How the Internet Happened

by Brian McCullough

  Netflix also benefited from being essentially the only game in town for a while. The incumbent rental behemoths—Blockbuster, Hollywood Video and Movie Gallery—were reluctant to embrace the new format. They had been burned by the earlier LaserDisc technology, which had only proven popular to a niche audience. When, in the summer of 2000, Netflix even offered to sell itself to Blockbuster for about $50 million with the express idea that Netflix would become the DVD channel for Blockbuster, thereby saving it from the costly transition of its inventory from VHS, Blockbuster said no.31 It still didn’t believe DVDs would catch on.

  Netflix was originally launched as a hybrid service. You could rent DVDs à la carte for $4 a piece plus $2 for shipping the disc to you in those little red envelopes.32 But you could also buy the DVDs to own, just as a dot-com named Reel.com was offering at the time. The problem was, neither option proved very profitable; as other ecommerce companies had learned, for such small-dollar items, the shipping costs really ate into margins. When Reed Hastings moved from being merely Netflix’s biggest investor and advisor to becoming its full-time CEO in 1999, he discontinued the retail sales, and the company began experimenting with different rental schemes in an effort to find profitability. The model Netflix settled on in September of 1999 was originally called the “Marquee” plan. Subscribers paid $15.95 per month for the privilege of renting four (later, just three) movies per month. If you finished one movie, you simply mailed it back and Netflix would send you another. There were no other fees—especially no dreaded late fees. Hastings called the new program a “near DVD-on-Demand service.”33 Netflix’s rental volume increased by 300% in just three months.

  Even though it stumbled upon the strategy accidentally, Netflix would cleverly use the “no late fees” mantra as a way to position itself as the populist champion of the consumer. “Movie renters are fed up with due dates and late fees,” Hastings would tell the press. “With no due dates, our customers can stock up on rental movies and always keep a few on the top of their television, ready for impulse viewing.”34 Certainly, fining your customers for using your service is never a popular business model, but late fees accounted for as much as 13.4% of revenues for a rental chain like Blockbuster.

  Eliminating the late fee made for friendly headlines, but it was not what made Netflix take off.35 What mattered was that Netflix too had learned the key lesson of retailing in the Internet Era: unlimited selection, (near-) instant gratification. Whereas a typical brick-and-mortar video rental store carried 3,000 titles, Netflix carried tens of thousands.36 With Netflix, you could almost always get a movie you wanted to see. Compare that to the experience at a typical Blockbuster, where limited inventory meant one in five shoppers went home empty-handed. New releases were often rented out by other customers, so typical Blockbuster customers had to visit a store five consecutive weekends before they could actually take home the movie they wanted.37 Blockbuster even had an internal term for this experience: “managed dissatisfaction.”

  And more important, Netflix’s combination of a website storefront coupled with postal delivery proved infinitely more suited to satisfying modern customers. Early on, Netflix introduced the “Queue.” You could browse the site and make a list of the movies you wanted to see, much like Amazon’s shopping cart. Every time you returned a DVD to Netflix, it would automatically ship out the next title in your queue. The average number of movies in a customer queue was around fifty and went a long way to endearing customers to the service by making it feel personal. “It’s our biggest switching cost,” Hastings would later say, a prime reason users stayed loyal.38 Netflix became a platform to cultivate your individual cinematic tastes.

  Netflix also concentrated on other advantages the web made possible. Copying Amazon’s Recommendation Engine, as well as DVD retail competitor Reel.com’s pioneering Movie Match technology, Netflix developed its CineMatch movie recommendation system. Users were prompted to check out movies they might like based on previous titles delivered from their queue. Netflix invested heavily in this technology, hiring mathematicians and computer scientists to tweak the algorithm to include recommendations based on the habits of subscribers with similar taste. Netflix’s recommendation engine proved so uncannily good at predicting what you might want to watch that, eventually, nearly 70% of the movies that users chose for their queues were recommended by the algorithms.39 This was convenient for Netflix because it allowed for greater inventory and cost controls than a brick-and-mortar store could realize. Whereas nearly three-fourths of total rentals at a Blockbuster were for new releases, at Netflix, seven out of ten DVDs rented by subscribers were titles with release dates older than thirteen weeks.40

  ■

  AS HAD HAPPENED WHEN Amazon first began to challenge the big-box retailers, most people assumed that once Blockbuster turned its attention to online video rentals, Netflix would be crushed. In 2002, Blockbuster spokespeople were already dismissing online rental services as “serving a niche market.”41 But soon enough, the entire rental industry began to feel the pressure of online competition. By 2002, Netflix had attracted 750,000 subscribers, which, while only representing 2% of the video rental market, nevertheless caused sales to slip at the rental chains, especially in areas where Netflix was popular.42 Perhaps two years too late, Blockbuster Online was rolled out as a direct web-and-postal-delivery competitor to Netflix in 2004. As Barnes & Noble had, Blockbuster tried to leverage its physical footprint. You could return movies by mail or at your local store. It also unveiled a “no late fees” program in 2005, which shut Netflix up about that feature, but had the simultaneous effect of costing the company about $600 million in lost revenue.43

  Unlike with Amazon and the booksellers, once Netflix began to eat into the video rental market, the decline of the retail rental industry came quickly. At its peak, Blockbuster had more than 10,000 stores and 50 million members.44 At one time, more people had Blockbuster cards than American Express cards.45 But by 2010, Blockbuster had only 25 million customers and 4,000 remaining stores.46 That same year, Netflix announced that it had attracted 13 million members and was mailing 2 million discs daily in the United States.47 Blockbuster filed for bankruptcy protection on September 23, 2010.

  Store closings, layoffs and bankruptcies are often the markers we use to measure the disruptive effects of ecommerce. There were nearly 25,000 individual video rental store locations at the industry’s height.48 At one point, 60,000 employees wore the blue shirts of Blockbuster.49 Video stores were once one of the most common retail storefronts in America; there was hardly a neighborhood without one. Today, the few video rental stores left are nearly museum pieces. Netflix won not because it eliminated late fees, but, again, because it understood how consumers’ expectations were changing and moved to satisfy those new expectations. Unlimited selection. Instant gratification.

  And Netflix deserves credit for continuing to move in that direction even after it had conquered DVD rentals. As early as 2002, Reed Hastings was telling Wired magazine, “The dream 20 years from now is to have a global entertainment distribution company that provides a unique channel for film producers and studios. . . . In five to ten years, we’ll have some downloadables as well as DVDs. By having both, we’ll offer a full service.”50 He was talking about video on demand. About Netflix becoming a studio and producing its own content. About streaming. All delivered via the Internet.

  “We named the company Netflix for a reason,” Reed Hastings has said on more than one occasion. “We didn’t name it DVDs-by-mail.”51

  13

  A THOUSAND FLOWERS, BLOOMING

  PayPal, AdWords, Google’s IPO and Blogs

  Netflix’s successful IPO on May 23, 2002, was an early sign that the Internet was not over as a wealth-generating machine, and though it was one of the first Internet companies to go public after the bust, it was not the first. That was probably PayPal.

  PayPal began life as Confinity, launched in July of 1999 by Peter Thiel and Max Levchin with the immo
dest proposal of disrupting the global financial system. From the first days of the web, people had wanted to use the Internet to create some form of ecurrency. “As far back as 1995, there were a hundred companies that used cool technologies for moving money and that were going to change the world,” Thiel recalled.1 In the midst of the bubble, there were well-funded digital money schemes like Flooz.com and Beenz.com that did not survive the nuclear winter. PayPal’s crucial insight was that payments in cash could be beamed directly to your virtual person: your email address. By the end of the nineties, everyone had an email address. PayPal simply turned your email address into a virtual bank account routing number. Need to send me $10? Use PayPal to send it to my email address.

  Where the virtual-bank-account-tied-to-your-email-address really found traction was among web users who were already doing a lot of virtual transactions over the web: eBay buyers and sellers. On eBay, 90% of transactions took place via check or money order.2 Credit card merchant accounts cost hundreds, even thousands, of dollars to set up, and were designed for actual businesses. But what if you just wanted to unload your used record collection on eBay? There was no mechanism to take easy payment via credit card for the eBay hobbyist.

  Enter PayPal. Sellers on eBay simply asked buyers to “PayPal” them the payment for a successful auction to their email address. PayPal would withdraw the funds from one, and forward to the other. Among the eBay community, PayPal quickly generated a strong network effect: the more sellers asked to be paid via PayPal, the more buyers were incentivized to sign up for a PayPal account, and vice versa. Just as Hotmail had advertised itself with every email sent, PayPal attracted users with every auction that was settled using its service. PayPal quickly registered 10,000 users only two months after launching, and 100,000 a mere month after that.3

  PayPal had early competition from another company that had neighboring offices in Palo Alto. X.com was founded by a serial entrepreneur named Elon Musk, who had a vision that was just as grandiose as Thiel and Levchin’s: a next-generation suite of banking and financial services that would be entirely virtual. For a while, the two competed fiercely for users, but in March of 2000, X.com and Confinity merged, eventually adopting the PayPal moniker for the combined company.

  PayPal was initially completely free to use, but the service eventually charged sellers 2.9% and 30 cents per transaction—still less than credit card companies charged small merchants, and without any of the overhead or complexity. Very quickly, PayPal discovered how lucrative merely acting as a commercial middleman could be. By the fourth quarter of 2001, PayPal was profitable, thanks to facilitating payment for roughly one-fourth of eBay’s total auctions. After a mere twenty-six months of operation, there were 12.8 million PayPal accounts. It had taken eBay more than four years to reach 10 million accounts.4

  On Friday, February 15, 2002, PayPal went public and enjoyed a 55% first-day pop. The financial press, which had been so instrumental in cheerleading for the bubble, now proved downright hostile to the return of Internet IPOs. “It’s an anachronism—straight out of 1999,” the New York Times quoted a stock analyst as saying. “It’s like we’ve kind of forgotten what got us into this situation in the first place.”5 But doubters were proved wrong a mere five months later when eBay acquired PayPal for $1.5 billion, one of the biggest acquisitions in the wake of the dot-com implosion.

  PayPal showed that the web was still fertile ground for innovation, but perhaps the greater legacy for the company was how it proved to be the finishing school for an entire generation of entrepreneurs who would go on to lead the renaissance of the technology industry. Elon Musk, of course, went on to found Tesla. Peter Thiel became the first major investor in Facebook. Early PayPal employee Jeremy Stoppelman founded Yelp. Max Levchin founded slide. And PayPal alumni had a hand in founding, funding or contributing to the development of so many subsequent companies (LinkedIn, YouTube, Yammer, Palantir, and Square, just to name a few) that folks in technology often refer to a “PayPal Mafia” that runs Silicon Valley today.

  ■

  THE SUCCESSES OF NETFLIX and PayPal were beginning to banish the ghosts of the dot-com bubble, but it wouldn’t be until the undisputed star of the final wave of web startups found its footing that people were willing to believe in the Internet again.

  Google was the one service that had the greatest impact after the nuclear winter, but there was one important trait that Google shared with the departed dot-coms: it wasn’t making very much money. It’s somewhat forgotten now, but Google existed for several years without much of a business plan. The vision Larry and Sergey had sold the venture capitalists involved a three-pronged strategy. First, Google would license its search technology to the major portals. Second, the company would sell its search technology as a product to enterprises. And third, there were some vague promises about selling ads against searches on its own website.

  The young company made major progress toward the first goal by finally convincing some of the portals to use Google’s results on their search pages. The first deal was struck with Net­scape for its Netcenter portal, but the really big coup came when Yahoo was finally convinced to use Google for its search results (previously, a company named Inktomi had been Yahoo’s search partner). The partnership with Yahoo was announced in June of 2000, and was an enormous deal for Google at the time. Part of the arrangement allowed for a “powered by Google” logo to appear on Yahoo’s search pages, thereby introducing the Google brand to millions of mainstream web users. Daily searches served by Google swelled from 18 million a day before the Yahoo deal, to 60 million a day afterward.6 In early 2001, Google would pass the 100-million-searches-per-day milestone, answering 1,000 queries a second.7 Yahoo seemed not to mind that Google was essentially stealing its search audience because at the time it didn’t feel search was a core product. It was still pursuing its portal strategy. Yahoo did, however, purchase a $10 million equity stake in its new partner, thereby tying the two companies closely together in ways that would later become important.

  What Yahoo didn’t know was how important the partnership would prove to be for Google’s overall product. Remember that Google’s algorithms improved in direct relation to the number of searches it performed and the amount of data Google’s computers could hoover up. The flood of queries from Yahoo not only took Google to the next level in terms of search market share, but many Google engineers would later credit the Yahoo traffic for fine-tuning Google’s search engine into its mature state. Google was essentially improving itself on the back of its biggest partner.

  But the problem for Google was that the Yahoo deal simply wasn’t lucrative. The fees that Yahoo coughed up were barely enough to cover the increased processing and bandwidth costs Google incurred to service the traffic. The Yahoo deal taught Google that licensing alone wouldn’t be a big enough home run to build a company around—or at least, not a very big company.

  The second leg of Google’s original strategy was proving little better. Google produced an actual hardware device, known as the Google Search Appliance, which was a rack-mounted box meant to be installed in corporate data centers. It was designed to provide corporations and other organizations with large amounts of data and the ability to organize, index and search that data the same way that Google did with the web. But even though Google continued to produce the Search Appliance until 2017, it never became a breakout hit.

  By the end of 2000, Google was in a bit of a crisis. With monthly expenses of more than $500,000, the $25 million from Kleiner Perkins and Sequoia was starting to run low, as Google launched international versions of its website and continued to hire, taking total headcount past 100.8 “There was a period where things were looking pretty bleak,” Google board member and investor Mike Moritz admitted later. “We were burning cash, and the enterprise was rejecting us. The big licenses were very hard to negotiate.”9 And since Google had yet to earn a dime on the 70 million daily searches it was getting on its own site, by January 2001, Google’s out-
of-control growth was actually a problem. While the service was becoming so popular that its very name was becoming a common verb, “There was genuine concern (at the board level) about where the revenues were going to come from,” says early Google investor Ram Shriram. To make matters worse, it appeared to Google’s venture backers that the company’s founders were reneging on their commitment to bring in a “grownup” CEO. If Page and Brin didn’t recruit someone who could turn Google into a real company with real prospects to generate cash, there were rumblings that either Kleiner or Sequoia (or both) might pull out of the investment.

  Of course, advertising, the third leg of Google’s theoretical business model, was still an option, but in spring 2001, the existing advertising model of throwing banner ads at the top of every web page had imploded. Web advertising in general was in a deep freeze, the overall online ad market plunging to $6 billion in 2002, down from $8.2 billion in 2000. All the surviving portals were suffering because of this state of affairs.10 In the midst of the freefall in its stock price, Google’s erstwhile partner Yahoo was forced to lower its revenue guidance to Wall Street by 25% twice in a single quarter as the dot-coms went bankrupt and advertisers ponied up 50% less for online ads.11

  Google had never really experimented with ads, because the company’s founders were originally firmly against the idea. In their 1998 academic paper introducing Backrub/PageRank, Page and Brin had attacked the very notion of search companies relying on advertising to generate revenue because it made them “inherently biased towards the advertisers and away from the needs of the consumers.”12 In other words, ads guaranteed bad search results.

  But at this very moment of crisis, a revolution in online advertising was taking place that would ultimately prove to be Google’s salvation.