Fall 2005

As the school year resumed in the fall of 2005, the company now named
Facebook had effectively blanketed the college market—85 percent of
American college students were users and a full 60 percent returned
to it daily. Now Zuckerberg wanted to broaden membership into new
demographics. But many in the company wondered whether it made
sense. "The debate was 'What's next?'" says board member Jim Breyer.
"Do we go international? Do we go young adult and keep the people
who are graduating? But we knew that if we were going to win big, we
had to start getting the hearts and minds of high schoolers."
Zuckerberg and his. co-founder Moskovitz, for their part, saw Facebook
on a slow march toward ubiquity. To them, high school was just
an obvious next step. This could be a huge leap in Facebook's audience.
And it was important to counter MySpace, which was making rapid inroads
in high schools. Once you knew how Zuckerberg felt, you knew
how the board was going to vote.
So Facebook had that summer started planning to include high
school students. Investor Breyer and Matt Cohler—the older peopleboth
argued that the Facebook brand was irrevocably associated with
college and that college students didn't want high schoolers in there
with them. They argued that a high school Facebook should operate
separately and under a different name. Facebook High was considered
promising, but "FacebookHigh.com" was owned by a speculator who
wanted too much money for it.
If high school students joined Facebook, how would the service
validate users? Protecting the culture of real names and genuine identity
was critical. The college-issued .edu email addresses had ensured
that people were who they said they were. That was the foundation that
enabled Facebook to protect its users7 information—you only shared
stuff with people you knew. More than half of all users had so much
faith in the security of their information that they included their cellphone
number in their profile.
However, only a small number of high schools, mostly private
ones, gave students email addresses. New general counsel Chris Kelly,
who had recently been hired, briefly launched a campaign to convince
high schools to issue email addresses to students as an online safety
measure. Then Facebook considered instituting its own national high
school email service. Finally it came up with a compromise. Part of
what authenticated you on Facebook was the people who, in effect,
vouched for you by being your online friends. So college freshmen and
sophomores were encouraged to invite their friends who were still in
high school. Then those users could invite their own friends. It meant
a slower start for the high school version of Facebook. The service created
separate "networks," or membership groups, for every one of the
country's 37,000 public and private secondary schools.
Initially, the high school site operated as a separate "Facebook."
Though high school users also logged in at Facebook.com, they
couldn't see college users7 profiles. Membership grew painfully slowly
at first, but by late October thousands of high school students were joining
the service each day. (Overall at that point, about 20,000 new users
were joining daily.)
Facebook was no longer just a college phenomenon. Zuckerberg,
with the strong support of Moskovitz, soon insisted that the two services
should be merged. By February 2006 they were ready to abandon that
distinction, so users could freely establish friendships or send messages
with anyone regardless of age or grade (the minimum age was set at thirteen).
Cohler and Breyer and many of the older employees remained
extremely worried that Facebook's appeal to college kids would plummet
when they saw high schoolers in there with them.
So it was a very dramatic day for them when they merged the
two systems. But it turned out college kids—the ones who noticed —
were generally pleased to be able to communicate with a larger universe
of potential friends. There was some griping as there always was
when Facebook expanded beyond what was seen as a formerly exclusive
cohort. One new group was called "You're Still in High School
and You're Friending Me? That's Awkward . . . Now Go Away." But
the data told Zuckerberg and his crew what they wanted to know.
It showed that lots of communication was developing between high
school and college kids and that overall activity was going up as a result
of the change. By April 2006, Facebook had over a million high
school users.
Facebook had outgrown its cramped warren of rooms above the China
Delight restaurant on Emerson Street in Palo Alto. The company decamped
for larger quarters one block away on University Avenue, not
far from Stanford and across the street from Google's original headquarters.
Facebook relocated to a modern glass office building, indicative of
a new gravitas for the company. Moving, however, involved improvisation
of the classic Facebook variety. Everybody carried their stuff themselves.
A short procession ensued as a row of T-shirted, unkempt young
engineers pushed their desk chairs, each one loaded with an extra-large
monitor, along the sidewalk for the one-block trip.
When Facebook reached 5 million users in October 2005, it held
another party at board member Peter Thiel's San Francisco club Frisson
to celebrate—only ten months after the one-million-user party there.
Every day brought more evidence that users were infatuated with the
service. At the beginning of the school year, Facebook had nearly doubled
the number of colleges where it operated—to over 1,800. At almost
every one, its penetration among students quickly surpassed 50 percent.
More than half of users were signing in at least once a day—an extraordinary
statistic for any Internet business. And in the office, the staff was
being bombarded with emailed pictures of quails.
Users had noticed the quote from Wedding Crashers at the bottom
of the search page that said, "I don't even know what a quail looks like,"
and they were trying to be helpful. Or else they were in on the joke. Or
both. It didn't matter. They cared.
the -ratebook.effect
Users were viewing 230 million pages daily on Facebook, and revenue
had climbed to about $1 million per month. Mostly it was coming
from ad networks that were placing low-priced display ads. Sponsored
groups like the ones run by Apple and Victoria's Secret were bringing in
thousands, and announcements at individual schools generated some
money as well. But since the company's costs each month were about
$1.5 million, Facebook was burning through its capital at the rate of
about $6 million per year. The money was mostly coming out of the
Accel investment, and Zuckerberg wasn't very concerned. Neither was
Moskovitz. Moskovitz kept working like a dog, but when he wasn't at his
desk he was driving proudly around in a new BMW 6-series sedan he'd
bought in September.
There was a sense among many at the company that they were participating
in something historic. Cohler, who unlike most of this crew
had actually received a degree, from Yale in music, saw analogies. "It
was one of those moments with a unique creative Zeitgeist," he says,
"like jazz in New York in the 1940s or punk in the 1970s, or the first
Viennese school of the late eighteenth century." The conviction that
this was history in the making led people to work even harder.
The history was not being made by Facebook alone. The company
was surrounded by other companies also creating a more social Internet.
Just around the corner was Ning, funded by Marc Andreessen and
building software that enabled anyone to create their own private little
social network. Up in San Francisco, forty-five minutes to the north,
Digg was inventing a new tool that allowed people to share articles and
other media they found on the Web. Other social networks like Bebo
and Hi5 were emerging there, too, some targeting the same users as
Facebook but in any case building clever products that were resonating
with users all over the world.
Moskovitz was more interested in user numbers than historical
analogies. Ever vigilant about competitors, he was worried that
MySpace had grown from about 6 million members in January to 24
million by now. "How are they doing it?" Moskovitz asked one day.
"Fuck MySpace," Zuckerberg replied.
He had a chance to express a similar disparaging view in slightly
more polite language directly to MySpace's leaders shortly thereafter.
Zuckerberg and Cohler flew down to Los Angeles, where they sat at
a restaurant with Ross Levinsohn, head of Fox's interactive group for
Rupert Murdoch's News Corp. He oversaw MySpace. Their competitor
was being solicitous again. Levinsohn was cultivating Zuckerberg
because he wanted to buy Facebook to add to his digital portfolio. But
Zuckerberg was, as usual, just stringing him along. In her book Stealing
MySpace, Julia Angwin recounts how Levinsohn seemed dubious Facebook
could handle its rapid growth. Zuckerberg was dismissive, both of
the comment and of Levinsohn's business. "That's the difference between
a Los Angeles company and a Silicon Valley company," he said.
"We built this to last, and these guys [at MySpace] don't have a clue."
A few weeks after it hit 5 million users, Facebook added a new feature
that would transform its service. It had succeeded up to that point by
being what one employee called "brain-dead simple"—all you were
able to do was fill in your own profile and scan the information others
had put into theirs. But there was one way to customize and modify
your profile that had become very popular. Though you were allowed
only one profile photo, students were frequently changing that photo,
sometimes more than once a day. They clearly wanted to be able to post
more photos.
Photo hosting was exploding on the Internet. Earlier that year Yahoo
had acquired Flickr, a pioneering service that allowed users to upload
photos for free, and was very creative with something called "tagging."
A tag was inserted by the photographer when he or she uploaded the
photo, to label it based upon its content. A single photo might be tagged
"landscape," "Venice," and "gondola." Users could search for photos
based on their tags.
A lengthy debate ensued about the wisdom of Facebook getting into
the photo-hosting and storage business. The earlier add-on Wirehog
application, which was intended partly to enable users to see photos
on one another's PCs, had fallen flat. During the brief period when
the Wirehog application was active, few users tried it. And Zuckerberg
worried that to tinker with Facebook's simplicity was risky when the service
was growing so rapidly just as it was. But finally Parker and others
convinced him it was worth a try to build a Facebook photos feature.
"The theory behind photos," says Parker, "was that it was an application
that would work better on top of Facebook than as a free-standing application."
Some of the company's best new arrivals took on the project. Aaron
Sittig oversaw the user interface and design. Engineer Scott Marlette
wrote the software. Managing the process was newly hired vice president
of product Doug Hirsch—the fruit of Robin Reed's recruiting labors.
At thirty-four, Hirsch was an online veteran who had been one of
the first thirty employees of Yahoo.
After a few weeks, Sittig, Marlette, and Hirsch quickly came up
with a well-designed if conventional photo-hosting service. Like many
on the Internet, it allowed users to upload photos and include them
in online albums, and enabled others to comment on them. But they
knew it wasn't exactly right. Hirsch, who had years of experience in Internet
product design, suggested they take a different approach, something
uniquely Facebook. "I wish there was just one really social feature
we could add to this," he said in a meeting. Sittig, a very serious young
man with blond bangs whose impeccable beach-boy good looks are
seldom graced by more than a fleeting and wry half-smile, considered
what that might mean. "I went back and thought a bit," he recalls, "and
I was thinking, 'You know, the thing I most care about in photos is, like,
who's in them.'"
It was a breakthrough. They decided that Facebook photos would
be tagged in just one way—with the names of the people in them. It
sounds elementary but it had never been done before. You would only
be able to tag people who had confirmed they were your friends. People
who were tagged received a message alerting them about it, and an icon
appeared next to their name on the lists of friends that appeared on each
user's page.
The photos team made two other important decisions. To see the
next photo, all you had to do was click anywhere on the photo you
were looking at. You didn't need to hit a little "next" button. They were
attempting to encourage that "Facebook trance" that kept people clicking
through pages on the service. It made looking at photos simple and
addictive. They also took a gamble and decided to compress photos
into much smaller digital files, so that when they appeared on Facebook
they were significantly lower in resolution than the originals. That
meant they would upload faster, so users could select a number of photos
on their PC and see them online within minutes.
Would people accept low-resolution photos? Would they use the
tags? On the day in late October when the team turned the Photos application
on, they nervously watched a big monitor that displayed every
picture as it was uploaded. The first image was a cartoon of a cat. They
looked at each other worriedly. Then in a minute or so they started seeing
photos of girls—girls in groups, girls at parties, girls shooting photos
of other girls. And these photos were being tagged! The girls just kept
coming. For every screenful of shots of girls there were only a few photos
of guys. Girls were celebrating their friendships. There was no limit
to how many photos people could upload, and girls were putting up
tons of them.
Ordinary photos had become, in effect, more articulate. They conveyed
a casual message. When it was tagged, a photo on Facebook expressed
and elaborated on your friend relationships. "Pretty quickly we
learned people were sharing these photos to basically say, 1 consider
these people part of my life, and I want to show everyone I'm close to
them/" says Sittig. Now there were two ways on Facebook to demonstrate
how popular you were: how many friends you had, and how many
times you had been tagged in photos.
Sittig, Marlette, and Hirsch had also stumbled onto a perfect new
use for photographs in the age of digital photography. More and more
people were starting to carry cell phones with built-in cameras, using
the cameras for quick snaps of daily activities. If you always had a camera
with you, you could take a picture simply to record something that
happened, then put it on Facebook to tell friends about it. The tags on
a photo automatically linked it to people throughout the site. This was
very different from the way photos were generally used on My Space.
MySpace was a world of carefully posed glamour shots, uploaded by
subjects to make them look attractive. In Facebook, photos were no
longer little amateur works of art, but rather a basic form of communication.
In short order the photos feature became the most popular photo
site on the Internet and the most popular feature of Facebook. A month
after it launched, 85 percent of the service's users had been tagged in
at least one photo. Everyone was being pulled in whether or not they
wanted to be. Most users had their profile set up so that if someone
tagged them in a photo they received an alert by email. Who wouldn't
go look at each new picture of themselves once they got that email?
After the photos feature launched people began to come back to Facebook
more often, since there was more often something new to see.
This thrilled Zuckerberg, whose primary measure of the service's success
was how often users returned. A full 70 percent of students were
now coming back every day, and 85 percent at least once a week. This
is astonishing customer loyalty for any Internet service, or any business
of any kind, for that matter.
Immediately the question shifted to whether Facebook could handle
all the new data and traffic. It put a massive burden on the storage
and servers. Within six weeks the photos application had consumed
all the storage that Facebook had planned to use for the coming six
months. Having data center software veteran Jeff Rothschild on hand
proved fortuitous. He stayed late night after night, trying to keep the
company's servers from "redlining"— exceeding their capacity and potentially
crashing. People from across the company were drafted to trek
to the data center and help plug in new servers. Marlette, considered by
most of his colleagues a programming genius, focused on rewriting the
photo software code to make it more robust and efficient. By late 2009
Facebook was hosting 30 billion photos, making it the world's largest
photo site by far.
The success of photos led to an epiphany for everyone at Facebook,
from Zuckerberg on down. The team had built what was otherwise a
plain-vanilla photo-hosting application. But the way they integrated it
with Facebook showed the magic of overlaying an ordinary online activity
with a set of social relationships.
FallZOOS
Facebook executives were seeing the Facebook Effect in action
themselves for the first time. Zuckerberg was beginning to talk about
what he would come to label the "social graph/' meaning the web of
relationships articulated inside Facebook as the result of users connecting
with their friends. With Facebook photos, your friends—your social
graph—provided more information, context, and a sense of companionship.
But it only worked because the photos were tagged with people's
names and Facebook alerted people when they were tagged. The
tags determined how the photos were distributed through the service.
"Watching the growth of tagging," says Cohler, "was the first 'aha' for
us about how the social graph could be used as a distribution system.
The mechanism of distribution was the relationships between people."
Perhaps applying the social graph to other online activities would
make them more interesting and useful, too. But how could Facebook
help make that happen? If photos were a new application on top of the
Facebook platform, what would some other applications be? Zuckerberg
found these to be enormously exciting questions, and they dovetailed
with thoughts he had discussed with Adam D'Angelo since even
before Thefacebook launched about how the entire Internet needed
to become more "social." It was the Wirehog dream finally coming to
fruition. "Watching what happened with photos," says Parker, "was a
key part of what led Mark's vision to crystallize. He was formulating a
broader and broader theory about what Facebook really was."
Harvard continued to figure in Facebook's story. Following the success
of photos, Zuckerberg began scheming to make more dramatic changes
in the service, but to implement them he would need a bunch of new
top-quality programmers. He had been frustrated by the people who were
applying in Silicon Valley. They just didn't fit in with Facebook's culture.
They were too corporate, not iconoclastic enough, and not in his view sufficiently
creative. So he combed through Facebook looking up old teaching
assistants and other computer science majors who had impressed him
at Harvard. He wrote out a list and gave it to Robin Reed, who started
calling them. It turned out that a bunch were living in Seattle.
In January 2006, Facebook hired four former computer science
teaching assistants from the Harvard classes oP03 and '04: three worked
at Microsoft and one at Amazon.com. One —Charlie Cheever, from
Microsoft—Zuckerberg thought of as a kindred spirit because Cheever
had been brought before Harvard's Administrative Board for downloading
student information into a database. Cheever let a few friends
search through his program to find out who roomed with whom, or
which dormitory that cute girl lived in. It was an escapade not unlike
Zuckerberg's with Facemash, but a year earlier.
This influx of programming hotshots immediately brought a new
rigor and focus to Facebook's engineering. Not only were they young
enough to understand the ethos of openness and transparency that was
at the heart of the company's values, but they had several years of experience
at the best software companies under their belt. They expected
nothing less than to participate in groundbreaking Internet innovation.