The Code

by Margaret O'Mara

  Marty Tenenbaum and Jim Bidzos had solved the problem of secure and reliable Internet shopping. Pierre Omidyar had shown it was possible to get people obsessively buying and selling with one another online. But all the VeriSigns in the world couldn’t fix Internet retail’s other big problem: the reluctance of consumers to buy something from a retailer sight unseen, at a possibly higher price, with no obvious way to get their money back if they didn’t like it.

  That was exactly the kind of thorny problem Jeff Bezos loved to solve.

  REGRET MINIMIZATION FRAMEWORK

  Jeffrey Preston Bezos’s path to the Internet economy started when David E. Shaw decided to use computers to make money on Wall Street. Shaw had a PhD in computer science from Stanford and a comfortable faculty job at Columbia when he was first lured to the Street for a high-paying job at Morgan Stanley at the height of the high-rolling Reagan years. He lasted eighteen months before he decided to leave to start a hedge fund of his own in 1988—a different kind of fund, one that applied the most sophisticated of algorithms to snap up the best deals and trade at a scale and speed only possible by computer. The trading floor of D. E. Shaw & Co. had more than four times as many Sun workstations as it had employees, machine and man working around the clock to extract more power and profit from the market. Massive returns had Shaw inundated with résumés; he only hired 1 percent of applicants. In December of 1990, he hired Jeff Bezos.16

  Born in New Mexico, raised in Texas and Florida, Bezos was another future titan who’d displayed early giftedness, from his aptitude for building electronic gadgets in elementary school to his photographic memory of every play sequence when he was defensive captain of the football team. A major influence on his young life was maternal grandfather Preston Gise, a missile specialist who had been part of the founding team at ARPA and later ran the entire Western region for the U.S. Atomic Energy Commission (his vast domain included the thousands of scientists of Los Alamos, Sandia, and Livermore Labs—the then employer of Ann Hardy, LaRoy Tymes, and many more).17

  After retirement, Gise returned to his West Texas ranch, where his grandson spent every summer. By day, Bezos learned the tough work of ranching—laying irrigation pipes, fixing machinery, vaccinating cattle—and gained an appreciation for what could be accomplished through resourcefulness and hard labor. By night, he would marvel at the vast constellations and galaxies carpeting the Southwestern sky, kindling his passion for space exploration. He briefly considered becoming an astronaut before heading to Princeton to study electrical engineering and computer science, and his fascination with outer space never diminished. Slipping the bounds of earth, shooting into the final frontier: that was the ultimate in self-reliance, the grandest exercise in long-range planning. His grandfather had been part of that great flurry of shoot-the-moon optimism in the early 1960s; the grandson wanted to recapture some of that hope and grand vision as he embarked on his own career.

  After graduation, Bezos had his pick of prestigious big-company tech jobs, but the emerging opportunities for computer scientists on Wall Street sounded more audacious and interesting. The people he admired most in the tech world were those who hadn’t played it safe, who’d been confident enough to push boundaries early on: Bill Gates at Microsoft, Alan Kay at PARC. A few years later, Bezos affixed a (mis)quote from Kay to his e-mail signature: “It’s easier to invent the future than to predict it.”18

  Although he presented an affable front, with his ever-present bark of a laugh and his eagerness to describe himself as a “nerd,” Jeff Bezos was a meticulous analyst, so much so that Shaw tasked him with assessing the Internet’s prospects early in 1994. Bezos was a big believer in finding the white spaces and jumping in, and the Internet clearly was a massive white space. Someone was going to make an immense amount of money selling things online. A few months into the research he started to ask: Why not him?

  Of course, it was a crazy idea. Bezos was raking in money in his current job, not to mention that it was the middle of the year. Walking away meant he’d lose his 1994 bonus. Was he really going to throw it away for the off chance that an Internet-based retail start-up might actually work? Bezos’s story of what happened next was one that he told again and again to curious reporters during those early years: urged by Shaw to take some time to think about the decision, Bezos decided to sketch out what he called a “regret minimization framework.” (“Only a nerd would call it that,” he’d chuckle.) If I got to age eighty and looked back on this decision, would I regret having tried this? Bezos remembered asking himself. The answer was a firm no. He knew he wanted to sell things on the Internet, and he also knew that he couldn’t do it in New York. For he’d figured out what he wanted to sell online—something that you didn’t necessarily have to touch and see first, something that wasn’t that expensive, and something that wasn’t breakable—a book. He would become an online bookseller, and he would do it from Seattle.

  As much as Seattleites might have liked to dream otherwise, and as much as reporters’ spin might have framed it as another heroic young man heading west to find his destiny, Bezos’s decision to move to the Emerald City was the unsentimental choice of a detail-driven Wall Street analyst. Most of the Internet early adopters were in California. But the rules of the Internet road were that customers in the same state as a firm’s headquarters would have to pay state sales tax on their online purchases. Scratch that; he wanted to be somewhere with a smaller population, and fewer taxpayers. The fact that Washington State didn’t have an income tax was an added bonus.

  Then there was the matter of logistics. The West Coast’s biggest book distributor was close enough to Seattle that his team wouldn’t need a warehouse of their own at first; they simply could order up the books and mail them from there. But there was another Seattle factor, one that signaled that Jeff Bezos had far more than books on his mind. The region had Microsoft and its thousands of software engineers, some of whom might be restless enough to leave the mothership and join a start-up. If this was going to be what he planned, he’d need a lot of them. “Life’s too short to hang out with people who aren’t resourceful,” Bezos once said. And he was the most resourceful of them all.19

  It took him 60 meetings and considerable powers of persuasion to raise his first $1 million from 22 investors—“anyone who knew anything about the book business did not invest,” Bezos remembered—but by the summer of 1995, Amazon .com was open for business. (Bezos toyed with calling it Relentless.com, but opted for something a little softer.) The early clientele were computer nerds, too, and computer manuals dominated the bestseller list, but that didn’t last long. It was immediately clear that selling products over the Internet fixed the chronic inventory problem that had sunk many a brick-and-mortar operator. Amazon didn’t have to keep things on the shelves in anticipation of customer demand, but it could fill that demand for any title, from the celebrated to the obscure. Consumers flocked to the service, and Michael Crichton page-turners quickly lapped the computer books on Amazon’s sales charts. “If it’s in print, it’s in stock,” was Amazon’s boastful motto, and it was right on point.20

  While Amazon was in Seattle, its ties into the Valley were so tight that it might have well been in Sunnyvale. The local presence of Microsoft, of Boeing, of the University of Washington—that was nice, but not yet particularly relevant for a garage-based operation that only had a handful of employees. Bezos needed the kind of start-up support that only the Valley could provide. Amazon became an early member of CommerceNet, adopting the software and security protocols so that nervous book buyers would stop putting checks in snail mail to pay for their new books. Then came the coup: an early funding round from John Doerr, who put $8 million and a management team into Amazon, ending up with 15 percent of what would one day become one of the world’s largest and richest companies. Little surprise that Bezos and Doerr would immediately hit it off, for Bezos, too, was a five-years-in-the-future sort of guy, “willing to plant seeds and wai
t a long time for them to turn into trees.”21

  Doerr didn’t have to wait too long for those trees, however. Between 1996 and 1997, sales jumped nearly tenfold, from $15 million to close to $150 million. Forty percent of business was from repeat customers. Amazon went public that year, with Bezos keeping control of 41 percent of the company. By the end of 1998, its customers were in the millions and its market valuation was a gobsmacking $30 billion—more than the century-old emperor of American retail, Sears.22

  The gee-whiz press stories about the nerdy guy in Seattle focused on the books he was selling and the frenzy he was creating on Wall Street. They paid less attention to the thing that was attracting so many customers and keeping them coming back for more: the data. For Amazon wasn’t just able to give its visitors the books they came to buy, but the books that they didn’t even realize they wanted. Carefully mining data from every transaction, Amazon was able to track its customers’ tastes with eerie accuracy. Bought a Stephen King thriller? Here are five other books you might like, and here is some music, too. “You might also like” became an addictive feature precisely because it was so on-target. Bezos might have seemed like a goofy book lover, but he was pure quant, and his work on Wall Street had shown him the extraordinary things computer models could do. The humming engine behind Amazon lay in its vast and tightly guarded rooms of blinking black server blades. From the very beginning, Amazon wasn’t a bookstore. It was a data platform.

  Jeff Bezos also gave Steve Jobs a run for his money as the most relentlessly on-message CEO the tech world had ever seen. As Amazon’s valuation climbed and Bezos’s net worth soared into the billions, his persona remained as earnest and humble as ever. He still drove his beat-up Honda Accord and sat at a desk made of an old door and blocky, nailed-on two-by-fours. Amazon rented grungy office space in marginal neighborhoods, their insides as messy and makeshift as a freshman dorm during finals week. It didn’t produce TV commercials, host splashy launch events, or even put a big sign out front.

  All was in service of the message that Bezos pounded home to every reporter, in every press release, and in every annual letter to his shareholders: the customer came first. Valley dot-coms might have outrageously luxe offices and holiday bacchanals. Amazon.com had door desks and pizza parties. “Everything connected with the company,” noted Adweek admiringly, “is carefully scripted to create the image of a scrappy underdog that cares more about people than profit.”23

  THE QUEEN

  As all this percolated out West, something happened back East that became one of the luckiest of the tech industry’s many lucky accidents: the stock market flooded with money. Many things triggered the flow of investment capital. The early 1990s recession ended; the pain of post–Cold War defense cuts faded. The Clinton economic program cut government spending and hacked away at the deficit, recapitalizing banks. Telecom deregulation and NAFTA opened new markets and lowered labor costs, increasing corporate profits. Bucking tradition and economic laws of gravity, Fed Chair Alan Greenspan kept interest rates low, creating incentives for investors and individuals alike to borrow, borrow, borrow.

  While money sloshed around the markets, the Valley’s PR mavens filled the business pages with stories about Internet companies and the whiz kids who ran them. As they read about the latest generation of endearing geeks—Andreessen, Yang, AOL’s Steve Case, and others—Wall Street investors got Internet fever. By April 1995, AOL’s market capitalization was close to $1.3 billion, or $640 per subscriber. A company’s actual capacity to earn a profit didn’t seem to matter in this wild world: AOL’s market cap was 7 times revenue; Netcom, an Internet service provider that went public in late 1994, had capitalization that was 14 times revenue. The market was still extremely volatile—nearly every article about Internet stocks included a worried quote from an expert who doubted their staying power—but people kept buying.24

  Then, on August 9, 1995, Netscape went public and Wall Street went nuts. The company was little more than a year old. It hadn’t made a dime of profit. But the excitement among investors was running at such a fever pitch that NASDAQ delayed the opening bell by ninety minutes that morning at the request of Netscape’s underwriters. When the market opened, the shares zoomed up to more than two and a half times their offering price, peaking at $75. Netscape came out of the day with a staggering valuation of $2.3 billion. On paper, Marc Andreessen now had a net worth of $80 million. Jim Clark was a billionaire. “People started drinking my Kool-Aid,” Clark gleefully told his biographer, Michael Lewis. “What the IPO did was give anarchy credibility.”25

  It also gave credibility, and celebrity, to a new generation of tech-industry analysts. The leader of the pack—the woman whose unerring instinct for Internet picks gave her the title “Queen of the Net”—was a thirty-five-year-old Morgan Stanley analyst named Mary Meeker.

  Just like Ben Rosen and his microcomputers one generation earlier, the Indiana native had been doing deep-dive research on the Internet while no one on Wall Street was paying much attention, and she was bullish about what she saw. She also had forged connections with key people in the Valley’s Internet revolution, including Doerr, whom she had met as Morgan Stanley took business and financial software maker Intuit public in 1993, and worked with again as the firm took Netscape public two years later. Meeker understood that the market was going to be big, and that it was going to turn tech into a fundamentally different business. “It’s a media market,” she told the San Jose Mercury News in late 1994, “and the winners will be companies that do the best job of editing and presenting information.”26

  Even after the Netscape earthquake, the tech market remained distressingly unpredictable, with stocks spiking up and down with alarming velocity. Meeker sailed above it all. “If I believe in the company,” she said. “I buy the stock.” And she believed in AOL, in Netscape, in Amazon, and in eBay. When she rated something as a “buy,” she stuck with it for the long haul. Meeker’s unflappability and her instinct for picking tech winners quickly generated a large and eager audience for her research. Morgan Stanley’s annual Internet report featuring Meeker’s insights had been released earlier in 1995 without making much of a stir beyond Wall Street. After Netscape’s IPO, the bank got so many requests for copies from unlikely places—schools, small investors—that they struck a deal with HarperCollins to publish it as a book the following spring.27

  Meeker wasn’t the only Internet-era star at Morgan Stanley. King of the dealmakers was mustachioed and assertive Frank Quattrone, who had been a technology banker in the firm’s Bay Area office since graduating with a Stanford MBA in 1981. He’d led the deals for Silicon Graphics and MIPS and Cisco, and he leveraged his Valley connections to land some of the juiciest deals of the boom, starting with Netscape in 1995. Quattrone left the bank in 1998 for Deutsche Bank and then Credit Suisse, earning upwards of $100 million per year.

  Then there was Morgan Stanley banker and close Meeker ally Ruth Porat, a Palo Alto native with deep connections to the tech world. Her physicist father, Dan, had worked at SLAC; her brother, Marc, was CEO of General Magic, an already legendary company that tried and failed in the early 1990s to build a pocket-sized computer—an iPhone before its time. Together, Ruth Porat and Meeker vetted nearly every Internet start-up that came down the chute in the late 1990s, and the bank became the primary manager on fifty of them.28

  Some observers wondered whether it was healthy to have tech’s most bullish analyst and its most connected bankers working together so closely; some sharp-elbowed bankers (including Quattrone) already had a reputation for pressuring their researchers to talk up a stock. But such qualms quickly got lost amid the enthusiastic din of a market that was going wild for all things Internet.29

  Soon it was hard to pass a newsstand without seeing magazine covers emblazoned with the smiling mugs of the newest generation of high-tech cowboys and cowgirls. Freshly-minted MBAs opted for jobs at tiny Silicon Valley dot-coms instead of the
Fortune 500, willing to make the risky bet that stock options would be a path to fame and fortune. The Wall Street IPO became a rite of passage for nearly any Silicon Valley start-up, no matter how new or untested. “Why do they do IPOs?” parried one broker. “Why do rock stars marry models? In part, it’s because they can.” Wall Street was clamoring for all things Internet, Meeker observed. “These companies would be silly not to take access to capital.”30

  CREATIVE ACCOUNTING

  The storming of Wall Street by new, mostly profitless dot-coms wasn’t embraced by everyone. Business columnists tut-tutted about “Netscape fever” as soon as Clark and Andreessen’s company shot out of the gate. “For someone who looks at the fundamentals, this really represents a dangerous sign of overspeculation,” warned one analyst.31

  There was reason to be concerned if you knew much about the fundamentals underneath the dot-com flash and pop. Not only were most of these firms profitless, but they used a particularly creative—and, frankly, deceptive—kind of accounting when reporting to Wall Street. As Valley firms had done ever since the first tilt-ups appeared amid the fruit orchards, the new crop of Internet companies gave employees stock options, luring in talent that start-up–stage enterprises otherwise couldn’t afford. This made the payroll line smaller than it otherwise might have been on the balance sheet, and built loyalty and drive among employees, whose job-hopping inclinations were tempered by the golden handcuffs of stock options that had yet to fully vest. Like any trait developed in relative isolation, the Valley’s use of stock-as-compensation became more noticeable and distinctive over time.