boston Globe was doing a story on Wal-Mart's warehouse center. A user had just taken a picture of it and put it on Flickr. The photo editor did a search on Flickr, found it, contacted the user and said, "OK, we'll pay you $150." That's a real opportunity, whether that's something we're able to facilitate and take a cut of, or not.

Koman: How long will you be in beta?

Butterfield: There are a few more features we still want to do--printing is one of them. We need to do one more round of infrastructure enhancement. We're essentially growing it as fast as we can and we want to be able to get a few steps ahead on the infrastructure side before we can really go and say, this is it. We don't really have a date yet, maybe first half of this year.

Koman: So, what's the business model?

Butterfield: There are two halves to it. The first is that things that are driven by individual users, so subscription fees for premium services. Actually, we're doing fantastically with that so far; we're really happy with the results. There are other things like printing, CD and DVD backups, and gifts--so if you take a bunch of photos at a wedding, you can pull them with friends and families photos and give them to the bride and groom. Those are probably secondary to subscription fees.

The other side is monetizing the whole collection of photos, essentially advertising. If you look at photos that are tagged with "Italy," you'll probably see ads for hotels in Italy, tours of Piedmont and whatever. There's a lot of contextual advertising you can do against big collections. And in addition, a bunch of different marketplace activities, allowing people to sell prints of their photos, letting people sell images to news media, and stuff like that. And probably all of those are secondary to advertising.

Koman: What parting advice would you give web service developers and Internet entrepreneurs?

Butterfield: Be opportunistic. Be open, really listen and assess what kind of response the product gets and evolve in a way that makes sense. Also, be open to working with people's existing products. We didn't make a Windows or OS X-based photo organization software package because Photoshop Album is already very popular, IPhoto ships on every Mac, so why try to fight that? We'd rather work with that.

Richard Koman is a freelancer writer and editor based in Sonoma County, California. He works on SiliconValleyWatcher, ZDNet blogs, and is a regular contributor to the O'Reilly Network.

Return to the O'Reilly Network.