The Platform

Mark Zuckerberg has had a particular obsession since Facebook's early
days. On the night that his early collaborator Sean Parker first met
Zuckerberg at that trendy Tribeca Chinese restaurant in May 2004, the
two got into a curious argument. Zuckerberg, in Parker's opinion, kept
derailing the discussion by repeatedly talking about how he wanted to
turn Thefacebook into a platform. What he meant was that he wanted
his nascent service to be a place where others could deploy software,
much as Microsoft's Windows or the Apple Macintosh were platforms
for applications created by others. Parker argued that it was way too
early to think about anything like that.
Kevin Efrusy of Accel Partners has a similar recollection. At one
of his very first meetings with Zuckerberg after Accel invested in
the company in late spring 2005, the young CEO asked for a favor.
"Kevin, I need to find someone to help me think through my platform
strategy."
"Huh? Yeah, well maybe someday we can be a platform," Efrusy
replied, haltingly. "But we're just a company with six people . . . I mean,
I guess I know a guy over at BEA [a business software company] who has
done some interesting platform work . . ."
Zuckerberg cut him off. "BEA? I was thinking more like Bill Gates.
Can you help me talk to Bill Gates?"
"Ummm . . . I don't know. Maybe Jim Breyer can help with that..."
A week passed. Efrusy was again at Zuckerberg's office. "Okay," said
Zuckerberg. "So I talked to him."
"Talked to who?"
"Bill Gates!"
the faeebook effect
Even in these early days, Zuckerberg was trying to imagine how
his little service could be more than just an Internet destination where
people went to communicate with each other.
Every great technology company goes through one or two key transitional
moments when its creators discover they have created something
different—and bigger—than they initially realized. Early on it dawned
on Bill Gates—then making bespoke software for little PC hardware
companies with partner Paul Allen—that software should be its own
industry. He later had a second epochal realization: that entire computers
could be built around an operating system. Microsoft subsequently
became the most profitable company in history. One night it hit Yahoo
founders Jerry Yang and Jeff Filo that they didn't have only a map to
the Internet. Their service could also be an unprecedented way to collect
detailed market research about Net users. Yahoo became the first
big ad-supported Net media company. Google's turnabout came when
founders Sergei Brin and Larry Page discovered they could direct user
searches not only toward websites but also toward a separate database
of advertising. Thus was born the most powerful business model of the
Internet era so far.
Zuckerberg's first eureka moment was when he and Moskovitz realized
their service could go beyond college. But another struck while
watching the stunning success of the photos application. It became
apparent something special was happening. "Our photo site lacks
features anyone else would build," Zuckerberg told me in early May
2007. "We don't store high-resolution photos. The printing function is
downright bad. And until recently you couldn't even change the order
of photos in an album. Yet somehow, this application became the most
trafficked photo site on the Internet, by far." And something similar was
going on with the application Faeebook engineers had quickly thrown
together to allow users to invite friends to events. It was garnering more
usage than Evite.com, which had been for years the leading website
for invitations.
"So why were photos and events so good?" he asked. "It was because
despite all their shortcomings they had one thing no one else had. And
that was integration with the social graph." This was Facebook's own
conceptual breakthrough, and Zuckerberg was proud of the term he
used to describe it. "We did some thinking and we decided that the core
value of Facebook is in the set of friend connections," he continued.
"We call that the social graph, in the mathematical sense of a series of
nodes and connections. The nodes are the individuals and the connections
are the friendships." Then his enthusiasm veered, it seemed at the
time, toward overstatement: "We have the most powerful distribution
mechanism that's been created in a generation." Zuckerberg immodestly
explained that this same power could be applied to any sort of application—
not just photos or events. His certitude was jarring.
By "distribution" he meant that by connecting with your friends on
Facebook you had assembled a network, this so-called social graph, and
it could be employed to distribute any sort of information. If you added
a photo, it told your friends. Ditto if you changed your relationship
status, or announced that you were heading to Mexico for the weekend.
But it could also tell your friends about any action you took using any
software to which your social graph was connected. So far, though, the
only applications that took advantage of this distribution capability were
photos, events, and a few others created by Facebook itself.
Most software companies, were they to conclude that they had such
an ability to create uniquely powerful applications, would create more
of them. They might make shopping applications on top of their social
graph, or games, or applications for businesses. Instead, Facebook
stopped building applications at all, at least for a while. In the fall of
2006 Zuckerberg set out to realize his long-held vision of a platform
for others to build applications on top of Facebook. He wanted to do
for the Web what Gates did for the personal computer: create a standard
software infrastructure that made it easier to build applications—
this time, applications that had a social component. "We want to make
Facebook into something of an operating system, so you can run full
applications," he explained.
COO Owen Van Natta, whom I also talked to in May 2007, had
his own way of describing this potential: "Take anything today on the
Internet and overlay a lens that is people you know and trust who have
their own perspective. That's what we will enable with platform. What
wouldn't potentially be more valuable when seen through that lens?"
Zuckerberg had thought about platforms almost since he first touched
a keyboard. He learned to program as an adolescent by coding functions
that worked on top of AOL, then the dominant online service.
A community of hackers—including Zuckerberg—turned AOL into a
platform whether its leaders wanted it to be one or not. Then when he
was a senior at Exeter, he teamed up with Adam D'Angelo and built
his software for listening to MP3s (audio files) called Synapse. Synapse
became popular in part because it allowed other programmers to build
companion programs, called plug-ins, that supplied additional features.
Synapse was, in effect, a mini-platform. And in his earlier, abandoned
obsession with his cherished Wirehog, Zuckerberg was thinking of
Facebook as a platform. Wirehog was in effect, if only briefly, the first
independent application to operate on top of Facebook.
Becoming a platform on which the applications of others can operate
is one of the great holy grails of technology. Microsoft dominated
the technology industry for almost two decades by positioning its Windows
software as the monopoly operating system platform for the PC
industry. Anyone who wanted to build a PC application had to use Windows.
(It was Bill Gates in fact who popularized this use of the word
"platform.")
Creating a platform enables a software company to become the
nexus of an ecosystem of partners that are dependent on its product.
And once a company is at the center of an entire ecosystem, it becomes
maddeningly difficult for competitors to dislodge it. Not only did Apple
succeed at this masterfully with its Macintosh operating system, but
it succeeded again, first with the iPod and then with its magnificent
iPhone.
By becoming a platform, Facebook also takes some of the burden
off itself to excel in everything it does. Facebook will never be able to
build the best application in every area its users are interested in. ComThe
Platform
panics that devote more resources to chat, for example, will continue
to outpace Facebook. I recently asked my seventeen-year-old daughter,
Clara, if she used Facebook's chat application, an ambitious addon
the service launched in mid-2008. No, she said, she still preferred
AIM and Apple's iChat (an answer many American teenagers would
give, despite their addiction to Facebook). "Facebook Chat feels like
using Morse code," she explained. It doesn't have enough features and
isn't easy enough to use. Zuckerberg decided that what Facebook did
uniquely well was maintain your personal profile and your network of
friend connections. Ultimately, almost everything else will be done by
other companies.
Facebook made the first move to turn itself into a platform back in
August 2006. The world barely noticed. The big news around then was
the News Feed scandal. Programmer Dave Fetterman spearheaded
something called the Facebook API, or application programming interface.
It enabled users to log in to other sites on the Web with their
Facebook username and password so the partner site could extract
their data, including their list of friends. Some at Facebook—mostly
older executives —objected to letting user data escape the confines of
the service in this way. They said the company was giving away something
valuable and getting nothing in return. But Zuckerberg pushed
it through. To demonstrate the API, Facebook built its own external
website application called Facebank, later renamed Moochspot, for
keeping track of small debts between friends.
While thousands of developers did fiddle around with the API, not
many used it, and very few Facebook users did.
The real problem with the API was that it didn't help outside application
partners very much, because it didn't include that vaunted "distribution."
It didn't take full advantage of the social graph. You could
pull your list of friends out of Facebook but you couldn't send information
you produced back inside to them. You and your friends could
keep track of debts on Moochspot, yet it didn't send any information
back to your profile.
the faeebook effect
But shortly Facebook triumphed with News Feed. It enabled your
friends to easily learn about your Facebook activities —including which
applications you installed on your profile. Only with the News Feed in
place could Facebook become a successful platform. Open registration
also helped lay the groundwork. Software developers would obviously
be more interested in applications on Facebook if it operated at a large
scale and included all sorts of people.
As soon as the News Feed brouhaha settled down, the company's priorities
turned to building the platform. D'Angelo and Charlie Cheever
did much of the critical programming work. Dave Morin got the job of
"platform marketing'7—working with potential developers. (In his previous
job at Apple, already a Facebook partisan, he had sought futilely
to get Facebook built into the Mac OS.) Morin and Fetterman visited
companies that had successfully created platforms, including eBay,
Apple, and Salesforce.com.
But despite all the external models, the team kept harking back
to one internal reference point. "We used photos as the model the entire
time," says Morin. "We just kept looking at it, asking, 'How do we
enable every application to do what photos does?'" Each profile page
included a box for photo albums. Clicking on a photo took a user to an
entire page, which looked much like a website. When you uploaded a
photo it updated your personal mini-feed on your profile as well as the
News Feeds of relevant friends. So the team decided to allow outside
developers similarly to place boxes on profile pages and to build full
pages inside Facebook. Actions in any application could, of course, generate
News Feed stories.
Carrying this logic even further, they arrived at the principle that
Facebook should not be able to do anything with its own applications
that outside developers couldn't do. It should be a level playing field,
Zuckerberg explained in 2007. "We want an ecosystem which doesn't
favor our own applications," he said. This policy was followed to such
an extreme that features were removed from Facebook's own photos
application because an outside developer would not have been able to
include them.
The company extended an extraordinary degree of freedom to its
new partners. Amazingly, it planned to let developers make money
with their applications, but would not charge them anything at all for
the right to operate inside Facebook. "People can develop on this for
free/' said Zuckerberg around the time the platform debuted, "and can
do whatever they want. They can build a business inside of Facebook.
They can run ads. They can have sponsorships. They can sell things,
they can link off to another site. We are just agnostic. There are going
to be companies whose only product is an application that lives within
Facebook."
But did it make Facebook a better business? That was not a priority.
"We don't force ourselves to answer the question how we're going to
make money off this right now so long as it's strengthening our market
position," he said back then. "We'll figure that out later."
So Zuckerberg saw it. But some of his colleagues, particularly the
ones who sold advertising for the site, were apoplectic. Why should its
application partners be allowed to compete with Facebook itself in selling
ads? There were plenty of angry meetings. But for all the venting,
Zuckerberg was unswayed. Activity on applications, he argued, would
generate more activity in Facebook. That would create more page views,
and even on application pages Facebook would reserve space to sell its
own ads. Zuckerberg also advocated a sort of corporate Darwinism. He
said he wanted outside apps to help keep Facebook honest by forcing it
to make its own remaining applications good enough to compete successfully.
Back then I talked to Zuckerberg in his private retreat—an allwhite
conference room furnished with midcentury modern furniture
from Design Within Reach, a few blocks down University Avenue. (He
didn't decorate it himself, but he liked it.) White Eames chairs, a white
Saarinen table with delicate metal legs, white curtains, white blinds,
gray rug and sofa, and a big black beanbag chair. Employees called
it the "interrogation room" both because Zuckerberg was known for
his probing questions and because its austerity evoked a prison cell. It
was Zuckerberg's twenty-third birthday when he and I met there. He
was barefoot and unshaven, wearing an A&W Root Beer T-shirt with
blue jeans. In a corner was an unopened box of Transformer robot toys.
Zuckerberg was drawing diagrams on the whiteboards that covered all
the walls, and at one point couldn't find an eraser. So he picked up a
knit hat from the floor and wiped the board with that.
In April, Zuckerberg had given a talk at a News Corp. executive
summit at a resort at Pebble Beach, two hours south of Palo Alto. Rupert
Murdoch had recently said a few things in public to suggest he wondered
if he had bought the wrong social network. At a gala dinner Zuckerberg
and Murdoch huddled together intently, while MySpace CEO
Chris DeWolfe sat nervously at a nearby table. Finally Zuckerberg got
up, announcing that he had to get back to take his girlfriend to a movie.
"After he left, the MySpace guys rushed over to Rupert," says blogger
and author Jeff Jarvis, who attended the dinner. "It was like 'Dad! Pay
attention to me!'"
Now as the platform launch approached, Zuckerberg made no
bones about the fact that it was intended partly to best MySpace, which
remained the dominant American social network. MySpace had recently
decreed that some third-party applications could not operate
there, and even shut one down merely on suspicion that it had been
selling advertising. "We just have such a different philosophy and view
of the world," Zuckerberg explained. "We're a technology company.
MySpace is a media company, and they view their job as owning and
distributing content."
To succeed with the platform launch, Facebook had to start promoting
itself to developers. Dave Morin and Matt Cohler crisscrossed the
globe visiting start-ups and big media companies alike, seeking to convince
them to make software for Facebook. A splashy launch event was
planned for May 24, 2007, at a big hall in San Francisco. Facebook
called the event f8, a name that subtly proclaimed it was Facebook's
"fate" to become a platform. Zuckerberg even emerged from his shell to
solicit advance attention from a journalist, me, whom he invited inside
the company for an exclusive story as he prepared for f8. I published
an article titled "Facebook's Plan to Hook Up the World" in Fortune
magazine and online at the very moment f8 began.
Facebook hired a veteran event planner named Michael Christman
to oversee the f8 logistics. On his first visit to the offices he was in a
lengthy meeting, sitting by the door in a big conference room that also
held a flat-screen TV and a Nintendo Wii machine. The door opened
and banged into Christman's back. Two young men appeared but
backed out when they realized the room was in use. A few minutes later
they came in again, hoping the meeting was over, and again knocked
his chair. They wanted to play the video game. When it happened a
third time, Christman turned and said sternly, "Boys, if you want to play
with the Wii, come in. But don't bang my chair again/' At that point,
Meagan Marks, a Facebook employee who was managing f8, said, "Michael,
this might be a good time to introduce you to our CEO, Mark
Zuckerberg."
The days leading up to f8 were a frenzy of excitement and near
panic. Employees were fueled by a sense that they were making history.
The graffiti-scrawled halls were abuzz with grand proclamations.
"We're gonna change the Internet!" "We're gonna make the Internet social!"
"We're gonna finally put people on the Internet!" "We're creating
a real economy on the Web!" Apple veteran Dave Morin remembers
driving home one night at 4 A.M. after a particularly intense planning
session, thinking, "This is what it must have been like building the first
Macintosh." To prepare for the new Facebook, Morin was reading Democracy
in America by Alexis de Tocqueville, the classic nineteenthcentury
observation of the U.S. political and economic system, as well
as Adam Smith's The Wealth of Nations. Modesty of ambition has never
characterized successful leaders at Facebook.
It had been a marathon of programming. Adam D'Angelo and his
team building the platform worked seven days a week for more than
three months. The night before f8 they were almost—but not quite —
ready. A core group crowded into a room at San Francisco's W Hotel
running through final fixes. Most hadn't slept for days. But a key piece
of the platform software still didn't work properly.
Some of the programmers took an alertness drug called Provisual
so they could stay up yet another night. They were semidelirious. They
joked they should mix Provisual with cocaine and call it Blow-visual.
Luckily the quality of their coding was higher than that of their humor.
But they made it through the night Just hours before f8 was scheduled
to begin, they flipped the switch. The software worked! Their brains
barely did.
Nobody outside Facebook knew what was coming, except the few
partners that had agreed to develop applications in advance. The company
had kept the purpose of f8 secret. The only thing most of Silicon
Valley knew was that Facebook would make a big announcement. Facebook
had never done anything like this before. Hundreds of journalists
crowded the front rows. It seemed every software and Internet company
in California, and many from farther away, sent a delegation.
As f8 began, the 750 people throughout the packed room strained
to see the diminutive Zuckerberg, wearing his standard T-shirt, fleece
jacket, and sandals. He walked out onstage and pronounced, "Together,
we're starting a movement!" It was a phrase suggested by hip San Francisco
strategy and marketing consulting firm Stone Yamashita.
Zuckerberg's platform demo was by far the best-rehearsed presentation
he had ever given. He'd slaved over his wording, but continued
modifying his slides until minutes before he was scheduled to appear.
He was extremely nervous. Everybody would be watching, even his parents,
who were in the audience. But he paid a price for his last-minute
modifications. When he got onstage the slides appeared in the wrong
order and his speech got out of synch. He paused and looked confused.
The event staff and Facebook's executives held their breath. "Well, this
worked in my office . . ." he joked. The tension defused. The correct
slide came up. He finished smoothly.
The platform wowed the crowd. It took Facebook way past MySpace.
No other consumer website had anything like this. Rapturous coverage
instantly began sprouting on blogs and journals all over.
A solid ecosystem had already started coming together. More
than forty companies demonstrated applications. Mighty Microsoft
showed two apps that helped integrate existing Internet software
with Facebook. The Washington Post (who else?) showed a "political
compass" to compare your political views to those of your friends.
Sean Parker teamed up with Zuckerberg's old Harvard dormmate Joe
Green to make an application called Causes, to help nonprofits raise
money. Another big partner at the platform launch was iLike, which
had previously built its own social network to share songs and musical
favorites.
Immediately afterward, f8 turned into an eight-hour public hackathon,
where any developer could work alongside Zuckerberg and
Facebook's programmers to build software on the fly. (That was another
reason it was called f8.) But when the event ended at twelve, the night
was not over for the Facebook crew.
They retired again to the W Hotel, where they proceeded to, as
they said, "push the platform live," meaning turn it on. Staffers scattered
throughout conference rooms to do various necessary tasks, while
Moskovitz and Morin sat on a sofa in the lobby working from their laptops
via the hotel Wi-Fi. Once the platform was working they crashed,
though not before a little partying, of course.
Dave Morin awoke blearily the following morning to find a string of
panicked messages on his cell phone. "We have so much traffic we don't
know what to do!" said one from an executive at iLike. "Can you help
us get more servers?" Apparently just about every application launched
the day before was having trouble under the strain of a massive influx
of users. Morin headed developer relations, so the companies wanted
his help. iLike's executives flew down from Seattle and rented a U-Haul
truck, which they drove around Silicon Valley borrowing servers from
various tech companies so they could handle the load. By Friday, the
day after f8, 40,000 Facebook users had installed the iLike application.
Two days later, the figure had soared to 400,000.
Morin got help from the company that ran Facebook's South San
Francisco data center. Facebook itself occupied a series of what are
called "cages"—fenced-in indoor enclosures full of servers and networking
equipment. An adjacent cage was made available to any developer
that needed help managing its traffic. Eventually Facebook did
a deal with a larger data center operator to open an entire facility for
application partners, which would be, in Internet lingo, "peered" with
Facebook's, meaning that in the electronic topography of the Net it was
essentially right next door.
The reaction to f8 across the tech industry was close to ecstatic.
Facebook's platform launch became—along with the launch of Apple's
iPhone a month later—one of the two most-discussed tech events of
the year. No longer was it possible to dismiss this upstart as merely a
plaything for college kids. The influential blog TechCrunch called the
platform "inspired thinking." Prior to f8, Zuckerberg and his crew had
hoped that in the subsequent year 5,000 applications might come onto
Facebook and half its users would install them. But within six months
250,000 developers were registered, operating 25,000 applications.
Just as Zuckerberg had predicted, Facebook gave applications an
unusual ability to acquire new users. This was the vaunted "distribution."
The News Feed told users when their friends had installed new
applications, so even the most modest app from a single developer with
no marketing budget could reach millions of users almost overnight if
it did something useful. Though the News Feed still was a selection
chosen by algorithm, Facebook tuned the software to make sure that
newly installed applications were announced. By six months later, half
of Facebook's users had at least one application on their profile.
Just about every software and Internet company was suddenly talking
about building an application for Facebook—from industry titans
to college kids in their dorm room. Facebook's platform infrastructure
made it almost as easy for such lone wolves to create an application
as for Microsoft. When it launched the platform, Facebook turned off
its own Courses application, which helped college students track one
another's class schedules. A New Jersey high school student named
Jake Jarvis, seeing opportunity, quickly wrote something similar and six
months later sold it for an amount his father says was "sufficient to pay
for a year in college."
The platform brought Facebook a gravitas it never before possessed.
It caused both technologists and ordinary users to sense that this service
was more than they'd reckoned. In Silicon Valley and among techies
worldwide, it suddenly became uncool not to have your own Facebook
profile.
The platform also changed the experience of being on Facebook.
There was a new expansiveness, an air of possibility. If adding the photos
application had made Facebook feel like a place where you wanted
to spend a lot of your time, turning it into a platform for applications
began to make it feel a bit like being on the Web itself. Facebook was
becoming its own self-contained universe.
For high school and college students it had long been routine to
spend the majority of their online time there. Now people of all sorts
and of all ages began to do the same. On the day of f8 — May 24, 2007 —
Facebook had 24 million active users, with 150,000 new ones joining
every day. The demographics were already spreading out, with 5 million
users between twenty-five and thirty-four, a million between thirty-five
and forty-four, and 200,000 over age sixty-five. Within a year Facebook
tripled to more than 70 million active users.
In all the complex and frenzied preparations for f8, Zuckerberg and
his team had given surprisingly little thought to exactly what kinds of
applications were likely to work best on Facebook. As is so often the
case at this company, driven by ideals and led by a CEO obsessed with
a long-term view, high-mindedness prevailed. The Facebook team assumed
that general-purpose applications with wide functional appeal
would play a big role in the new ecosystem. When they prepared for f8
by taking proprietary features out of their own photos app, for example,
they believed that someone might come along with a better one and
successfully compete against them. Their idea was that this should be
a forum for the best, most functional, most sophisticated applications.
When I was reporting in 2007 prior to f8, Facebook had me speak with
one close ally of the company, who told me, "Facebook is creating the
opportunity to build a whole generation of Adobes and Electronic Arts
and Intuits that live within Facebook." These were the giants of the
industry. As usual, the company was aiming high.
Facebook, however, is nothing more than the collective actions of
its users. What happens there depends on what Facebook users are interested
in, not, in the end, what Mark Zuckerberg thinks they ought
the fa.cebook effect
to be interested in. With Facebook's platform, he learned that lesson a
bit painfully.
A frenzy of new applications quickly emerged on Facebook, but
they were hardly high-minded. The ones that took off fastest were
mostly silly, but intrinsically social in a way that games on the Web had
never been before. One of the first really hot apps was one called Fluff
Friends. It didn't do much more than let you electronically "pet" a virtual
dog or cat, but when you pet your friend's dog your photo would
show up on their profile. It was a new way to send a simple message,
which 5 million people did. Another similar app enabled you to give
your friends a "vampire bite." Food Fight helped you throw food at your
friends, and reached 2 million users in just a few weeks. A silly little
app called Graffiti—which let you scribble on friends' pages—became
the number-two application. A couple of young guys in San Francisco
wrote it in a couple days in their apartment.
These were genuinely social applications—they successfully
brought offline behavior into this new online world. It's just that it was
the kind of behavior that reflected the penchants of the people who still
were the overwhelming majority of Facebook's users—teenagers and
college kids.
A few weeks before f8, Morin had coffee with Mark Pincus, the erstwhile
founder ofTribe.net, co-owner of the sixdegrees social networking
patent, and early investor in Facebook. Pincus told Morin excitedly that
he intended to build a poker application for the new platform. "It won't
work," Morin asserted dourly. "Games aren't viral." Pincus went ahead
and launched Texas HoldEm Poker on Facebook, starting a company
called Zynga, which was headed for huge success. Zuckerberg himself
was disappointed at the silliness of many of these apps. He wanted his
company to help people communicate things that mattered, not make
it easier to play around.
Then came the phenomenon called Scrabulous. Two brothers
from Kolkata, India, Rajat and Jayant Agarwalla, built a blatant imitation
of the classic board game Scrabble for Facebook. You could
play multiple games with as many friends as you wanted, taking a turn
whenever it was convenient. Scrabulous was a sensation. The magaThe
Platform
zine PC World rated Scrabulous number fifteen on its list of "The 100
Best Products of 2008," just behind Craigslist and ahead of the Nintendo
Wii. (Facebook itself was number three.) Some days as many as
342,000 people played.
Scrabulous got even Mark Zuckerberg's attention. He had never
succeeded in convincing his grandparents to join Facebook, but now
they finally agreed—so they could play Scrabulous with him there. His
antipathy toward games on Facebook began to crack. It was apparent to
him that you were interacting with people you cared about when you
played Scrabulous. And after all, Scrabble was a game of words and the
intellect—the kind of game played by people who went to Harvard.
All this excitement did not go over well with the owners of Scrabble,
however. Shortly after Scrabulous launched, Hasbro, which owns the
rights in the United States and Canada, tried to buy the online game,
reportedly for as much as $10 million. The Agarwalla brothers refused
to sell, then Hasbro sued them. The game was shut down. Meanwhile,
Mattel, which sells Scrabble in the rest of the world, launched its own
Facebook version for use outside North America. Eventually Hasbro
launched a legal U.S. Scrabble for Facebook, and the Agarwallas reworked
their game to resemble Scrabble less and renamed it Lexulous.
It remains popular.
Pincus's Texas HoldEm was the next game to take off. Morin was
right in a sense—games didn't spread virally as quickly as some kinds
of applications. But they engendered extraordinary loyalty—once a
user starts playing, they frequently come back. Zynga later added other
games, including Farmville and Mafia Wars, both of which now have
millions of users. Pincus raised money from venture capitalists and
invested aggressively. Zynga is now the largest application company
on Facebook, with about 250 employees and over $200 million in annual
revenues. And Pincus says Zynga is profitable. Texas HoldEm had
20.3 million active users on Facebook in December 2009, making it
by far the most popular poker site anywhere on the Internet. But even
more impressive is the game called Farmville, also created by Zynga.
On Farmville, a player manages and grows a farm, tending crops and
feeding animals, etc. You trade with your neighbors and join in a comthe
facebook effect
munity of farmers all trying to build the biggest and most productive
farm. It has about 80 million total users. Overall Zynga has 241 million
total active users for all its games as of February 2010, according to the
research firm Inside Network.
Games are now the most successful type of application on Facebook,
drawing in phenomenal numbers of players. It makes sense, since
gaming is a fundamentally social activity. Facebook enables you to play
any game with any of your friends on the service. As of February 2010
there were twelve games on Facebook with more than 20 million players,
according to the company. The highly complex World of Warcraft
for years dominated online multiplayer gaming, with at most 11.5 million
players. But gaming on Facebook is a more casual activity. "We
now have tens of millions of people playing games who don't identify
themselves as gamers," says Gareth Davis, who oversees the gaming portion
of the Facebook platform. "They play games here because they
want to have fun with their friends."
Davis is working with every major console game manufacturer to
enable classic video games to connect with Facebook and incorporate
a social element. "In three years every game will be social," he predicts.
"Every single device—whether it's a console or a phone or a TV—will
connect with Facebook and be able to incorporate and share your Facebook
data." One game by a company called Social Gaming Network
enables people to play tennis. For a racket they swing an iPhone that is
connected to Facebook. Their adversary on Facebook can be anywhere
else in the world, swinging his or her own iPhone.
Games and silly applications continued to surge throughout the Facebook
platform's first year, but the company was finding it wasn't simple
to manage and police its ecosystem of partners. Since anyone could
create an app, the platform attracted quite a few players who were less
idealistic than Zuckerberg and more interested in making a quick buck.
A competition ensued among applications for users at all costs. Applications
were designed as much to get new users as to be fun or valuable.
The key was to figure out how to manipulate Facebook's software
so that messages would go into people's News Feed inviting them to
download an app. Applications became very clever about generating
stories that would flood everyone's home pages. One called Funwall let
you create little animations or download video onto your profile. That
was fine and good. But it had an insidious interface that used ambiguous
language to trick many users into sending invitations to every single
friend. Even tech industry sophisticates fell for it.
Facebook kept trying to weed out spammers and encourage more
reputable applications. But changes intended to punish malfeasance
often impaired legitimate apps. Says Morin: "We had to learn a lot about
developer relations and policy setting and things that we just didn't understand.
We sort of stumbled our way through becoming good at dealing
with developers."
The company implemented a variety of new rules to try to police
applications and make them behave. It urged users to complain about
spam. It changed the software to reduce the number of application stories
flowing into a user's News Feed. And it hired an industry veteran
to head up the platform. Ben Ling, a slender and flamboyant Chinese-
American, had been running the payment system called Google Checkout.
He was the highest-level employee Facebook had ever lured away
from Google. Executives called him a "rock star."
By the summer of 2008 the problems had gotten completely out
of hand. Facebook's platform was like the Wild West. So at a second
f8 that July, Facebook announced a variety of refinements and rule
changes, including a rating system. Now Facebook could weed out apps
by "verifying" the good ones. Facebook wanted to encourage the apps
that were the most fun or useful. Despite all the fluff, a fair number
of substantial and useful applications did get traction. A popular one
called Visual Bookshelf let you list books you've read, rate them, and
write short reviews.
But Zuckerberg's favorite Facebook application was the Parkerand-
Green-created Causes. It was driven by high motives—to help
nonprofits raise money. Facebook users who make a donation create a
story in their friends' News Feeds. Ideally that inspires friends to make
their own donations. Explains Joe Green: "Social recognition matters
in charity, too. People who make big donations like to get their names
on hospital buildings. So we at Causes allow you to show what you care
about on your Facebook profile." He says it's like wearing a yellow rubber
Livestrong bracelet. Users responded strongly. Causes has remained
among the largest applications on Facebook.
Now the platform ecosystem has become substantial. There are
more than 500,000 applications operating on Facebook, created by over
1 million registered developers from 180 countries. More than 250 of
these applications have at least one million active users every month.
Investors have high hopes for this new type of software company. The
top five Facebook applications companies alone—Zynga, Playfish,
Rock You!, CrowdStar, and Causes—have raised approximately $359
million in investment capital between them. That includes a giant
$180 million infusion into Zynga in late 2009 by private investors led
by the Russian firm Digital Sky Technologies. Justin Smith, who runs
Inside Facebook, which is devoted to the Facebook developer community,
estimates that there are about fifty venture-funded software companies
with substantial revenues whose primary business is building and
operating applications on Facebook. Zynga is the largest. About 200
smaller companies, comprised of two to four developers each, have annual
revenues of several hundred thousand dollars. At least another 300
solo operators have written a Facebook application that earns enough
to support them.
Facebook application companies are doing so well that their estimated
aggregate revenue in 2009 was roughly the same amount as
Facebook's itself—slightly over $500 million. These applications generate
revenue in several ways. Selling advertising generates $200 million
for applications companies. Apps often host ads promoting other Facebook
apps, and get paid about fifty cents on average each time a user
clicks through and installs another app.
Transactions inside applications create even more revenue. Justin
Smith of Inside Facebook estimates there were $300 million in such
transactions in 2009. Much of this is spent to buy an upgrade to a more
advanced level of a game, or to buy virtual goods, like a fancier shoe
to kick your friend in KickMania. Playfish's Pet Society game, where
users set up houses to display their pets, releases new virtual items every
Monday. On Valentine's Day in 2009 the company sold five million
images of roses that players could give to their friends. Each one cost
about two dollars. In Zynga's Texas HoldEm, players who want more
chips than they're allocated each day pay real money to get them, even
though there is no way to remove winnings from Facebook. Numerous
Facebook games have revenue exceeding $3 million per month.
Savvy marketers have also realized that Facebook applications are
a free way to get in front of consumers. That's why the Washington
Post Company did its Political Compass. When Bob Dylan released a
new album in 2008, his record label created an application that used
old footage of him as a young man holding a series of signs. Facebook
users could put their own message on the signs and then host the film
on their profile.
In turning its network into a platform for whatever any outside developer
wants to build, Facebook has created many new capabilities but
also a new set of risks. For all their usefulness and entertainment value,
applications on Facebook are often cavalier about how they treat user
data. Frequently when users install an app they give it essentially blanket
permission to extract data from their profile. But once that data is in
the hands of the developer the user loses all control of what happens to
it. Facebook has just begun to take steps to deal with this problem. The
limits for what is and isn't acceptable remain unclear, and predatory
applications continue to arise that take unnecessary liberties, often in
order to make personal data available to outside marketers who pay for
access to it. It's another piece in the complicated puzzle of Facebook
privacy. Says Marc Rotenberg of the Electronic Privacy Information
Center: "Facebook and its business partners learn lots about us, but we
know very little about them or about what information of ours is collected
and how it's used."
As more and more software companies embrace the platform and as
Facebook's dominance of social network computing spreads around the
globe, the company's platform strategy is rapidly evolving. Its long-term
plan is that fewer and fewer applications will operate inside Facebook's
own walls. Now a service called Facebook Connect enables any website
to tap into users' information and network of friends, and send reports
of user activity back into News Feeds on Facebook. The company increasingly
is encouraging partners to tap into Facebook that way. So
far more than 80,000 websites already do, including about half of the
largest ones worldwide. Zuckerberg's long-desired platform strategy has
been paying off.