Viral Loop
[ From Facebook to Twitter, How Today’s Smartest Businesses Grow Themselves ]
Adam L. Penenberg
To the Penengirls,
Charlotte, Lila, and Sophie
Contents
Prologue
An Insanely Viral Scheme
How the Guys from Hot or Not Rode a Simple Idea to a Fortune
Introduction
Viral President
Positive Feedback Loop, Spreadable Concepts, and the Three Categories of Viral Expansion Loops
I
Viral Businesses
1 Tupperware and Ponzi Schemes—the Original Viral Models
Party Plans, Referral Networks, and Sizzlemanship
2 The First Online Viral Expansion Loop
Mosaic, Netscape, Network Effects, and the Spark that Touched Off the Internet Boom
3 The Spreadable Product as New Business Paradigm
Viral Plain, Ning’s Double Viral Loop, Your Digital Self, and Crushing Crushlink
II
Viral Marketing
4 The Perpetual Viral Advertisement
Viral Tag: P.S. I Love You. Get Your Free Email at Hotmail
5 When the Audience Decides What’s Good
Collective Curation; Boom, Bust, and Beyond, and the Viral Reality Show of Your Life
6 Viral Video as Marketing Strategy (Psst. Pass It On…)
Letting Go of Your Brand, Joke Cycles, and Not Acting Like Some Guy in His Mid-40s Trying to Be a Hipster
III
Viral Networks
7 eBay and the Viral Growth Conundrum
Faster, Pussycat. Scale! Scale! Scale!
8 PayPal: The First Stackable Network
Viral Synergy, Greedy Inducements, Scalable Fraud, and Battle of the Network All Stars
9 Flickr, YouTube, MySpace
Spreadable Spreadables, Stackable Stackables, and the Point of Nondisplacement
10 Tweaking the Viral Coefficient
Bebo, Patterns in Virality, and Critical Mass
11 Viral Clusters
Facebook, the Social Graph, and Thingamajigs
12 The Search for a New Ad Unit
Death to the Traditional Banner Advertisement, the Arms Race Between Marketers and Consumers, and Time (Not Clicks)
Epilogue
The Viral Loop Company Index
Acknowledgments
Notes
Searchable Terms
Copyright
Prologue
An Insanely Viral Scheme
How the Guys from Hot or Not Rode a Simple Idea to a Fortune
Two Heinekens into a lazy Tuesday afternoon, James Hong, a twenty-seven-year-old dot-com refugee from Mountain View, California, was listening to his roommate, Jim Young, a Berkeley graduate student in electrical engineering, wax on about a woman he had spotted at a party the previous weekend. Young, also twenty-seven, insisted she was a “perfect 10.” Hong didn’t believe him. He knew his roommate had a thing for “goth,” while Hong’s own tastes were more Abercrombie & Fitch. What the world needed, they agreed, was a metric to reliably rate someone’s looks.
The two out-of-pocket entrepreneurs weren’t just in a frat-like buzz obsessing over women who likely wouldn’t give them a second look. This exchange was the logical outgrowth of a broader discussion they had been having about Web services, something that had long been on Hong’s mind. At the time, October 2000, Web services focused on business-to-business applications, but Hong wondered about a consumer play—a product that would appeal to regular people. A year earlier, he and Young had toyed with the idea of creating a dating site with collaborative filtering to better cluster people: users would scan a series of pictures, then express an opinion as to how attractive the person was. This, they reasoned, could increase the efficiency of dating, although it didn’t end up spawning an actual business. Now he and his roommate were ginning up more ideas for a start-up.
Young’s “perfect ten” comment got Hong thinking. What if you could rate someone’s looks by opening it up to a vote? Most people believed they were above average, which was, statistically speaking, impossible. Hong, an athletic extrovert with a buzz cut and zingy wit, rated himself a 7. And his skinny partner with the modish, foppish hair? About the same, but who really knew? In the way the value of a house or used car was whatever someone was willing to pay—that is, whatever the market would bear—a person’s attractiveness could be based on whatever a large community of people judged it to be. For it to work, all they would need were pictures and a grid from 1 to 10 for site visitors to click on. Then they would crunch the aggregate data and spit out the results. Hong didn’t know many people who had a whiteboard in their living room, but he and Young did, and they quickly drew up plans. Burned out on his dissertation, Young figured it could be an entertaining diversion and suggested they build a simple website instead of a vastly more complicated Web service.
Hong scoured the Web for candid shots—he wanted real people, not Sports Illustrated swimsuit models—while Young got cracking on the digital infrastructure, cobbling it together with Linux code. A couple of days later, while visiting his parents, Hong was tinkering with the site when in walked his father, a Taiwanese-born engineer who wore pocket protectors and had given him a slide rule for his thirteenth birthday. If he found out what his unemployed son was working on, he would shake his head and mutter that he was wasting his education. Hong told him it was all Young’s idea.
Peering over his son’s shoulder, his father looked at the photo on the screen. “Oh, she’s hot. She’s an eight,” he said. Hong could scarcely believe his ears. He was half convinced his father had engaged in sex only three times in his life—when he, his brother, and his sister were conceived. Without prompting, his father clicked on another photo. “Not so hot,” he clucked. Then click…. Click, click, click, until he pored through the entire cache of photos—forty in all.
His father was hooked. Because of the way Hong and his partner designed the site, a user couldn’t find out the poll results until he voted and the next page loaded. They purposely placed the aggregate score on the left side over a smaller version of the picture while the next picture, significantly larger, sat front and center. The idea was to train a user’s eyes to flow from the middle to the left. Because there was always one more photo to be judged, a user would feel compelled to cast another vote. It was like the old Lay’s potato chip ad: “Betcha can’t eat just one.” The interface demanded engagement, and this in turn made it sticky. If a site retained a visitor’s interest, it had the potential to be popular and spread by powerful word-of-mouth endorsement.
When his father was out of earshot, Hong phoned his roommate. “Dude, we have to launch this thing right now!”
[ GOING VIRAL… ]
On Monday, October 9, 2000, Am I Hot or Not went live, not that anyone but its two founders knew about it. Hong emailed forty-two friends (partly, he said, because forty-two was the answer to “life, the universe and everything” in the cult sci-fi classic The Hitchhiker’s Guide to the Galaxy). “Here’s a website that Jim and I made—be nice,” he wrote and inserted a link to their own pictures. Soon after, visitors pinged the site, quickly spreading well beyond Hong’s forty-two friends. It didn’t take an epidemiologist to recognize a viral spread in the making. To test it, Hong slipped on a pair of Rollerblades and skated across the street to an office park where Tellme, a software call center and services provider, was located. He glided over to an office worker in the parking lot.
“Hey, have you seen this website where you can rate people if they are hot or not?” Hong asked.
He hadn’t. Hong told him and bladed home. Inside ten minutes Tellme’s IP address appeared in the site logs. Hong monitored the man
’s activities as he clicked through the photos. Suddenly more Tellme IP addresses popped up. Clearly the man had shared the link with coworkers, who then passed it to others. A drizzle turned into a full-blown storm, with the hits coming in fast and furious from all over the country. Traffic swarmed their borrowed server; the site slowed to a crawl. It took half a minute for each page to load.
Hong was hosting the site on his brother’s machine, which sat in a data center. The more bandwidth they consumed, the more they would have to pay. Hong knew this would quickly become unsustainable. By the end of the first day thirty-seven thousand unique visitors had found their way to the site and two hundred people had uploaded photos. On day two, they broke one hundred thousand in addition to the mass of returning visitors from the first day. At the current run rate, Hong estimated that Am I Hot or Not would cost $150,000 in bandwidth in their first year. The more successful they were, the more likely they could go bankrupt. Hong, a self-described pop culture junkie with an MBA from Berkeley, was broke, $60,000 in debt, and living rent-free in his roommate’s parents’ house. He had no idea how the site would make money and frankly didn’t care. All he knew was that he had stumbled into an insanely viral scheme, but he had to figure out how to keep it going before the levees gave way and it drowned in its own popularity.
It would only get worse before it could get better. Salon.com was planning an article for the following day. The reporter contacted Hong after a friend had emailed her the link with the comment, “I think this is quite viral.” Hong begged her to push the story back a few days until he could solve their bandwidth woes. She refused. (Slow news day, she explained.) In her essay, she called Hong and Young’s brainchild “nothing more than a virtual meat market,” “indescribably horrible…and yet utterly addictive,” “fashion police with a twist,” a site that “manages to throw the whole idealistic notion of ‘beauty comes from within’ right out the window in about three seconds flat.” She uploaded a five-year-old photo of herself, claiming unconvincingly that she didn’t want to know how she would fare. Her mix of opprobrium and self-loathing found a hungry audience. Readers slammed Am I Hot or Not’s server. Now Hong and Young were fielding two viral storms.
[ TOO BIG, TOO FAST ]
Survival depended on solving their scaling problem. Step one was to confront the colossal amount of bandwidth the photos soaked up. Around midnight, while commiserating with Young at the drive-thru of an In-N-Out Burger, Hong had an idea. “We don’t need to host the pictures,” he said. “We’ll let Yahoo do it!” Based on his analysis of the server logs, Hong believed twenty-five new pictures a day would be enough to attract fresh traffic to the site—eventually people got bored and moved on. They would effectively cap the number of new pictures. While they were at it, they would direct users to post their photos on Yahoo GeoCities and submit the URLs (the Web addresses) to Am I Hot or Not. By 3:30 a.m. Hong and Young had transferred their entire collection to Yahoo.
Then they addressed the next challenge: offloading the site from their clotted server and out of the data center. At 4:00 a.m. they drove to Berkeley where Young had an office. They pulled the plug and reinstalled the site on a cheap 400-megahertz Celeron PC that Hong had gotten free for opening an eTrade account. So no one would turn it off, he popped the top off a case of pushpins and mounted it over the PC’s power switch. Then they secreted it under Young’s desk, stashed in a corner and buried by books. To the uninitiated it looked like a pile of stuff with an Ethernet cable poking out.
It was five in the morning and Am I Hot or Not had been down for two hours. They wondered if the outage might have killed it. As soon as they switched it back on, they got their answer. Bam bam bam, IP addresses pounded into their logs. Three hours later Young’s adviser, the dean of the Berkeley engineering department, informed them the university’s information technology department was up in arms. Their single PC was weighing down the entire network. Amused by their plight, the professor promised to cover for them, but they had twenty-four hours to set up shop elsewhere.
Beg, borrow, steal, whatever it took. Hong searched for a new home, settling on Rackspace, a Web-hosting firm. Although he and Young had no money, they were getting heaps of press. The Guardian, the New York Times, and news outlets from around the world found Am I Hot or Not irresistible. Hong cultivated the media because he knew their concept would be easy to steal. The more stories, the more spikes in traffic, the harder it would be to dislodge them when the inevitable copycat arrived on the scene. He cold-called the head of Rackspace business development to propose a trade. “I know you guys want to go public and it’s great to get your name out,” Hong said. “Your whole value proposition is that you can help companies scale by outsourcing. If you can help us, we will have all these upcoming interviews, and we can be a poster child for you.” The Rackspace executive agreed, and every day that week Hong called to request more machines. By its eighth day the site was fielding 1.8 million page views per day, and both Hong and Young, who had slept eight hours over those eight days, were literally shaking.
[ SHOW ME THE MONEY ]
Now that Hong and Young had, for the time being, solved the site’s scaling challenges, they needed to figure out how to monetize it. With the kind of traffic they attracted—within six weeks of launch the site had 3 million page views a day and posted more than three thousand photos—advertising made the most sense. These were the days predating automated ad servers like Google AdSense, which meant they had to enlist an online advertising network. But a number of users had been uploading nude photos, while pornographers were tricking users into passing on their emails so they could spam them. Hong knew that they couldn’t attract advertisers if there was objectionable material, so he and Young came up with the motto “Fun, Clean and Real,” and issued basic rules: no celebrities, minors, models, or porn; no group photos, ads, or anything with contact information such as email addresses or phone numbers.
They created a community monitoring system: users could click a link under an inappropriate picture and, based on an algorithm, anything that was tagged too often was deleted. Ready to approach ad networks, Hong emailed the founder of DoubleClick, who replied that when he visited the site, the first picture he encountered was of a naked woman. What we need, Hong told Young, was a more intensive moderation system: someone to approve the photos before they posted. Originally he appointed his parents, who were retired and had plenty of free time. After a few days he asked how it was going.
“Oh, it’s very interesting,” his father replied. “Mom saw a picture of a guy and a girl and another girl and they were doing…”
“Dude,” Hong told his partner, “my parents can’t do this anymore. They’re looking at porn all day.”
They turned to their community to act as moderators. Each applicant was required to pen an essay. Those chosen to moderate were instructed to reject inappropriate pictures, ads, or anything with someone’s contact information. With hundreds of users trolling the site, they could filter what was posted. After this went into effect, it didn’t take long to land an advertising network. Although the click-through rates were low—on the order of 0.2 percent—the immense number of page views still made it a moneymaking proposition. Within two months they had counted 7 million page views a day and had collected 130,000 photos. By year’s end, three months after the site’s launch, they had broken into Nielsen’s Top 25 advertising domains on the Web and had generated $100,000 in advertising revenue.
But they were nowhere near Easy Street. Hong received a cease-and-desist letter after Howard Stern, on his radio show, mistakenly called their site Am I Hot, the name of a far racier site pushing skin and cleavage. Hong had done his due diligence and looked for similar-sounding domains before launch. It wasn’t his fault Stern had made a mistake, but he wanted to avoid a battle over the name. He agreed to change it to Hot or Not, and they promised to redirect traffic to the new domain for three months. Hong made good use of all the press coverage to rebrand the site,
and traffic barely hiccupped. In the meantime, he chopped their overhead. He approached Ofoto about an affiliate deal, telling an executive how he had dispatched people to Yahoo to upload photos. Hong could just as easily send them to Ofoto, and by hosting them, Ofoto would have the lead in offering additional services. The agreement he struck meant that Hong had moved something from costing money to making it free to actually generating a profit.
[ REVENUE STREAMS ]
Just when things seemed to be humming along, the Web ad market stumbled, a victim of the dot-com bust, and advertising rates dropped. “Can we charge for anything?” Hong wondered. The answer was staring him right in the face. After they had launched the community monitoring system to stamp out pornography, they received emails from users complaining that they could no longer meet people through the site. Since it made sense to stoke the community, Hong and Young set up Meet Me, which allowed active users to commingle online. It was a hedge against porn operators, since a member could no longer simply post an ad and wait for email to roll in. A user had to engage in the community.
In April 2001, they instituted a $6 per month fee to join Meet Me, figuring it was cheap enough to qualify as an impulse buy. In its first month it generated $25,000 in revenue; by year’s end it had brought in $600,000. Their success spurred them to work even harder; most nights neither got much sleep. They turned it into a race: could Hong bring in people faster than Young’s system could handle? Hong’s job was to create a bottleneck and Young’s was to clear it. That first half-year felt like it lasted a day. Practically every waking moment they considered ways to make the site run faster and better. They were featured in People, Time, and Newsweek and by the end of the year they were on Entertainment Weekly’s It List. They were profiled in the New Yorker. Hong became the first in his MBA class to make it into the Wall Street Journal.
Viral Loop Page 1