What were you doing 12 weeks ago?
Twelve weeks ago was Nirav Tolia's last day on a pretty enviable job. He =
is
27, and in addition to managing the marketing of Yahoo!'s E-commerce
properties, he had represented the company on television more than 100
times. Almost nobody leaves Yahoo!, but Nirav Tolia had just heard a real=
ly
interesting start-up idea from a friend, Naval Ravikant, who had recently
left @Home. It took Tolia about one day to decide, and the following morn=
ing
he resigned.
Soon after Tolia's last day at Yahoo! he and Ravikant were joined by one =
of
the highest ranking engineers at Netscape, Ramanathan Guha, who was cooki=
ng
up an idea very similar to Ravikant's on the back burner of his big brain.
By the end of the week, they were five strong. Eleven weeks ago, despite =
not
having a single line of code written or even a paper sketch of the Web si=
te
they wanted to build, they got $8 million in seed financing from venture
capitalists -- half from Benchmark Capital, which had financed Ebay, and
half from August Capital, where Naval Ravikant was camping out as
entrepreneur in residence. This gave the start-up what is believed to be =
one
of the highest seed-round valuations ever.
Nine weeks ago, they brought on Lou Montulli and Aleksander Totic, two o=
f
the original six founding engineers of Netscape. Eight weeks ago, they mo=
ved
into a second-floor gabled loft in Mountain View, Calif., and began grind=
ing
out 15-hour days, seven days a week -- but of course these guys had done
that before. In a spring when it had started to seem to some Silicon Vall=
ey
veterans that all the big original ideas were gone, theirs was a lightnin=
g
rod for talent. The new director of business development, Dion Lim, began=
to
cut deals with other Web sites to import their data feeds. Seven weeks ag=
o,
they started hiring category managers. Six weeks ago, it became clear to
Guha that enough of the original programming was already done, and he cou=
ld
switch hats from coder to manager. Five weeks ago, their venture capitali=
st
from Benchmark, Bill Gurley, came by the office for his first look-see. H=
e
was blown away: "This is, unequivocally, the fastest I have ever seen a
start-up move."
Four weeks ago, they began to cut distribution deals; two weeks ago, they
settled on their marketing plan, and now, having reached a critical mass =
of
31 people, they are set to launch their Web site. In 12 weeks, the amount=
of
time it might take an average person to decide what kind of hedge to plan=
t
in the backyard, they built a company from scratch. An instant company, o=
r
what is being called in Silicon Valley a "second-generation Web company."
Not so long ago, it seemed incredible that a Web company could be born in=
a
mere two years. But rather than going back to normal, the pace of creatio=
n
in Silicon Valley now seems to be speeding up even more. Any Web company
that starts out today and takes two years to get up and running is likely=
to
be left in the dust.
In first-generation Internet companies, the founder and a few college
buddies moved into a garage that they decorated with Nerf guns and green
army men. In second-generation Internet companies, the staff coalesces no=
t
from friendships but from respect for mutually complementary skill sets.
They skip the garage phase, engage two real-estate brokers and make
simultaneous bids on three office spaces, hoping one comes through. They
move in over the weekend and by Monday have it decorated with Nerf guns a=
nd
green army men. In first-generation Internet companies, the staff resigne=
d
from monolithic software corporations or took leave of business school or
jumped ship on brand-manager positions at Procter & Gamble. Nobody had
Internet experience; they learned by making mistakes, of which there were
many. The purpose of the Internet was unclear. Now, companies are being
formed by staff members who have years of know-how. And they see the
Internet, above all, as a place to buy things. Some $301 billion was
generated by the Internet economy in 1998, with an annual growth rate ove=
r
the past four years of 174 percent.
Because of the way high-tech employees are compensated, there are likely =
to
be a great number of second-generation start-ups in the next year. The
notorious stock options that add up to so much paper wealth usually take
four years to fully vest. For the early movers on the Internet, the four
years are coming up. And the golden handcuffs are coming off. This
particular second-generation Internet company has managed to recruit top
people who were still handcuffed -- what people in the Valley call "the
unhirables." Naval Ravikant walked away from what at the time was $4 mill=
ion
worth of unvested @Home stock options. Ramanathan Guha walked away from
probably more than $4 million (a figure he is contractually forbidden to
confirm) at America Online, which had acquired Netscape. Sabrina Berry's
previous employer, CommTouch Software, was planning to go public. Berry
walked away from all of her shares in that company. For Lou Montulli to
join, he resigned from a hot, well-financed start-up called Geocast Netwo=
rk
Systems. At the time Nirav Tolia left Yahoo!, the unvested options he lef=
t
on the table were worth $10 million.
But, he insisted, it didn't matter to him if it was $20 or $20 million --=
he
has a dream to pursue.
"People are going to think I'm nuts," said Tolia, rolling his eyes. "Can =
we
not talk about this anymore?" said Guha. "It's painful to dwell on." Six
weeks ago, the engineering team walked to In-N-Out Burger for lunch. They
crossed an overpass above Highway 101 and paused at the rail. "This team =
has
been responsible for many of the key features of Netscape Navigator," Guh=
a
said. "These guys can point to any of these cars passing underneath and s=
ay:
'That driver has almost certainly used my code. And that driver. And that
one.' We want to build something that has that kind of influence. We want=
to
build a site that everyone will use."
So what's the idea that is inspiring so many to jump? Until this week,
they've kept everything secret, operating under the code name "Round One.=
"
In fact, not even people who come in to interview for a position learn th=
e
idea their first day. Several hours of vague conversation seem to be lead=
ing
up to the grand presentation, but alas, the applicant is sent home with a
preliminary offer, setting out salary and options and title -- and no cle=
ar
sense of what the company will do. If the candidate is sold on the team,
then she or he comes back for a second round. Only at the end of that nex=
t
day does she sit down in front of a whiteboard with Ravikant and Tolia an=
d
hear something like this:
As the Web becomes an infinite supply of goods and services, goes the pit=
ch,
people crave guidance on what and where to buy. So far, the great number =
of
on-line shopping guides present quantitative, machine-sorted and
machine-generated data: comparisons of product prices and specifications.
But what consumers need (Ravikant and Tolia contend) is a recommendation
that gets beyond that: the advice of someone they trust, someone just lik=
e
them.
Their solution is a Web site, Epinions.com, which they envision as a sort=
of
Zagat-for-everything, a site consisting entirely of consumer opinions or
reviews of anything you can buy. Epinions.com itself will sell nothing at
all -- it has no warehouse, no trucks on the back end. The money would co=
me
from deals Epinions.com cuts with companies that do sell things: every ti=
me
an "E-pinion" prompts a reader to click "Buy," the company will earn a ti=
ny
commission on the resulting sale. At the start, the E-pinions on
Epinions.com will be culled from existing sources, guiding users through
aggregations of expertise from the four corners of the Web. But the key t=
o
the whole idea is to make Epinions.com participatory, taking advantage of
what I call the Tom Sawyer model. Write and post a short review of any
product on Epinions.com, and you can earn a few pennies every time the
review is read by another user. By letting readers rate the usefulness of
the E-pinions, the most trusted ones will float to the top of every
category. As Ebay is a marketplace for products, Epinions.com seeks to be=
a
marketplace for ideas. If it catches on, like Ebay, then everything
snowballs, and these hobbyist-reviewers function as sliver-time virtual
employees who do all the work for you. "Everybody is an expert at
something," they kept repeating around the Epinions.com office; they hope
their site will be the place where everyone shares their expertise.
Similar logic has been welling up in the collective unconscious of Silico=
n
Valley, and most E-commerce sites are already adding some form of E-pinio=
n
to their Web pages. Productopia, Deja.com, Cnet, Amazon.com -- everyone's
hiring editors and bringing back the old-fashioned, well-trusted written
word. Of course, sites that both sell goods and review them are subject t=
o
criticisms of bias. Epinions.com.com would be the first company to start =
up
doing E-pinions and only E-pinions, hoping to be as Jell-o is to flavored
gelatin.
And that's it. Then again, what was Yahoo! at the start but just a Yellow
Pages to the Web? The point is that job recruits with demonstrable talent
are buying in to give it a go. And they know that in the short and
unpredictable history of Internet businesses, success has often come down=
to
getting the details right, fast. "We don't need any more strategists," sa=
ys
Mike Speiser, a McKinsey consulting alumnus who learned to curtail his ow=
n
inclination to heady analysis. This was 10 weeks ago.
"We need closers," agrees Nirav Tolia. "We need bulldogs. We need enginee=
rs
who are execution machines," says Guha. "This is not a strategy play. Thi=
s
is an execution play." First-generation start-ups raise small seed rounds=
to
develop a "proof-of-concept version," at which point the start-up has to =
go
back to dog-and-pony shows, negotiating for more money. Again the second
generation is different, faster. Prototypes, demos, alphas -- the languag=
e
of the hustle -- those words aren't even in the Epinions.com vocabulary.
Every minute spent dancing for investors is a minute stolen from the
finished product. Ravikant and Tolia's business plan (which consisted of =
16
sparse slides) had no financial projections and no budget. They negotiate=
d
for $8 million, enough that they wouldn't have to go back for more until
well after launch. They had no idea what it would cost to pull together t=
he
E-pinions they would need to stock the site, but they budgeted $5 million=
,
just to be safe.
Who's afraid of competition? 'We're a start-up,' boasts 25-year-old Naval
Ravikant, an Epinions.com founder. 'We've got focus. Nobody will be able =
to
move as fast as us. I pity the fools!' The group's biggest fear was the
wrath of prominent venture capitalists who did not get an opportunity for=
a
cut of the deal. A slightly rattled Tolia played me several phone message=
s
left on his answering machine by furious V.C.'s. One of the advantages of
combining August Capital and Benchmark is that they occupy the same
two-story building. When the terms of the valuation were set with August,
Ravikant and Tolia walked upstairs to Bill Gurley's office at Benchmark.
Gurley had joined Benchmark only a month before, and Epinions.com would b=
e
one of his first big plays for his new employer.
"I need to know if you're in," Tolia said. Gurley was calm. He recounted
some of the internal discussion among the Benchmark partners. One partner=
,
Gurley offered, had scored the idea a 6.5 and the team a 9.5 on a scale o=
f 1
to 10. But he wouldn't tell them details of how the final vote was scored.
"So where does that leave us?" Tolia asked.
"Don't worry, it's done," said Gurley.
"Should I contact your lawyer or something? Draw up term sheets?" "We do=
n't
do term sheets here," Gurley responded, offering his palm. "We do
handshakes." In those first crucial weeks, the Benchmark investment was
like having a Hertz Club Gold pass. Every service provider is overbooked =
in
Silicon Valley -- realtors, phone-system installers, furniture suppliers,
headhunters. Dropping the Benchmark name was the way to impress vendors
without sharing the idea. Everyone wants to do business with what may bec=
ome
the next Ebay, dreaming they'll be rewarded with friends-and-family share=
s
when the time comes.
Of course, they can't do everything. There was that first weekend in thei=
r
new digs, when the parts for their desks arrived from Home Depot -- 25 so=
lid
wood doors, 100 4-by-4 legs and 400 metal braces. Despite this formidable
team of engineering talent, in eight hours of off-and-on tinkering they
couldn't correctly assemble any desks. Finally they called a carpenter wh=
o
had done this before, and he started building two desks an hour.
Fortunately, what they don't know about desks, they do know about code.
Hiring staff with seasoned Internet experience has allowed Epinions.com t=
o
delegate like crazy. "We need to be told what to do but not how to do it,=
"
said Luke Knowland, who had done it before at Wired Digital.
"It would take four very bright first-generation engineers a full year to
program this site," Guha estimated. "But because we've done it before, we
can write most of the code in six weeks." Everything is faster. Zero dra=
g
is optimal. For a while, new applicants would jokingly be asked about the=
ir
"drag coefficient." Since the office is a full hour's commute from San
Francisco, an apartment in the city was a full unit of drag. A spouse? Dr=
ag
coefficient of one. Kids? A half point per. Then they recognized that suc=
h
talk, even in jest, could be taken as discriminatory in a hiring situatio=
n.
On the business-development side, "I no longer have to waste months
evangelizing," says Dion Lim, who has been cutting deals to aggregate
opinion material from existing Web sites. A couple of years ago, the proc=
ess
would have been slow and painful. "Now, I just call, and they have a
syndication rate scale and a preferred data-feed format," he says.
Meanwhile, Epinions.com has kept up constant reconnaissance on the
competitors it will be jockeying with this fall, despite those competitor=
s'
best efforts to keep their strategies secret. The Valley has what it call=
s
the "whisper circuit," which is not so much wild gossip as the ability to
call in old favors and threaten to pull people's teeth. A lot of
whisper-circuit surveillance leaks out the back door of companies through
their engineers, who often refuse to lie on principle or are very bad at =
it
when they try.
Through the whisper circuit the company learned that one potential
competitor was trying to wiggle out of a partnership so that it could
overhaul its product toward something like Epinions.com. The team learned
that a top job applicant, on the verge of accepting its offer, had been
grilled so hard by another venture capitalist that he cracked and spilled
the Epinions.com idea. (The offer was retracted.) Another V.C. was trying=
to
discredit Epinions.com by telling people he'd turned down its deal, which
he'd never seen. And it was on the whisper circuit that the Epinions.com
team learned that Amazon.com had started flying writers and editors to
Seattle and offering them positions as category editors to cover a wide
range of products -- food, video games and so on. The whispering was
specific -- that Amazon.com was offering a $65,000 salary, a 10-percent
signing bonus and options that could be worth $1 million in four years.
(Amazon declines to confirm or deny those details.) The entrance of
Amazon.com onto the scene seemed like bad news for Epinions.com. Everyone
lost sleep that night.
But their exuberance returned with dawn. "Amazon is supersmart," said Nav=
al,
marching out of Benchmark Capital's Sand Hill Road offices with his
teammates in tow. "But we're a start-up. We've got focus. Nobody will be
able to move as fast as us. I pity the fools!" Other than that first nigh=
t
with Amazon, I haven't lost a single hour's sleep over our competition,"
Nirav Tolia said four weeks ago, when he was only rationing himself four
hours a night anyway. "All the sleep I've lost has been over our internal
conflicts."
Indeed. By hiring so many bulldogs and execution machines who were all us=
ed
to being No. 1, Tolia feared the competition between employees would tear
the company apart. For the first month, without a product to obsess about=
,
they focused on their responsibilities, and the closest proxy for their
responsibilities was their title. That they had given up so much money to=
be
here made them a little testy -- they wanted constant assurance that thei=
r
career decision wasn't a mistake.
Everyone kept demanding an org chart, preferably with his or her name in =
a
box near the top. In first-generation Web companies, the premise was that=
no
task was beneath you: you did whatever it took to succeed. This wisdom se=
ems
not to have been passed down. "How do we go from a team of champions to a
championship team?" Tolia kept asking. Bill Gurley had turned Naval Ravik=
ant
on to complexity theory. "Truly alive systems exist only at what is calle=
d
'the edge of chaos,"' Ravikant said in one meeting. So though it was caus=
ing
him to lose hair, he was running the company on the edge of chaos, rallyi=
ng
people to risk making mistakes. "I don't want to be a company that plays =
it
safe." He gave his employees an org chart, and then another one every wee=
k.
Their titles became vague, more fungible.
Going through the start-up experience usually bonds a team together. Ther=
e
are those occasional "Breakfast Club"-like days when workers' inner lives
get revealed to each other. This bedrock of goodwill gets the team throug=
h
hard times later. Going through it at second-generation speed only allows
brief bonding moments. Mike Speiser covered Internet companies as an
investment-banking research analyst, but he hadn't worked at one before
Epinions.com. A few weeks ago he said: "You know what I miss? I miss thos=
e
good old days, when we had the run of the place at August Capital, hangin=
g
out and brainstorming." Those halcyon days, six weeks earlier. Nirav Tol=
ia
came up with what he thought would be a solution to distract the champion=
s
from their fiefdoms. At the all-hands meeting five weeks ago, Tolia
announced that he would shave his head if the company met its offical lau=
nch
date. This is a guy whose E-mail was "[email protected]" for a good
reason -- a hair is never out of place on his head. "When you're wonderin=
g
why you're here at 2 in the morning, think about my cue ball," he said.
Everybody howled with laughter. Then Aleksander Totic went over to his
computer and pulled up an ancient Web page, from way back in 1994. Digita=
l
photographs were posted from the period when the original Netscape engine=
ers
shipped Navigator 1.0. There at the top of the page was a picture of Lou
Montulli -- who is even more of a sharp dresser than Tolia -- with his he=
ad
completely shaved. Then everyone really laughed. "If we're going to be a
second-generation Web company, Nirav's going to have to come up with
something better," Totic chuckled.
atching an instant company get built has been slightly disorienting. Sili=
con
Valley is sustained by the myth that you can come here from anywhere with
sheer smarts and a firm handshake and make good. Second-generation Intern=
et
companies seem to seriously tip the favor to those already here. Four wee=
ks
ago on the whisper circuit, Tolia learned that an entrepreneur from Arizo=
na
was in town to shop a business plan for a company, called Publicopinion.c=
om,
with some of the same basic concepts, like rating reviews. Tolia took the
challenge seriously -- Publicopinion.com already had a prototype on line =
and
needed financing to take the next step. But the truth is that if the guy
from Arizona is only now trying to get an audience with venture capitalis=
ts,
he probably doesn't have a chance to catch up.
After Ravikant left @Home, he would still see old colleagues at parties. =
The
comment he heard from them time and time again was: "It's amazing you wal=
ked
away from all that money. I wish I was brave enough to take the chance." =
So
why did they walk away from all that money? Take it as a given that they =
all
believe in the commercial viability of the idea, but beyond that, their
comments are all over the map. One guy talked blatantly about wanting "pl=
ane
money," and how you weren't even a player in the Valley with less than $1=
00
million. A few plead that they just want to live the start-up experience,
and the money they've earned has bought them the unconditional freedom to
pursue that dream.
Now they are at the takeoff point, and their first-generation experience
can't help them. The next 12 weeks will be an even greater challenge: the
goal now is to turn a brand-new site into a hive, one that has 80 percent=
of
all E-commerce categories covered well in advance of the crucial Christma=
s
buying season. They are blindly gambling that they have the right incenti=
ves
and the right filtering mechanisms in place.
Ready. Fire. Aim.
Photos and Credits at
http://www.nytimes.com/library/magazine/home/19990711mag-tech-startup.htm=
l
Tushar J. Mehta
email: [email protected]
voicemail: (206) 405-2335