The Flickrization of Yahoo!
How the founders of a hot young photo-sharing site are helping to change the focus of the search engine giant--and turning its fight with Google into a battle of man vs. machine.
By Erick Schonfeld

(Business 2.0) – "I have never seen so many people with cameras," says Jerry Yang. "It is kind of scary." It's a perfect September evening at Yahoo headquarters in Sunnyvale, Calif. Yang, co-founder and chief Yahoo, decked out in a giant foam sheriff's hat, is surrounded by roughly 250 photographers. But these are not paparazzi here to profile the billionaire. They are, in fact, far more interested in taking digital snapshots of one another.

This act of mass photography is a party for Flickr, the rapidly growing photo-sharing site acquired by Yahoo in March. Flickr co-founder Stewart Butterfield posted an open invitation on the Flickr blog, and all of these people--customers, really--just showed up. Butterfield and his co-founder and wife, Caterina Fake, have quite a fan club. "Flickr has changed so many lives. The number of friends I have has exploded," enthuses Deborah Lattimore, an entrepreneur who owns a transcription service.

Behind Lattimore is a huge screen with hundreds of constantly changing photos, a sample of the 14,000 or so images uploaded to Flickr every hour. Peer closely and you'll see snaps of the party that guests have already sent in via cell phone. "I look at Flickr with envy," Yang says. "It feels like where the Web is going."

What Yang envies is the community of 1.5 million rabidly loyal users Flickr has cultivated and the vast amount of content they've created. Of the 60 million photos uploaded to the site so far, more than 80 percent are public, meaning that anyone can look at them. More than half have been "tagged" with user-created labels, making them searchable. To use Flickr is to belong to the culture of participation sweeping the Web--where you write your own blog, produce your own podcast, and post your personal photos for all to see. If this is where the Web is going, Yang wants to make sure Yahoo gets there first.

Indeed, the Flickr purchase helped ignite a larger strategy. Thanks to a new generation of managers like Butterfield and Fake, Yahoo is starting to see how user-generated content, or "social media," is a key weapon in its war against Google. That upstart in neighboring Mountain View may have a better reputation for search, it may dominate online advertising, and it may always win when it comes to machines and math. But Yahoo has 191 million registered users. What would happen if it could form deep, lasting, Flickr-like bonds with them--and get them to apply tags not just to photos, but to the entire Web?

For one thing, the company could make a lot more money. More rabid users mean more ads and more premium subscriptions such as Flickr Pro, whose users pay $25 a year to take their photo-uploading allowance from 20 megabytes to 2 gigabytes a month. Such user fees account for 13 percent of Yahoo's revenues. Together, ads and fees added up to $3.8 billion in revenues and $1.2 billion in profits for the first nine months of this year.

It's a strategy that comes right from the top. Social media "is going to be a gigantic piece of what we do," says Yahoo CEO Terry Semel. "I don't think old media is what people are going to spend most of their time doing on the Internet. This paradigm needs its own inventions, its own methods, its own way to go forward." That doesn't mean Yahoo is ignoring traditional mass media--quite the opposite, in fact--but when it comes to the future of social media on the Web, Semel is winning praise for being one of Silicon Valley's savviest CEOs. "Yahoo has done the best job of the large guys of getting the concept," says tech guru Esther Dyson, who was an early investor in Flickr.

Semel may get it, but if you want to see this revolution in action, you have to talk to the young guns he's been hiring. Many of the champions of social media inside Yahoo--including Flickr's Butterfield and Fake, senior technologist Bradley Horowitz, and the head of Yahoo's developer network, Toni Schneider--are former startup founders recently acquired or hired. These entrepreneurs are sprinkling their social-media DNA all over the company, in a process some insiders are calling the "Flickrization" of Yahoo.

The Flickrizers' most ambitious goal is to turn Web searching itself into a social event--the idea being that you can find what you're looking for faster if you first see pages saved and tagged by people you know and trust. Done well, it could play as the triumph of the humans over Google's cold mechanical approach.

This is an especially attractive idea to Yahoo veterans, since it harks back to the vision Yang and Yahoo co-founder David Filo had in their Stanford University dorm rooms: Categorize the Web and recommend the best sites for its users, using human editors. That vision had to be abandoned when the Web got too large. But this time the users and the editors will be one and the same, there will be enough to tackle the entire Internet--and Yahoo won't have to pay them.

With hindsight, it's easy to trace a line through Yahoo's history, from Filo and Yang's human-filtered Web on one end to Flickr on the other, and believe that the company has always had some form of social media in its blood. But there were moments when Yahoo was decidedly unfriendly toward users' content. In 1999 it bought GeoCities, the popular purveyor of personal homepages, and proceeded to raze the 41 virtual neighborhoods into which its sites were clustered. Worse, Yahoo's new terms of service said the company owned rights to all GeoCities content, even members' photographs. Thousands of GeoCities residents left in protest. Yahoo quickly rescinded the terms of service.

Still, the company managed to get the user-generated formula right most of the time. It had a forum for special interests in Yahoo Groups, which now boasts 70 million members. And asking people to post their own reviews, as Amazon and Netflix know, is one of the best ways to grow your site's social value; Yahoo has more than 6 billion user-generated ratings on Yahoo Music alone.

Before it bought Flickr, Yahoo had just launched a fledgling social-networking service called Yahoo 360 that lets members write their own blogs. A few months earlier, the company had introduced the world of blogs to its members by adding the ability to subscribe to RSS feeds in My Yahoo. But the biggest change to Yahoo began in a New York hotel in November 2003.

That's where Butterfield was suffering from food poisoning when he had a feverish dream about a multiplayer game built around sharing digital photographs. The vision stuck with him. Three months later Butterfield and Fake, a former art director at, built the first version of Flickr in a Vancouver, British Columbia, loft.

At first the site centered on instant messaging, with a tray of photos that could be dragged into the conversation. Early members were gamers, bloggers, and digital-camera enthusiasts who set the tone by posting captivating pictures. They started to arrange local meetings to take more. Butterfield tweaked the technology, Fake nurtured the community, and membership went up to 250,000 people by the time Yahoo came knocking a year later.

The feature that first caught the eye of Horowitz, senior director of Yahoo's technology development group, was all that voluntary tagging Flickr users were doing. Horowitz is the former CTO of Virage, a video search company sold to Autonomy for $25 million. (He quit Autonomy and joined Yahoo in 2004.) Prior to Virage, Horowitz had spent years at the MIT Media Lab studying the science of computer vision and image recognition. "It is a very hard problem," he says. "Biological vision systems are beautiful, complicated things."

What struck him was that Flickr solves the problem in a very elegant way: Instead of teaching computers to identify images, Flickr gets people to do the heavy lifting. Most users describe their photos with tags and make them public for the benefit of friends and family, without realizing that they're greasing the wheels of a great social media machine. Add together all those labels and you have millions of keywords--a gold mine of image search. For a good time, try sampling the 94,000 photos Flickr users have tagged with the word "fun."

Encouraged to look at the startup by an e-mail from a Yahoo engineer and Flickr fanatic in Bangalore, India, Horowitz invited Butterfield and Fake to Silicon Valley in late 2004. They had lunch in the Yahoo cafeteria and immediately hit it off. "I met Stewart and Caterina and fell in love," Horowitz recalls. "It was beyond Flickr. I saw them as kindred spirits, entrepreneurs who could infect Yahoo with that small-company focus."

The feeling was mutual. Butterfield and Fake had several offers on the table for Flickr, reportedly including one from Google. Yahoo was their top pick. By March, Horowitz had persuaded his top brass to buy Flickr for an undisclosed sum, estimated to be around $30 million. But first, COO Dan Rosensweig had to overcome a lot of internal grumbling. There was no real business behind Flickr, and no unique technology either. So why did Yahoo need it? Says Rosensweig, "We could only justify it when we realized how big the vision could be if applied to the Yahoo network."

Once Yahoo committed to Flickr, those inside the company who believed in the promise of social media were even more encouraged to speak up, and Yahoo's executives were more likely to pay attention. Since the Flickr acquisition, Yahoo has added, an online events database created entirely by its users, and Konfabulator, which lets people design their own desktop software widgets and was bought with an eye toward eventually allowing users to slice and dice Yahoo content any way they want. The company has also launched beta sites for blog search and podcast search.

Such was the hunger for social media projects that Butterfield and Fake were in high demand from the moment they arrived in Sunnyvale. The pair were pulled into meetings with other groups--Yahoo Travel, Shopping, Local, 360--that wanted help becoming more Flickrized. Fake spoke to the folks at Yahoo Autos about the best ways to build a community around custom rides for auto enthusiasts. The travel group, which has more than 250,000 in-depth travel guides written by readers, wanted to tap the knowledge of its audience even further. "I spend so much of my time on stuff that is not Flickr that it stresses me out," Butterfield says.

If Yahoo is to get social media into every nook and cranny of its business, extending it to search could have the greatest benefit of all. Targeted search ads are projected to be a $9 billion industry this year, according to S.G. Cowen. To that end, Fake took over My Web in September. Still in beta, My Web is Yahoo's take on social bookmarking (a concept pioneered in 2003 by the startup It lets you save bookmarks, tag them, group them, and share them with your network of friends. The more you and your friends use it, the better the My Web experience will supposedly become.

If social search pans out, it could give Yahoo a much-needed edge over Google. Google takes an automated approach to search, throwing armies of Ph.D.s and thousands of servers at the problem. It wants to make search more relevant by creating better algorithms. Yahoo also does algorithmic search, but it can't beat Google at that game. So it's gambling that tapping into the collective intelligence of its audience will produce more relevant search results. "It is not who has the bigger index," Horowitz says, taking a swipe at his rival. "We hear a lot about efforts to index all the artifacts of human knowledge, but the actual bulk of human knowledge lives in people's heads."

That, at least, is the theory. In practice, however, tagging search results and bookmarks may still be too geeky an activity for Yahoo's average Joe. Thus far, My Web has seen tepid growth in the number of pages saved (about 300,000) and tags applied (fewer than 90,000). That might not seem bad for a product still in beta, but My Web is seeing little month-to-month growth. (, by contrast, has 10 million saved pages and half a million tags.)

Social search requires people to change their habits, and My Web works well only if you and a bunch of your friends use Yahoo. "They need to be much more open about letting people outside Yahoo see those tags," says Charlene Li, an analyst at Forrester Research. Fake is trying to get more pages saved by touting a "Save to My Web" button that bloggers and publishers can stick on their websites. She also plans to make it easier to connect people with similar interests. "We learned from Flickr that you need to make the experience very delightful," Fake says.

Perhaps the largest pothole in Yahoo's road to social media, however, is its business model. Users are encouraged to stay within the Yahoo network of pages as much as possible so they can be served more ads and sold more services. For now, it works. People spend more time on Yahoo--25 minutes a day--than on any major portal besides America Online, which benefits from a captive audience of dial-up users. But the tricky thing about the culture of participation is that users want to choose how they participate. For example, many bloggers are setting up feeds that put Flickr photos directly on their blogs, meaning that Yahoo doesn't get any pageviews--or ad revenues--from them. "I still think Yahoo has a heritage to overcome," says tech book publisher and pundit Tim O'Reilly, referring to the decade-old habit of directing traffic to its site.

Schneider, the vice president in charge of Yahoo's developer network, is trying to change that. He is another young gun, a former startup CEO whose e-mail company, Oddpost, was bought by Yahoo for $29 million (see "The New Road to Riches," October 2004) and has since become the basis for Yahoo's next-generation e-mail service. Now he's encouraging programmers to create applications based on Yahoo content that will not necessarily be hosted on Yahoo. He's doing this by opening up the company's APIs, or application programming interfaces, which is a bit like handing over the blueprints of your house to anyone who walks by and inviting them to construct an addition. "Why not make sure the next set of great startups is built on top of Yahoo?" Schneider says.

The trouble is that startups tend not to like APIs when there are strings attached. And attaching strings is exactly what Yahoo did at first, when it insisted that programs built on top of the Yahoo Maps code must be hosted on the Yahoo network. Since Google did not have the same requirement for the Google Maps API, many developers simply went out and built Google Maps-based applications instead. That, O'Reilly says, is a prime example of Yahoo's heritage problem.

Schneider acknowledges the issue and is working to fix it. In November he made Yahoo Maps as untethered as Google Maps. And he has high hopes for a new business model based on the Yahoo Publisher Network, launched in beta last summer. YPN, as it's known, serves up ads to small websites and blogs and pays them every time someone clicks on an ad, in much the same way Google's AdSense does. The difference is that YPN lets websites tag themselves, Flickr-style, so as to get more relevant ads in return.

In the long run, thanks to YPN, Yahoo content will become more nomadic. Schneider wants YPN to syndicate reviews, ratings, and Flickr feeds. He is even considering opening up an API to YPN itself so users can experiment with building their own creative advertising systems. In theory, someone could figure out how to deliver audio and video ads via YPN.

There are many more visions of cool user content on the horizon, like the idea of a Flickr for videos. Clearly, though, there's a limit to how fast Yahoo can absorb all this radical change. Butterfield compares the company to a centipede: "You can take a few feet off the ground and change a few things, but you cannot take all the feet off the ground at the same time." Still, with the support of Semel and the passion of true believers like Schneider, the Flickrization of Yahoo is proceeding at a startling pace. What the company looks like in the future will depend entirely on how many users show up for the party.