Harry Jho read the email but did not believe it. A human from YouTube? He had long since given up on speaking to an actual human from the company. Once at an event an employee handed him a business card, a promising sign. He looked down to see an email, support@google.com, and no name.
Jho worked on Wall Street as a securities lawyer for Bank of America. With his wife, Sona, he also ran Mother Goose Club, a YouTube channel where colorful animals sang nursery rhymes. After the couple taught English in Korea in the 1990s, Harry moved into banking, and Sona, who held a Harvard graduate degree in education, produced shows for public-access stations and the local PBS. The Jhos, who were Korean American, had two young children and noticed how few faces on kids TV looked like theirs. As educators they saw television’s pedagogical flaws. To learn, kids should see lips move, but Barney’s mouth never did. Baby Einstein mostly showed toys. So they started Mother Goose Club, investing in a studio and hiring a diverse set of actors to don animal costumes and sing “The Itsy Bitsy Spider” and “Hickory Dickory Dock.” It was like Teletubbies, only less trippy and inane. The Jhos planned to sell DVDs to parents, ginning up interest for a possible TV show. YouTube offered a convenient place to store clips, and in 2008 Harry Jho started an account there, not thinking much of it.
Two years in, though, he started checking the account’s numbers after leaving work. One thousand views. He checked the next day. Ten thousand. He couldn’t find many other videos for kids on YouTube. Maybe, instead of television, he thought, we can be the first to do this. At least before Nickelodeon took the field over.
It was the spring of 2011 when the human from YouTube emailed, extending an invitation to Google’s Manhattan office. The staffer showed Jho designs for the site redesign and shared some tips. Finally, Jho asked the question he was itching to ask, “Why did you call us?”
“You might be the biggest YouTuber in New York,” the staffer replied. This was news to Jho.
Harry and Sona Jho, soft-spoken professionals who wore glasses and sensible clothes, looked more like PTA parents than YouTube stars. But the iPad had debuted a year before, a handy device for frazzled parents of toddlers. Soon enough, YouTube would add an auto-play function that mechanically teed up one video after another. After the Google meeting the Jhos saw even more traffic on their site. YouTube let them into its ads program. A year later YouTube switched to prioritize watch time, and very quickly Mother Goose Club got company. It began with BluCollection, an anonymous account that only posted videos of a man’s hands scooting toy figurines across a floor. The Jhos watched as these clips appeared in the sidebar next to theirs, one by one. Similar videos followed, carpeting the entire sidebar. Then they saw these videos take over YouTube.
Parents and bureaucrats have always cared what the kids are watching. In the 1970s a federation of advocates and educators who had helped put Sesame Street on air pushed for tighter regulation of commercial activity on children’s TV, worried that kids could not distinguish programs from ads. Saturday morning cartoons were forbidden to pitch products. A 1990 law, the so-called Kidvid rules, went further, requiring broadcasters reaching children to air certain hours of educational programs and place time limits on how often commercials were aired. Networks tried bending the rules, but regulators held the threat of a license removal.
Then cyberspace arrived. An alarmist 1995 Time cover showed a blond boy at a keyboard, his eyes lit in horror-schlock glow, above the menacing word “Cyberporn.” “When kids are plugged in,” Time asked, “will they be exposed to the seamiest side of human sexuality?” Lawmakers governing the modern internet were so obsessed with threats of sex and violence that they ignored other concerns, such as the balance of educational content in media and the potential developmental impacts of unchecked consumerism. Privacy activists, worried about the creeping panopticon of web trackers like Google’s cookies, pushed Congress to regulate children’s browsing. Websites were openly inviting kids to share the sorts of personal details marketers valued. “Good citizens of the Web,” read a promotional site for the movie Batman Forever, “help Commissioner Gordon with the Gotham Census.” A minor victory for activists came in 1998 with the Children’s Online Privacy Protection Act (COPPA), which prohibited websites from collecting information from minors under thirteen for use in targeted advertising. But the law gave enforcement to a different agency (FTC) from the one overseeing television (FCC) and had none of the rules concerning educational programming or commercials that TV had. Old media also had rules about talent. Certain states, like California, restricted the hours child actors could appear in TV or movies and set safeguards for their earnings. The internet didn’t.
But kids were clearly heading online, and the massive kid’s entertainment complex was eagerly coming with them.
YouTube had seen this migration early. Before the Google acquisition, YouTube vice president Kevin Donahue, a former Cartoon Network producer, pitched YouTube’s founders on a kids’ version of their site. They directed him to the company lawyer, who shot down the idea. COPPA required websites to do acrobatics to pull something like this off, and YouTube was then so thinly staffed that it needed all its legal resources for copyright issues. YouTube required uploaders to check a box stating that they were over thirteen. The site’s terms of service declared it was only for people above that age, and so, on paper, it was.
Google had arrived at similar conclusions. Yahoo, its old archnemesis, once ran a kids’ website (Yahooligans), and a few times a year someone at Google would propose a kid-friendly version of Google search. The idea never made it past the sticking point: How do we decide what is kid-friendly? A few parents on staff did spot nursery rhymes, ABCs, and toy clips clearly made for toddlers—even before BluCollection, the Jhos were not alone—and fretted about their quality. “Kind of total crap,” recalled one mother there.
Any proposals for cleaning the crap had to go through Hunter Walk, the product manager who once tried to scrub the comments section. Walk embraced YouTube’s youth culture cachet and knew the kids’ world; he interned at Mattel during business school and once worked at a children’s bookstore. Yet when colleagues pitched a kid-friendly YouTube, he said no. YouTube simply didn’t have enough premium kids’ material to make this anything but a lousy version of cable, he reasoned. (Viacom owned Nickelodeon, so YouTube wasn’t getting its shows.) YouTube had some TV classics—the company played a Sesame Street clip at the second VidCon—but called these “nostalgic,” not kids’ content. Staff knew juvenilia like Fred were big hits but had convinced themselves that Fred’s audience was mostly teens, bored with TV, and that anyone under thirteen watched with adult supervision, like the site’s fine print said they must.
But enough stuff clearly made for grade-schoolers or even preschoolers was piling onto the site that some at YouTube felt they had to do something. Walk took a break from YouTube in early 2011, after a painful repetitive strain injury to his wrists, and he returned that fall with a half-baked idea: “YouTube for Good”—a patchwork initiative to improve features for activists, nonprofits, and schools. Most school districts blocked YouTube, wary of its internet free-for-all. But educational videos were blossoming: the Green brothers had their shows, and Salman Khan, a hedge fund analyst, had started uploading mathematics lessons as the Khan Academy channel, part of a Silicon Valley wave devoted to upending higher education. So Walk led an effort to give these creators a name (EduTubers), tools, and attention. He lobbied schools and politicians on YouTube’s benefits for students. If schools let YouTube in, the company reasoned, more quality kids’ material could flourish.
Other parts of Walk’s YouTube for Good project were designed to make the site look more respectable. Coders built a tool to blur faces in videos so that protesters could upload footage with less fear of repercussions. One team worked on a feature to thread tweets and other online chatter about YouTube videos onto the site and its home page, giving it the feel of a news site. Earlier, to sell the channels idea, Walk had christened YouTube “the living room for the world” (a screw-you to TV). Now he framed it as the world’s classroom and town square.
Still, as the company tried to prop up its educational and wholesome sides, it was blindsided by a strange beast born within its walls, charging hard in another direction.
The beast, as always, appeared first in the data. Some of YouTube’s coolhunters, the old home-page curators, had settled into a marketing unit called YouTube Trends. Each week the unit sent around the “What’s Trending” report, a roundup of the site’s emerging fads. Business staff also monitored a chart of the site’s top one hundred ad earners. One odd channel started landing in the trending reports and soaring up the earnings chart.
DisneyCollectorBR: “Giant Princess Kinder Surprise Eggs Disney Frozen Elsa Anna Minnie Mickey Play-Doh Huevos Sorpresa.” March 24, 2014. 14:27.
“Hey, guys, Disney Collector here.” We see two dozen toy eggs of various sizes, bearing recognizable figures on their wrappers from various children’s entertainment franchises. The voice is feminine, lilting, slightly accented. We don’t see her face. The camera holds tight on her hands, fingernails painted black with delicate little Disney princess portraits on top. She announces each egg methodically. “Mickey Mouuuuse.” She begins to unwrap the toy eggs, first peeling back the foil casing, a soft, crisp sound. Then the chocolaty layer, a satisfying crackle. Then the tiny plastic capsule holding a toy, a treasure. Then another.
YouTube had never seen a force like DisneyCollectorBR. By that summer the channel’s most popular video, a four-minute unwrapping, had 90 million views. Overall its videos were watched a whopping 2.4 billion times. Tubefilter, an online Billboard for YouTube, placed DisneyCollectorBR as the third most viewed channel behind PewDiePie and Katy Perry. Soon the channel claimed gold. A research firm estimated it raked in as much as $13 million a year from YouTube ads. The videos contained something uncanny and new, tapping into neurons in children’s brains in a way that few fully understood. Certainly, no one at YouTube did. “I think she disappears in the mind of the children,” one marketer told a reporter. Unboxing videos had begun years before in tech reviewer circles, with footage treating iPods and smartphones as fetish items. Now the Kinder Egg, a marginal product developed in Italy, took on totemic significance. U.S. officials banned it, citing the small toys inside as choking hazards, so YouTubers chasing DisneyCollectorBR’s trend started buying these eggs on eBay, like contraband.
At Maker staff watched this fad closely and dubbed the sites “hands channels.” Other YouTubers preferred “the faceless ones.” Like earlier YouTube hits, these channels sought views using Google’s central corridor, search. Look at the mishmash written beneath DisneyCollectorBR’s video:
Princess egg, frozen eggs, Scooby doo, hello kitty, angry birds, sofia the first, winnie the pooh, toy story, playdoh surprise
A keyword soup. Toy unboxing titles followed similar logic. “Choco Toys Surprise Mashems & Fashems DC Marvel Avengers Batman Hulk IRON MAN”; “Disney Baby Pop-up Pals Easter Eggs SURPRISE Mickey Goofy Donald Pluto Dumbo.” These titles weren’t made for the intended viewer or even their parents. These were made for algorithms—for machines to scrape and absorb. Disney, like many media giants, refused to put prized material on YouTube. So when people typed into the search bar “Frozen Elsa” or “Marvel Avengers”—Disney had bought Marvel in 2009—machines showed them the faceless ones.
One person on YouTube’s Hollywood team watched the trend in dismay, thinking, The algorithm is really effed up. More sinister stuff prowled on YouTube’s outskirts. A public relations staffer once looked up a sampling of videos viewers had flagged for YouTube’s moderation unit. There was Bugs Bunny, a clip easily appealing to young kids, edited into a violent first-person shooter game. “Maybe it wouldn’t traumatize kids,” the staffer decided. “But it’s really weird.” It was relatively benign compared with the weird to come.
Most faceless channels, like DisneyCollectorBR, were also anonymous. Early stars used pseudonymous handles but typically sought fame with real names or at least faces. They had managers, agents, hangers-on, Twitter profiles. To earn ad money, YouTubers had to provide the company with a legal name and an email address, but YouTube liked to keep this information separate from its staff for security reasons. YouTube faced an unprecedented situation with DisneyCollectorBR: the company knew next to nothing about its most popular channel.
A human at YouTube now called Harry Jho with a different question: “Do you know who they are?”
Jho had never quit his banking job even after Google’s money had begun flowing in, because it never flowed steadily enough. Some months, during summers or holidays, the Jhos’ channel made $700,000 from YouTube ads, a huge haul. But at other times it dropped to $150,000. How could they hire a big staff and ensure steady salaries? If YouTube were their sole income, “we would have gone crazy from the stress,” Jho recalled.
A Maker representative called him once his channel took off. Jho was interested, until the second and third calls, each from a new Maker recruiter, and Jho realized they were simply dialing down a list. He went to YouTube events—VidCon and Kidscreen Summit, a children’s entertainment equivalent—and asked other creators what value they saw from joining these MCNs. Not much. YouTube had become more welcoming, but the company didn’t distinguish his programming from that of others, or at least its machines didn’t. Once, the Mother Goose Club YouTube page was overtaken with promotions for a new horror film, an Exorcist spin-off. Right beside “Skip to My Lou” was a thumbnail of a demon-possessed girl shrieking. We’re a kid’s channel, Jho thought. No one wants to see that. He tried in vain to complain to YouTube. Eventually, he found a fix: if he bought ads for his own channel to run on YouTube pages, not only did the Exorcist disappear, but his traffic shot up.
By 2014, Kinder Eggs had overtaken YouTube. The formula to get into the related videos sidebar—and, thus, to get in front of kids—looked clear. The Jhos held a meeting in their Manhattan office. They looked at the columns of bright, keyword-stuffed videos from DisneyCollectorBR and its countless imitators. “It’s really cheap to make these videos,” Jho observed. “We could set up a room. Go buy these toys for a couple thousand bucks.” They looked at the columns again.
Finally, a friend who was there in the office piped in. “This is just like porn,” he said. “This is toy porn.”
They dropped the proposal.
Something else happened as the faceless ones spread on YouTube: more people at the company started having children, and those with older kids heard them ask incessantly for their own screens.
Silicon Valley had long had a widespread, ironic parenting philosophy rooted in avoiding its own inventions. Steve Jobs reportedly limited his children’s use of technology. Like Jobs, staff at YouTube spent hours at work reviewing code and business plans all geared to maximize time spent on YouTube and then went home and told their kids to get off YouTube. The site’s addictive appeal for sponge-brained preadolescents was obvious. Some employees felt as if they were working at a cigarette company. YouTube’s chiefs, loyal Googlers, attempted to measure this problem and turn it into a metric.
They invented categories: “Delicious” and “Nutritious.” Much of YouTube mirrored the old scornful derision of TV as “bubble gum for the eyes”—colorful, sugary, craved. But plenty of videos were educational and healthy, too, YouTube thought. (It surely had more of those than TV did, if you added the hours up.) Delicious videos certainly improved watch time, although some staff expressed concern that this type of viewing could be fleeting. After gorging on snacks, don’t you feel guilty? What if people felt the same way after watching hours of unboxing? What if parents, after seeing kids glued to the faceless ones, took away the candy? “We’re only going to be successful long-term if people are happy at the end of their sessions,” Mehrotra told staff during a meeting in San Bruno, “not just at the beginning and middle.” (After Walk stepped away, Mehrotra had taken over product as well as engineering.) YouTube gauged how long people watched, but staff wanted something more qualitative to measure Nutritious videos. The YouTube for Good team created a survey to run at the end of some videos. It showed little checkboxes and asked, “What is a better use of an hour of your time?”
Reading a book
Going to the gym
Watching television
Watching YouTube
YouTube wasn’t TV; it couldn’t give Nutritious videos a prime-time slot. Instead, Walk proposed assigning them a “goodness score,” granting educational footage from creators like Khan Academy and the Green brothers more weight in YouTube’s search and discovery system. In meetings and internal correspondence YouTube referred to this as adding “broccoli” to the site (or sometimes “chocolate-covered broccoli”). Some drafted broccoli OKRs. The Torso division, which managed its ever-sprawling creator class, drew plans to get 30 percent of watch time from Nutritious videos. Coders working on YouTube search and ads all discussed the effort.
Then, in a fateful twist, these discussions petered out. No company-wide objectives and key results were set.
The same sticking point that had hobbled Google’s kid search reappeared: What exactly is Nutritious? How do we decide? Can we program quality into algorithms? Should we? A few working on YouTube’s education projects pivoted to developing a special app for kids. Walk left the company. No metrics were established. “If you can’t figure out how to measure it,” recalled one executive there, “you just pretend it doesn’t exist.”
Besides, bigger, uglier brawls were starting to demand more of everyone’s time in San Bruno—brawls outside Google and within.