by Gina Keating
Everyone had a say in the two main tasks: building a Web site with an enticing and easy-to-use customer interface and a smoothly functioning back end; and getting DVDs into customers’ hands quickly and in one piece.
Kish and Meyer set to work on the Web site. Kish sketched out crayon drawings to show Meyer and Boris Droutman in color the steps of the journey she wanted customers to take from their arrival in the virtual store until they paid and left the site with the promise of receiving their purchases quickly.
Meanwhile, Vita Droutman worked with Jim Cook to create programs to ring up sales and rental transactions, maintain inventory, move DVDs in and out of shopping carts, and charge customers’ credit cards.
Boris and Vita met as teenagers and immigrated to the United States separately as college students—Vita as a political refugee. They were working as programmers at separate companies when Meyer, who had worked with Vita at auditor and professional advisory firm KPMG, asked them both to help with the new start-up. Vita, petite and spirited, joined first, and the tall and laid-back Boris came along a few months later, after working only at night to help set up the new company’s content management system.
They had initially hesitated to leave two steady jobs to work at the same start-up, but the chance to work together, and with Meyer, was too tempting. Besides, they were still in their twenties—the youngest team members—and like most programmers in Silicon Valley in those days, they were always fielding calls from employment recruiters.
The push-pull of turning the marketing department’s concepts for the user interface, inventory management, shipping, and even the credit card application into intuitive and consumer-friendly software was so satisfying that the grueling hours of coding seemed like play. Marketing and technology waltzed together in a harmony that the young technologists had never before experienced.
Meyer insisted on customizing Oracle programs to one day handle search and fulfillment operations for up to ten million users. Vita Droutman thought the number was a bit ridiculous—they all believed they’d be lucky to get one hundred orders in the first month. Meyer also instructed Vita and Boris to build the platforms that would accommodate later features that they didn’t have time or money to build before launch.
The list included a recommendation engine to suggest movies based on customers’ previous choices, a reminder function to retrieve titles that customers had flagged on earlier visits, and a subscription plan.
The initial business plan called for both selling and renting DVDs on an à la carte basis. The company would charge the video stores’ going rate for VHS rental—four dollars plus two dollars shipping for a single disk, and three dollars for each additional one. Renters could keep movies for seven days and return them in a postage-paid mailer. If customers liked a particular title they could buy it for 30 percent below the retail price. Their hook would be something that retail stores could not offer—a selection that included almost every DVD ever released.
The title count in late 1997 was an underwhelming five hundred DVDs of mostly older movies. Only Warner Home Video had risked releasing new titles onto DVD, and strictly because its home video chief, Warren Lieberfarb, was pushing the format.
Before he took over finance and operations, Jim Cook had reviewed Randolph’s business plan as a favor to Kish, a former colleague at Intuit. He had had serious doubts about its viability.
Based on his experiences as VP of finance and operations at the Internet Shopping Network, Cook saw problems with nearly every aspect of the proposed video rental operation: the labor costs for filling orders were too high and the costs of replacing broken and lost DVDs were unpredictable. DVD players were still too expensive and titles too few to interest mainstream consumers. The technology was iffy—even duplication houses were having trouble standardizing copies so that all DVDs worked in all players.
Randolph and Kish went away from their first meetings with Cook and worked out reasonable fixes for each of the obstacles he raised and returned with a modified plan.
The moment that Cook decided he wanted to join the venture came when Randolph asked him pointedly, “Jim, don’t you agree that when we solve all these ‘what ifs?’ it will create such a high barrier to entry that there’s no way anybody else could solve this stuff, because it’s so hard to do?”
Cook spent the first three months as the new operations chief at the general mail facility on Meridien Road in San Jose, learning everything he could about the U.S. postal system. The ability to deliver product economically, promptly, and in one piece would make or break their new company, so mastering the post office was critical.
Smith took charge of designing the mailers, and she enlisted an outside design firm to work rapidly through dozens of possible layouts, sizes, inks, and paper stock, as they learned from their mistakes.
She pressed Cook, Kish, and Lowe to recruit friends and family members in cities across the country to help field-test each version. They mailed test shipments back and forth for weeks, asking their testers to report by e-mail about the condition of the envelopes and the DVDs. So few people had DVD players at the time that they often had to have their testers send the disks back so they could pop them into the office’s players to see if the disks had been damaged in the mail.
The team argued over whether to leave the company’s name off the packaging to discourage theft of the DVDs, and they fussed over minute details in the placement of the crucial bar code that let the mailer bypass the high-speed drum sorters that frequently ripped open envelopes and crushed disks. San Jose post office officials even allowed Cook to dump trays of DVDs into its sorters, day after day, to watch what happened.
Slowly, over dozens of iterations, the elements of a viable fulfillment operation took shape. Randolph discovered “skip shipping,” a way to bypass all automation by sorting mail into twenty-seven bags by zone and delivering them straight to the post office freight docks. Once Netflix went live, Cook generally made the last post office run, stuffing mailbags into his Merkur Scorpio and dropping them at the freight bays before the 9:00 P.M. cutoff.
The result of their experiments was a three-part mailer constructed of stiff but light cardboard that could both fit up to three disks at a time and serve as a return mailer after the original address flap was torn off.
Cook also tested layouts for the tiny DVD storage vault and timed workers to see how fast they could find titles, pack and label them, and sort them into mailbags. The most efficient configuration turned out to be a miniature reproduction of a Blockbuster store. DVDs hung in glassine envelopes from pegboards that covered the walls and rows of shelves set up in the middle of the room. The aisles were so narrow that only one person could pass at a time. It probably did not meet Occupational Safety and Health Administration standards, but Cook figured they’d have plenty of time to perfect the fulfillment operation as orders slowly grew.
The process of naming the company dragged on for several weeks. Randolph incorporated it as Kibble Inc., in a nod to their contention that, above all, the “dog” had to want to eat the “kibble.” Randolph, Kish, and her husband, Kirby, created a working list of possible names during a brainstorming session shortly after they moved to the Scotts Valley office.
They decided the name should consist of two syllables—something about the Internet and something about movies. One afternoon they scribbled side-by-side columns of movie terms and Internet slang on a whiteboard in Randolph’s office and matched them up.
Randolph left the list on the whiteboard and encouraged team members to add to it. His personal favorite was Replay.com, but Directpix.com, NowShowing.com, Netflix.com, eFlix.com, and CinemaCenter.com were all contenders. The Droutmans and Meyer favored Luna, after the black Labrador retriever who accompanied Randolph to work each day, as well as for the lunatic nature of their venture.
With time running short to design and print a new logo
for PR and promotions, Smith pushed them to decide. Most of the team already favored one choice, so one day, without fanfare, they decided: “We’re Netflix.” They paired the new name—with a capital F to emphasize the film connection—with a purple-and-white logo that depicted an unspooling reel of film.
In January 1998, Randolph and Smith began formulating a launch and customer acquisition strategy with the help of Corey Bridges, a bright and hyperkinetic twenty-nine-year-old product manager who had worked with them at Borland. Bridges was coming off a stressful stint at Netscape Communications, where he had helped launch the company’s iconic Internet browser. His specialty was tapping Usenet newsgroups and discussion groups, a forerunner of blogs, to spread the word about new products among early technology adopters.
Randolph and Smith had had to work hard to persuade Bridges to delay his plan to become a Hollywood screenwriter and instead help market Netflix. With less than two months to go before the planned launch, Bridges gave in, mainly because he admired Randolph and had enjoyed working with the calmly capable Smith at Borland. “I’ll only work fifty hours a week—sixty at the most,” he warned them.
His engineer roommate at the University of California–Berkeley had turned Bridges on to Usenet, a university-based precursor to the Internet, and its tightly focused newsgroups that ran ongoing discussions of topics ranging from esoteric scientific theories to purebred puppies. Bridges watched as the Internet developed, how quirky online communities formed and acquired their own tastemakers, bullies, and sheep, just like in the real world. The dynamic of the virtual societies fascinated him, and he studied them carefully.
DVD player owners were the ideal test market; they were among the most discriminating early technology adopters, already talking about their newest toy online.
He set out to influence the influencers—he’d drop into discussion groups posing as a consumer to observe, and then he’d contact the main players secretly to say, “I’m working on something interesting that I think you’ll like.”
Once he had hooked a couple dozen influencers, he recruited them for a closed beta test of the site with the promise that if they gave Netflix feedback, they could be the first to post the news on launch day.
Every one of them said yes. “It’s like shooting fish in a barrel,” Bridges exulted to Smith and Randolph. “I can post something that literally every potential customer of Netflix will see.”
By early spring, the Web site and back-end systems that enabled customers to search the inventory of films and order them were in place. By later standards it was rudimentary: tiny graphics and short blurbs about each film set on a vast white field. They had to cut a deal to use movie descriptions generated by allmovie.com, a cinephile Web site based in Ann Arbor, Michigan, after the studios refused to grant Netflix permission to use box art and copy on its Web site. The Netflix team decided to scan in photos and titles from the boxes anyway and wait for the cease-and-desist letters.
The staff had grown with new programmers crammed together at the folding tables and marketing hires sharing Kish’s cramped office. The dress code was beyond relaxed. Randolph sometimes showed up in jeans and a T-shirt that he had dropped on the floor before crawling into bed for a few hours’ sleep, only to put back on in the morning. Kish slept at the office on the increasingly frequent nights she worked too late to drive home to Redwood Shores.
A stultifying funk built up from too many bodies crammed into a small space with a balky ventilation system. Mountains of discarded DVD jewel cases piled up in one corner as the DVD inventory grew and the vault readied for action.
They all had a lot at stake, and it started to show in arguments that often escalated to shouting matches. They were not a typical start-up staff of kids just out of college—most had had seniority at big software names, and all had taken big pay cuts to work at Netflix. They had done so gladly to realize a shared dream of a consumer-oriented company that reflected their ideals and carried their intellectual DNA.
It felt more like a family than a company, because they were all so passionate about every decision, Vita Droutman reflected in March 1998, as launch day loomed. On April 14, 1998, six months after the first line of code was written, they were ready. Netflix went live.
Hastings, now in his second semester of Stanford’s graduate education program, came to watch the launch but stayed on the fringes. He and Randolph had conferred frequently over the previous six months, but he was in the office so little that many of the newer employees barely knew him.
That morning, Randolph and Smith sat for two press conferences over the phone with a full roster of journalists calling in from mainstream newspapers such as the San Jose Mercury News, venture capital magazines such as Red Herring and Upside, and the tech Web site CNET.
Meyer pushed the site up, and Bridges unmuzzled his newsgroup influencers. Curious visitors filed into the virtual store. The site worked exactly as planned. As the visitor count ticked higher, Meyer began getting nervous.
The servers reached capacity about ninety minutes after launch—and crashed. Meyer sent Boris Droutman to a nearby Fry’s Electronics, in company controller Greg Julian’s battered Toyota pickup truck, to buy ten new computers to boost capacity while he worked on a fix to bring the site back up.
In the vault, the laser printer jammed, unable to handle the heavy flow of orders. The pegboards were in disarray; orders were half completed and lying in piles on benches as workers squeezed through the narrow aisles, trying to keep up.
Everyone who wasn’t working on fixing the capacity problem pitched in to help with the orders that flooded in each time the site came back up. By nightfall they had received more than one hundred orders to ship more than five hundred disks, and Meyer was still struggling to keep the site online.
“We have to put a page up on the site saying something like ‘The store is too crowded; come back later,’” Hastings told them.
That’s funny, Smith thought. It’s the Internet. It can’t get too crowded. Nor had it occurred to her until that moment that on the Internet the store never closes.
CHAPTER TWO
THE GOOD, THE BAD, AND THE UGLY
(1998–1999)
IT HAD BEEN AN ABSTRACT thing; a concept they invented, converted into a business plan, and launched into existence in the space of a year. The Netflix team set out to challenge an established industry, and the reality of that wish came calling as soon as the launch was over.
The team had not expected the level of success or attention that they received from day one, and they quickly started to feel the responsibility of not screwing it up. So this is what it’s like to have a tiger by the tail, Kish thought.
Lowe came home the night of the launch and told his wife he was unsure, after all the work that had gone into launching Netflix, what came next. “It’s like having a baby,” she told him. “Now that it’s here, what do you do with it?”
Like a real infant, Netflix gave its caretakers many more sleepless nights, and suffered from growing pains associated with disrupting two business paradigms—the VHS format and the bricks-and-mortar rental—in its first year of life.
“We had difficulty using the Netflix site last week and apparently weren’t alone—site carried message ‘Sorry, due to extreme opening week demand, the Netflix Store may be slow,’” Audio Week wrote the week after the launch. “Spokeswoman said site has had ‘overwhelming’ number of visitors since April 14 launch but couldn’t provide specifics on hits. However, she said company tripled site capacity to accommodate demand, and we saw improvement by week’s end.”
By launch day DVD players were selling far faster than videocassette recorders—four hundred thousand units in the first six months after the optical format’s U.S. release in March 1997. It had taken VHS two years to get into half as many households. Prices of players were falling fast, to an average $580 in April 1998 from $1
,100 a year earlier.
Initially wary, movie studios finally accepted the new format and were releasing DVD titles at a clip of one hundred per month. The number of titles in Netflix’s vault climbed to fifteen hundred by August. Blockbuster and Hollywood Video, correctly viewing the new format as a threat, refused to stock it in stores, ceding the field to Netflix throughout the summer of 1998. Every factor that could have tripped Netflix up had instead fallen exactly the right way.
“Be sure to bookmark Netflix.com to revisit as their popular titles become available,” CNN correspondent Dennis Michael urged viewers of Showbiz Today that summer.
Consumer adoption of the DVD format surged, and the Netflix team found itself swept along. In the first four months the vault mailed out and got back twenty thousand rented DVDs, and Netflix hit $100,000 in monthly revenue, becoming in theory a $1 million company.
The launch-day crash of the Netflix site was a harbinger of the work that still awaited Randolph and his team in the months that followed. He and Kish and Meyer had made trade-offs to get the site up and running within the time Hastings had specified when he funded them. In prelaunch brainstorming sessions they had outlined but not built a suite of features they felt were important to make the site compelling and easy to use.
Kish and Meyer wanted a function to remind customers about movies they wanted to see, inspired by Kish’s trips to bookstores to scope out new titles so that she could borrow them from the library.
In internal memos they called the feature The List, but it came to be known as the Queue when Meyer built it a year later. It would be difficult to build, because it included “intelligent” functions that allowed customers to sort through and prioritize inventory in individual accounts. Meyer had not wanted to spend time on it until more critical functions were added.