by Gina Keating
In its place, Meyer and Kish came up with a Remind Me icon—an extended index finger with a red bow tied around it—to give customers a way to designate films they were interested in. The programmers disliked the rudimentary graphic and ribbed Kish about it, nicknaming it “The Bloody Finger.”
Lowe wanted to build in a digital shopping assistant, based on his observation of customers at his Video Droid stores; they often avoided asking his opinions about movies and went straight to clerks who shared their tastes.
Ideally, the digital shopping assistant would have a personality and a photo and could point customers to movies they would like in Netflix’s (future) vast library. They weren’t able to build the shopping assistant for the launch either, but the idea became the basis for a recommendation engine that also had been on an early drawing board.
Another of Lowe’s dicta—never falsely raise customers’ hopes by placing empty VHS boxes on shelves, as Blockbuster did—took shape as a search engine quirk that led customers deep into the catalog and away from newly released titles. Meyer programmed it to guide people to films that were in stock, as opposed to showing them everything that was available.
The decision to deemphasize new movies also sprang from problems Lowe encountered in negotiating a break on DVD prices to mitigate the soaring cost of building Netflix’s inventory.
The typical reaction he got from the studios when he came to meetings representing Netflix was “Okay, kind of an interesting idea, but this will never work. We’ll sell you some movies but don’t expect any price break.”
With the studios unwilling to drop the fifteen-dollar wholesale DVD price, Netflix learned to be judicious in how it drove interest and in which films it featured on the Web site. Instead of featuring the latest or most popular DVD releases, movie promotions centered on lists of older films inspired by holidays, popular actors, or news events.
To generate more rentals of the older and obscure movies that made up the bulk of their early inventory, they implemented a movie ratings system that grouped customers together in “mentor groups,” which were designed to direct customers to films they might never have considered. It was known as collaborative filtering and worked on the premise that, if two customers rated ten films the same way, they would likely enjoy other films that each rated highly.
Netflix’s focus on niche and foreign films sharpened with Lowe’s discovery that many of the service’s early customers were Indian students and immigrant technologists whose selection of Bollywood films was limited to what they could find at local Indian markets. At Lowe’s request, the Web site began surveying customers, and it found a deep interest in Hindi films. It was a good lesson, and soon Netflix became known in other immigrant communities, as well as among cinephiles, as the best source for Japanese anime and Chinese martial arts movies. The inventory initially included soft-core pornography but nothing with an X rating—not because of any moral objection, but because Randolph wanted to wait until federal trade laws clarified the patchwork of the states’ obscenity laws.
“The last thing we want is to be dragged into court on some obscenity charge just because some small-town DA is behind in the polls,” he said.
By July, the operation had outgrown the Scotts Valley office, and Randolph was looking for a bigger space. Bridges’s newsgroup influencers still drove most of the Web site traffic, but he and Smith had begun thinking about new ways to use their meager advertising budget to get mainstream customers in the door.
Smith initiated an affiliates program after the launch, copied from Amazon, to pay newsgroup tastemakers who sent followers to Netflix. She and Bridges persuaded Randolph to spend money on a few critical Web sites, such as Harry Knowles’s aintitcool.com (which evolved from Knowles’s newsgroup postings) and Bill Hunt’s digitalbits.com, in exchange for regular mentions by Knowles and Hunt to their young, hard-core movie fans.
It was tough at first to find effective ways to reach people who both owned DVD players and shopped online. Randolph mused that out of a thousand banner ads, perhaps one would reach a target consumer. Besides, Hastings often disparaged start-ups that waged expensive advertising campaigns before they had anything to sell.
Online commerce still accounted for less than 1 percent of U.S. retail sales at the time, and Randolph knew he had to use conventional tactics to find more mainstream customers. Before the launch, Randolph, Smith, and Kish designed a plan to place coupons into DVD player boxes, and went out to sell it to the mainly Japan-based electronics manufacturers.
Randolph persuaded Kirby Kish, then working for a company that made microprocessors for DVD players, to tap his U.S.-based contacts at DVD manufacturers.
In January 1998, Kirby approached representatives for Sony, Toshiba, Pioneer, Panasonic, and Philips at the Consumer Electronics Show—the premiere U.S. hardware show—in Las Vegas, cornering them in corridors and speed talking through five-minute pitch meetings.
Without a Web site, or even screen shots, to illustrate Netflix’s novel rental plan, Kish ran into a wall of skepticism. The manufacturers’ representatives did not shut him down immediately, but their dismissal of Netflix’s business model was disheartening. “Why do you think you can compete with Blockbuster?” one rep asked him. “I don’t understand that concept—this is not how people get VHS tapes,” another said, waving him away.
Finally, Kish persuaded his Toshiba contact to meet with Randolph at Toshiba’s New Jersey offices. Randolph convinced the corporation’s U.S. representatives that the proposition would offer Toshiba a leg up on market leader Sony, which would not even take his calls. The Toshiba deal later opened doors to Hewlett-Packard and Apple, whose new Internet-ready HP Pavilion and Apple PowerBook featured DVD-ROM drives. Months later, Sony agreed to a meeting when those deals became public.
Manufacturers quickly discovered that Netflix offered a way out of a dilemma that was holding down sales: Consumers did not want to buy DVD players because DVDs were not widely available at stores. Retailers did not want to stock DVDs because no one had DVD players.
By including a Netflix coupon in the box a DVD player manufacturer could promise consumers access to a library of more than one thousand titles.
It took most of their first year to get the coupons in the coveted bill of materials (BOM), but by the end of 1998 customers were beginning to claim the free rentals that came with their DVD players.
Lowe, who had taken over the negotiations with electronics manufacturers, now leveraged his entertainment industry contacts on Netflix’s behalf. He persuaded David Bishop, president of a newly formed nonprofit home entertainment lobbying organization called the Digital Entertainment Group, to mention Netflix in a speech to the media and the industry in Las Vegas in June 1998.
The brief mention—about how Netflix was promoting DVDs—was enough to stir interest in the convention hall. A stream of industry executives stopped and talked to Lowe about his new start-up. Little by little, Lowe built Netflix up as a player in the new digital entertainment ecosystem.
Lowe extracted a joint promotion from Warner, whose home entertainment chief Warren Lieberfarb was later billed as the father of the DVD, for the release of Warner’s LA Confidential and Boogie Nights titles, as well as twenty-five thousand dollars for a full-page ad in the San Francisco Chronicle.
Lieberfarb meant to undercut video rental and shift home entertainment revenue back to the studios by retailing movies in the new digital format for as little as twenty-five dollars. He had backed the DVD aggressively, making Warner the first to release new movies in the new format; and ending the studio’s use of the VHS format in 2000.
Despite Lieberfarb’s strong-arm tactics, to force the adoption of a universal DVD standard, he initially had difficulty getting industry players to endorse the format over a competing technology: DIVX was backed by Circuit City and an entertainment law firm representing rival movie stud
ios. Ironically, Lieberfarb had to beat back DIVX by providing movie rental companies, including Netflix, with promotional kits that included DVD players and free rentals to acquaint consumers with DVD technology.
For four dollars, customers could watch DIVX disks for forty-eight hours after activating them in compatible players. The disks then became unplayable unless the buyer paid an additional fee for more time. The player had to be connected to a modem to verify that the heavily encrypted disks were playable.
Circuit City had backed the format with a $100 million investment that included “market development funds” paid to studios and retailers. DIVX was set for a mid-September 1998 launch.
Although Randolph considered Warner Home Video a major supporter, he was at first determined to be agnostic in the DVD versus DIVX format war.
Corey Bridges in particular argued strenuously against carrying DIVX, especially in light of Netflix’s budget strains. He placed the format high on a list of offensive consumer products. “You pay for it, but it’s not yours,” he complained.
Bridges was convinced that Netflix could not survive a format war, and he conceived his own “dark ops” plan to kill DIVX.
He had created an extensive roster of aliases, or “sock puppets,” that he used to make newsgroup postings, sometimes to prod discussions toward a certain topic or opinion. He now turned to these sock puppets, planted like spies inside a foreign government, to conduct highly choreographed arguments designed to discredit and vilify DIVX and to draw real consumers into the fray. To Bridges’s delight, consumer sentiment soon turned against the closed system.
Randolph didn’t exactly encourage him, but didn’t stop him either, and even suggested with winks and sidelong looks that his “people” carry other messages to the online forums.
By late 1998, Randolph had changed course and bet Netflix’s future on the success of DVD, mainly because the cost of stocking both formats was too great.
Since Hastings’s initial investment, he and Randolph had tapped Hastings’s contacts in the venture capital community to secure operating funds from a couple of angel investors. In August, Netflix scored a $6 million investment from Institutional Venture Partners, but those funds were barely enough to carry the company through the end of 1998.
Netflix was losing money on every transaction, mainly because of the high cost of both the free-rental coupon program and the labor-intensive mailing operation. And Randolph had picked up a troubling trend in the constant market research he and Kish conducted via the Web site: Most customers were not returning to rent at Netflix after exhausting their free rentals.
Their strongest revenue growth came from the sales of DVDs—a business that was doomed as soon as bigger retailers like Amazon or Walmart started carrying the format. He and Kish tested initiatives, one after another, aimed at customer retention—buy a punch card, buy ten rentals at once, two-week rentals, ninety-nine-cents-per-day rentals—but nothing worked.
One evening, as Randolph stood looking out a window at the Scotts Valley office into the parking lot, the responsibility of running a company upon which so many car payments and mortgages and lives were depending nearly overwhelmed him.
The Internet Recording Media Association forecast eight hundred thousand DVD player sales in North America in 1998, with first-adopters buying fifteen to twenty DVD titles, and an installed base of 8.6 million players by 2002. Why couldn’t Netflix find a viable business model? he wondered.
Randolph later recalled in a letter to protégé Matt Mireles, who went on to found and become chief executive of SpeakerText, that Randolph “pissed blood for years keeping Netflix alive while we figured that shit out—as did every successful entrepreneur in the valley.”
A few self-inflicted wounds that autumn added to a sense that they might be in over their heads in trying to ride an unproven technology into an unexplored world of online commerce.
In September, Lowe and Randolph decided to burn DVD copies of President Clinton’s grand jury testimony in the Monica Lewinsky matter and sell them on Netflix. When the DVD duplication house got backed up with other orders, Lowe agreed to forgo DVD labels to speed up delivery of the already sold disks.
Netflix took more than a thousand orders for the Clinton DVD—spurred by news stories about the offer in the Washington Post and New York Times. Randolph and Lowe counted the promotion a big success until a few days later, when they learned that, due to a mix-up at the dub house, a few hundred customers received DVDs containing hard-core Chinese pornography instead of the Clinton testimony.
Another rookie mistake cost them more than a little embarrassment: Netflix had not assigned unique numbers for each DVD player coupon, to prevent fraud and duplication when setting up the fast-growing program, because the company’s databases could not handle a matrix so large.
They expected a small amount of theft of free rentals, but a review of the shipping database—after vault workers realized that they were shipping dozens of free DVDs to one particular address—revealed that some customers had used the codes over and over. There was no telling how much the coupon fraud had cost in shipping fees, but there wasn’t much they could do to stop it without penalizing honest customers.
Around the same time, Hastings began showing up at the Scotts Valley offices more often; he admitted that he had become disenchanted with classmates who seemed more interested in the pay bump that came with a graduate degree than in solving major problems in education.
Plus, Hastings’s forays into political activism were becoming much more interesting. He helped organize the Palo Alto–based Technology Network lobby group, and became its first president. TechNet had a short agenda: to limit shareholder lawsuits against tech companies over stock volatility; to expand foreign visa quotas for high-tech workers; and to institute education reforms to improve students’ math and science aptitude.
Hastings also helped write a California ballot initiative expanding the number of charter schools in the state, and threw the weight of TechNet behind it. The move infuriated the powerful California Teachers Association. However, the threat of an initiative backed by the money and political might of Silicon Valley broke a years-long deadlock over charter schools and spurred a compromise plan that passed the California legislature in late April, a couple of weeks after Netflix launched.
Under Hastings’s leadership, TechNet endorsed and raised millions of dollars for high-profile Democrats, including the embattled President Bill Clinton. By late 1998 Hastings could claim concrete victories for TechNet’s agenda, including congressional legislation addressing both the securities litigation and visa issues, but it had come at a price. Conservatives in the fledgling organization were furious at the overtly political agenda—tilted toward liberals—that Hastings had brought to the group.
As some at TechNet agitated for his ouster, Hastings took a renewed interest in Netflix, which by then was attracting almost as much press as he had received at TechNet. He announced his resignation from TechNet in January 1999, saying his new start-up was claiming too much of his attention,
The start-up Hastings joined in Scotts Valley had the atmosphere of a laboratory of mad scientists—a creative, unstructured workspace still furnished with worktables. There were no set hours or meetings with agendas—staff members showed up when they needed to be there and stayed as long as it took to finish a project. Meetings were called by consensus—usually an hour or two beforehand. Everyone had an opinion about every project underway, because they could all hear and see what everyone else was working on.
Randolph had worked with all his senior managers before and trusted them to find their own ways to reach a collective goal. Rather than acting as an overlord ordering around submissive subordinates, he preferred to set the company’s direction and tone and occasionally step in to align the disparate departments.
It was an extremely creative time, he belie
ved, because communication among the staff and management was so unhindered. Randolph thought his chemistry with Hastings was energizing—they represented the yin and yang of the cerebral and the intuitive. The freewheeling and sometimes volatile discussions that accompanied many staff meetings were invigorating and vital to proving the best ideas.
For Hastings, the strife was unnecessary and irritating. He was not one to lose control of his feelings but nevertheless clearly disliked being opposed. In the coming years, those who disagreed with him too often would find their personal capital dwindling and themselves marginalized at Netflix.
Shortly after Hastings arrived, he announced without preamble at an executive staff meeting that he wanted to run Netflix, and that he and Randolph would act as co–chief executives. From the wan look on Randolph’s face it seemed clear to some in the room that Hastings either had not discussed the idea with him or had sprung it on Randolph shortly before the meeting.
Hastings then turned to Lisa Battaglia Reiss, the human resources manager that Randolph had recently hired from Borland, and laid her off in front of her shocked coworkers, saying he wanted to bring in Patty McCord, his longtime personnel manager from Pure Atria. McCord managed to smooth over his worst gaffes at Pure Atria, and people who knew Hastings well believed that she acted as an emotional sense organ to prevent him from alienating staff—mainly nonengineers—who did not respond well to his blunt observations and pointed criticism.
Hastings would often say in interviews years later that he had had the good fortune to remedy at Netflix the mistakes he made as a young CEO at Pure Atria. Those lessons included ruthlessly weeding out bureaucratic habits that slowed agility and focusing on one or two core competencies, but it apparently did not extend to treating his employees with more care.
And yet, Hasting was never mean-spirited. His constancy in demanding the best from his employees, and in acting in the company’s interests, won him their admiration and loyalty. It was simply as if everything, including human relationships, broke down to a mathematical equation. Hastings kept an employee who challenged or irritated him only as long as the cost to the company of eliminating that person was too great.