by Mark Bergen · 5 Sep 2022 · 642pp · 141,888 words
one goal above others. * * * • • • Cristos Goodrow had sent his email right as YouTube’s leaders convened. For maximum attention, he sent it to all of them and gave it a compelling subject line: “Watch time, and only watch time.” Goodrow came to YouTube after two decades programming software for companies across Silicon Valley and for Google
…
gridlock and lack of clear marching orders. In his prognosis he proposed rewiring YouTube’s machines to favor just one outcome: how long people stayed with videos. “All other things being equal, our goal is to increase watch time,” he wrote in the email. Goodrow liked to debate this idea with fellow
…
rising nemesis, Facebook, flourished not just by racking up accounts but by keeping people engaged. TV certainly did. So YouTube would keep people engaged by promoting videos that racked up the most watch time. Will it make the boat go faster? Then do it! Now they just needed a big, hairy, audacious goal
…
?” he asked. “When could we reasonably do that?” * * * • • • Mehrotra announced the new OKR the following year at YouTube’s annual leadership summit in Los Angeles: YouTube would work to get one billion hours of watch time every day within four years. “Look, I know what you’re all thinking,” Mehrotra began. Impossible. Once trained
…
unit plotted trajectories on a chart, which they called their “Break the Internet Graph.” Goodrow’s coders began rewiring YouTube’s search and recommendation system to promote videos that generated the best watch time, not the most views. Only videos that made their boat go faster. One team was not so enthusiastic. Bing
…
email in early 2012 from Mehrotra inviting him to a room in San Bruno nicknamed the product den. Before setting its audacious goal, YouTube wanted to test prioritizing watch time in its algorithm. Inside the den, Chen was told he had five days to prepare an explanation of the planned change for creators
…
“Your algorithm has a bug.” O’Brien was responsible for YouTube’s search product and neatly fit Google’s product manager mold—a fast talker with software know-how and better people skills than coders. Once YouTube switched its system to favor watch time, some staff were not prepared for the immediate upheaval it
…
the clip, which had been selected to run ads, nearly halved. It didn’t help that no one had updated the“analytics” web dashboard for YouTubers to show “watch time” as a figure. They saw “views” and saw those views plummeting but had no idea why. The company had to convince
…
the redesign that drove viewers away from hitmakers like him. Months later, when YouTube moved to favor watch time, that trend worsened. Penna’s management company, Big Frame, tracked stats for several YouTubers and watched them all drop practically overnight. YouTube’s blog post introducing the switch had less than helpful advice: How can you
…
be this now.” Really? How do you know? No one really did. But in a few months, the main lever became clear: only watch time. When he began on YouTube, Wong carefully studied the science of virality, testing the right ingredients. Now its formula felt painfully simple. “Okay,” he realized. “We’re just
…
Disneyland, and Wong saw swarms of attendees all start to hold cameras at arm’s length, pointed back at themselves. * * * • • • No YouTube genre benefited from the switch to favor watch time more than “Let’s Play.” That was the name given to footage of video gamers filming themselves playing popular or deeply bizarre
…
had raised $1.5 million in venture capital. But many of Maker’s channels had scripted, costly productions, and YouTube’s ad intake didn’t always cover expenses. Once YouTube switched to watch time, the economic tilt became impossible to ignore. Gamers filmed themselves playing and talking. Maybe they edited a bit, added some
…
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
…
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
…
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
…
movies, a foil to Apple’s iTunes. Rubin’s coders also controlled YouTube’s app on Android phones. Several directors at YouTube felt that it should run Google’s music service instead—music videos, after the watch-time transition, were exploding—and that YouTube should control its own app. They pushed Kamangar, who usually avoided
…
to practice, fail and laugh at yourself,” a fashion observer explained in The New York Times the year Nilsen debuted. After the watch-time transition, beauty gurus shot up in YouTube’s charts. Nilsen could post ten- or fifteen-minute-long videos with relatively little editing. She expanded beyond makeup tips and skin
…
the concept certainly did. More viewers moved from one recommended clip to the next, bringing in more watch time and, just as critically, more data. At San Bruno, YouTube staff rarely watched videos, but they watched video data constantly. In particular they paid attention to the seesaw of data on ads and viewership
…
told his coders he was willing to take a 1 percent drop in watch time for a 2 percent increase in ads, but nothing more. Engineers ran tests for Dallas, tweaking experiences for certain viewers without telling them. At YouTube Stats, a meeting Mehrotra held every Friday, they presented the befuddling results:
…
the machines found a way to show more ads and improve watch time. “How can it possibly be positive on both?” Mehrotra asked. “No idea,” an
…
met with Google’s networking staff, which invited her to do so: they were freaked out about the strain YouTube’s hefty watch-time goal was placing on company servers and wanted to curb the plans to relieve stress on bandwidth. There’s no evidence anyone warned her about
…
2005, mixing regular lefty jabs at the press and politicians with clickable tabloid fare. Uygur saw few conservative shock jocks on YouTube, until around the watch-time transition, when they “started popping up all over the place.” Many popped up with videos mocking Uygur’s show or his name; tagging footage with “
…
careful not to overuse ugly slurs or call for outright violence, the kind of invective YouTube removed. If the Brain network was set to maximize watch time, which it was, those sorts of videos might perform very well. YouTube had begun to filter videos promoting Islamist terror, restricting certain clips by age or deleting
…
produce twenty videos a week across her channels, a new kind of juggling. And then YouTube’s second generation and all the demanding social apps arrived. Kay adapted to YouTube’s watch-time change, uploading gaming and makeup videos. At YouTube events she learned that her fans were mostly teenage girls. Kay, then in her
…
, the company converted a 40,000-square-foot airplane hangar into a state-of-the-art production studio for select creators called YouTube Space. But YouTube’s algorithm still wanted the opposite. It desired watch time and daily views; videos that delivered that were usually made cheap. But they rose to the top. One
…
tracked videos about particular world or historical events. The French engineer intentionally noted how such a thing could improve YouTube watch time. Chaslot won praise for it from peers but couldn’t find any interested YouTube managers, and he soon received a negative performance review (a “ding”). Google let him go. He had more
…
the channels that uncritically cheered Trump, such as Alex Jones, under the guise of commentary or punditry. Bundled together, they had more watch time than legitimate news outlets on YouTube. This is a crisis, the staffer pleaded. If YouTube brass agreed, they didn’t say so. But a certifiable crisis came soon enough, and
…
using these surveys and thumbs next to videos to gauge satisfaction. To fix its quality crisis years before, YouTube had switched its gears from views to watch time, but that didn’t cut it anymore. (YouTube never specified the precise equation for its ranking system to outsiders.) When videos suggested the earth was flat
…
popular but what kind of videos were made. Also, when the company wanted to, it went in and turned the dials. Consider Minecraft. After the watch-time transition, YouTube’s audience clearly loved Minecraft, heaving the niche game into the mainstream. At one point, in May 2015, fourteen slots on
…
. (Wikipedia, after Wojcicki spoke, said it had not been informed of this plan.) Wojcicki also introduced a term she had begun using frequently at YouTube. Its algorithms favored watch time, daily viewers, and satisfaction, but they had added a fourth metric. “We’re starting to build in that concept of responsibility,” she told
…
website she documented this corporate crackdown, which she viewed as retaliation for her outspoken challenge to the meat industry. She posted three screenshots of her YouTube dashboard, showing watch time, views, and subscribers on her videos and how they kept falling. One post listed 307,658 minutes of
…
watch time and 366,591 views. “Your estimated revenue,” the YouTube dashboard read, “$0.10.” This she circled in red pixels. “There is no equal growth opportunity on YouTube,” her website blared in bright, frantic text. “Your channel will grow if
…
their videos and use an overall pool of ad money instead, doling out checks based on engagement—the likes, comments, and watch time videos got. This felt fairer and more sustainable. YouTube briefed a few creators on its ambitious plan. In March, Wojcicki presented it to her staff, telling them, “Please don’t
…
. Nearly a decade after tilting its system toward longer videos, YouTube was now paying for shorter ones. Of course, the main algorithmic metric for Shorts, like that for all of YouTube, remained watch time. Most signs indicated that TikTok did chip away at YouTube’s dominance. A 2021 report revealed that for the first time
…
2020, 388 and quality content, 175 responsibility metric, 328 screeners’ role in training, 320 and skeptics of YouTube, 223 skin-detection by, 255–56 titles of content chosen for, 172 watch time favored in, 156–60 and YouTube Kids app, 238, 244–45 Allen & Company (investment bank), 49 Alphabet, 257 alt-right, 263, 269
…
viewers, 252, 254 emphasis on growth of, 91 and initiative to recruit female viewers, 369 and length of viewing sessions, 252 (see also watch time of audience) loyalty to YouTube, 394 number of videos watched daily, 49, 140 satisfaction ratings of, 296–97 See also engagement of users Auletta, Ken, 97 authoritative sources
…
of, 323–24 video responses between, 39 volume of uploaded material, 6, 49, 140, 215–16, 389 and watch time of users, 157, 158–60 and Wojcicki, 261, 373 women’s experience as, 303 and YouTube Creator Summit, 250–51, 253, 262, 289–90 See also partner program; payment and income of creators
…
employees, 317–20, 327, 349 diversity hiring in, 301 as parents, 174 and perks at YouTube offices, 148 poached from Yahoo, 52 and Wojcicki, 211–13 engagement of users emphasis placed on, 154, 158–59 (see also watch time of audience) and machine learning applied to advertising, 191 payments based on (Moneyball proposal
…
, 388 neural networks in, 233–35 as new feature, 23 Reinforce program behind, 298 and right-wing content, 223, 224, 227 and watch time, 154, 155 See also algorithms of YouTube re-creation aesthetic, 27 Reddit, 218, 270 Redstone, Sumner, 60, 62, 76, 253–54 refugees, 264 Reinforce program, 298 related videos sidebar
…
and word of funded channels, 133 Walmart, 286 war crimes, archives of, 296 Warner, Mark, 341 Warren, Elizabeth, 365 watch time of audience and billion-hours goal of YouTube, 228, 270 and COVID-19, 376 and “Delicious”/ “Nutritious” content, 174 and engagement-based payments, 337 and machine learning, 191–92, 233 of pro
by W. David Marx · 18 Nov 2025 · 642pp · 142,332 words
ultimately the success of its hosts, Facebook and YouTube. These platforms, not individual websites, had become the backbone of the internet, reshaping the logic of cultural creation. After brief experiments in elite curation, platforms embraced a laissez-faire approach, prioritizing user numbers and “watch time” over quality content. The twentieth-century content industry
by Geert Mak · 27 Oct 2021 · 722pp · 223,701 words
were followed keenly – in that sense, the continent was far more united than in 1999. But the scenes in the Commons gradually became impossible to watch. Time and again the British premier was chased around the arena like a wounded bull, to laughter and jeers, speared from all sides till she bled
by Glyn Moody · 26 Sep 2022 · 295pp · 66,912 words
YouTube’s chief business officer, Robert Kyncl, told a conference: ‘We are roughly neck-and-neck with Netflix on revenue, actually we are slightly larger and growing faster.’597 Kyncl also revealed that video represents 25% of YouTube ‘watch time’, 50% is YouTube creators and 25% is music
by David de Cremer · 25 May 2020 · 241pp · 70,307 words
/customer-service-trends.html 34 Hoffman, P. (1986). ‘The Unity of Descartes’ Man,’ The Philosophical Review 95, 339-369. 35 Google Duplex (2018). https://www.youtube.com/watch?v=D5VN 56jQMWM Chapter 2: The Leadership Challenge in the Algorithm Age The machine age arrived a long time ago, but today’s
…
example of how utilitarian companies really employ their algorithms is the discussion surrounding how the YouTube algorithm makes recommendations to viewers. The metric that YouTube uses to decide on their recommendations for you as a customer (i.e. watch time) is not aimed at helping customers to get what they want, but rather to
by Rana Foroohar · 5 Nov 2019 · 380pp · 109,724 words
(often correctly) that this was what would keep them coming back and watching more—thus allowing YouTube to make more money from the advertising sold against that content. But because the subtler algorithms resulted in lower “watch time” than the original ones, the project was dropped. Chaslot was gutted; he believed that these
…
-popping content that pays off in shorter—albeit more immediately profitable—bursts. But the powers that be disagreed. Their mentality, according to Chaslot, was that “watch time was an easy metric, and that if users want racist content, ‘well, what can you do?’ ” This was a culture in which the metrics were
…
undermining the fabric of democracy.3 A spokesperson at YouTube, which doesn’t contradict the basic facts of Chaslot’s account, told me in 2018 that the company’s recommendation system has “changed substantially over time” and now includes other metrics beyond watch time, including consumer surveys and the number of shares and
…
‘Cult of Travis,’ ” Financial Times, March 9, 2017. 2. Video of Kalanick arguing with an Uber driver over fares can be accessed here: https://www.youtube.com/watch?v=gTEDYCkNqns. 3. Katy Steinmetz and Matt Vella, “Uber Fail: Upheaval at the World’s Most Valuable Startup Is a Wake-Up Call
by Rob Reich, Mehran Sahami and Jeremy M. Weinstein · 6 Sep 2021
, “the true scarce commodity is increasingly human attention.” When users spend more of their valuable time watching YouTube videos, they must perforce be happier with those videos. It’s a virtuous circle: More satisfied viewership (watch time) begets more advertising, which incentivizes more content creators, which draws more viewership. Our true currency wasn
…
’t views or clicks—it was watch time. The logic was undeniable. YouTube needed a new core metric. To argue for this new metric, he
…
wrote an email to the YouTube executive team arguing that “Watch time, and only watch time” should be the objective to improve at YouTube. In essence, he equated watch time with user happiness: if a person
…
lawn, or even smoking—make us happy or contribute to our well-being. Yet the focus on watch time ultimately became the basis of one of YouTube’s most significant objectives: to reach 1 billion hours of watch time per day by 2016—a goal it ultimately surpassed. To be fair, Goodrow notes that in
…
pursuing its goal, YouTube did sometimes take actions that had a negative impact on watch time if the company believed that the action was in the user’s interest: “For example, we made it a policy to stop
…
recommending click-baity videos.” But he followed up by saying “We never did anything without measuring impact on watch time.” Left out seem to be questions such as whether it’s really healthy for children (or adults, for that matter) to watch an endless stream
…
of videos; whether conspiracy theory videos by flat-earthers should be recommended with the same gusto as more benign videos; or what the race for watch time could do to the ecosystem of content producers—who are paid by advertisers when their videos are watched—who might create more outrageous videos in
…
order to have their content be the centerpiece of the user’s coveted watch time. Doerr himself acknowledged that management systems such as OKRs can have their faults, writing, “Like any management system, OKRs may be executed well or badly
…
difficult to factor it into the metric being optimized. It’s easier to just assume that watching more videos must be making users happier. Measuring watch time is straightforward; determining whether users are actually happier, more factually informed, or politically radicalized, is not. OKRs are just a more recent manifestation of the
…
her article a decade later, she could easily have included more examples from the tech world, such as whether an extreme focus on increasing video watch time may have left little room for considering the political, social, and health impacts of millions of people being glued to their screens. Of course organizations
…
slide”: Ibid., 7. “the marriage of Google”: Ibid., 11. “I think it’s worked out”: Ibid., xi. “As Microsoft CEO Satya Nadella”: Ibid., 161. “Watch time, and only watch time”: Ibid. “For example, we made it”: Ibid., 164. “Like any management system”: Ibid., 9. “Goals Gone Wild”: Lisa D. Ordóñez et al., “Goals Gone
…
Waal, Frans de, 92 Wales, Jimmy, 195 Walker, Darren, 180 Wall Street Journal, 42–43 Warren, Elizabeth, 181, 256 washing machines and laundry, 157–58 watch time metric, 34 Watchdog.net, xxiii Weapons of Math Destruction (O’Neil), 98 Weinberg, Gabriel, 135–36 Weinstein, Jeremy, xv–xvi, 72 Weld, William, 130 Western
by Kelly Weill · 22 Feb 2022
that time, YouTube veered away from recommending videos based on their relevance to someone’s previous viewing habits and started recommending videos that viewers were likely to spend more time watching. “It’s not trying to optimize for relevance,” Chaslot told me. “It’s trying to optimize for watch time, or at
…
was working there.” Which videos kept people on the website the longest? “Extreme videos are extremely good for watch time,” he said. The bizarre, the fringe, and the impossible lured in the most viewers. So YouTube’s recommendation algorithm, at least before a major overhaul in 2019, prioritized the strange. The recommendations often
by Sarah Frier · 13 Apr 2020 · 484pp · 114,613 words
win. The effect had already played out in other parts of the internet, where user-generated content reigns. On YouTube, the site’s algorithm gradually started to reward creators according to watch time, thinking that a longer time spent on a video meant it was engaging enough to be displayed higher in searches
by Peter H. Diamandis and Steven Kotler · 28 Jan 2020 · 501pp · 114,888 words
, 2018). 2010 speech at NYU law school: “The Virtues of Virtual Reality: How Immersive Technology Can Reduce Bias,” April 26, 2019 (video). See: https://www.youtube.com/watch?v=vXxfkkINq8M. In 2016, when Nintendo’s Pokemon GO was downloaded almost a billion times: Lauren Musni, “Pokémon GO Surpasses the 1 Billion
…
,” 9To5Google, May 6, 2019. the soft-spoken Google CEO Sundar Pichai may have stolen back the crown: “Keynote (Google I/O ’18).” See: https://www.youtube.com/watch?v=ogfYd705cRs. According to a recent Zendesk study: “The Impact of Customer Service on Customer Lifetime Value 2013.” See: https://www.zendesk.com
…
.cc.gatech.edu/~riedl/pubs/guzdial-fdg15.pdf. See also: “Artificial Intelligence System for Crowdsourcing Interactive Fiction” (video), September 1, 2016. See: https://www.youtube.com/watch?time_continue=1&v=znqw17aOrCs. From Passive to Active Video games with user-generated gameplay content: “Category: Video Games with User-Generated Gameplay Content,” Wikipedia. See
…
Kurzweil: Ray Kurzweil, author interview, 2018. See also this conversation between Peter and Ray where they discuss the concept of longevity escape velocity: https://www.youtube.com/watch?time_continue=2&v=SaOfLtoaKqw. Aubrey de Grey: Kira Peikoff, “Anti-Aging Pioneer Aubrey de Grey: ‘People in Middle Age Now Have a Fair Chance
by Nicolas Pineault · 6 Dec 2017
by Max Fisher · 5 Sep 2022 · 439pp · 131,081 words
by Charlotte Alter · 18 Feb 2020 · 504pp · 129,087 words
by Mollie Hemingway · 11 Oct 2021 · 595pp · 143,394 words
by Eswar S. Prasad · 27 Sep 2021 · 661pp · 185,701 words
by Clay Shirky · 9 Jun 2010 · 236pp · 66,081 words
by Laura Bates · 2 Sep 2020 · 364pp · 119,398 words
by Jonathan Taplin · 17 Apr 2017 · 222pp · 70,132 words
by Jamie Bartlett · 4 Apr 2018 · 170pp · 49,193 words
by Reid Hoffman and Chris Yeh · 14 Apr 2018 · 286pp · 87,401 words
by Michael Shellenberger · 28 Jun 2020
by James Nestor · 25 May 2020 · 365pp · 96,573 words
by Susan Linn · 12 Sep 2022 · 415pp · 102,982 words
by Laura Shin · 22 Feb 2022 · 506pp · 151,753 words
by Paul Scharre · 18 Jan 2023
by Daron Acemoglu and Simon Johnson · 15 May 2023 · 619pp · 177,548 words
by Brian Dumaine · 11 May 2020 · 411pp · 98,128 words
by Gary Vaynerchuk · 30 Jan 2018
by John Doerr · 23 Apr 2018 · 280pp · 71,268 words
by Dan Heath · 3 Mar 2020
by Ken Auletta · 4 Jun 2018 · 379pp · 109,223 words
by Yascha Mounk · 15 Feb 2018 · 497pp · 123,778 words
by Geoffrey Cain · 15 Mar 2020 · 540pp · 119,731 words
by Frank Pasquale · 14 May 2020 · 1,172pp · 114,305 words
by Scott J. Shapiro · 523pp · 154,042 words
by Richard Seymour · 20 Aug 2019 · 297pp · 83,651 words
by Talia Lavin · 14 Jul 2020 · 231pp · 71,299 words
by Eliot Higgins · 2 Mar 2021 · 277pp · 70,506 words
by Yolande Strengers and Jenny Kennedy · 14 Apr 2020
by Noreena Hertz · 13 May 2020 · 506pp · 133,134 words
by Manuel Castells · 31 Aug 1996 · 843pp · 223,858 words
by Jill Abramson · 5 Feb 2019 · 788pp · 223,004 words
by Alex Moazed and Nicholas L. Johnson · 30 May 2016 · 324pp · 89,875 words
by Nicholas Carr · 5 Sep 2016 · 391pp · 105,382 words
by Simon Clark and Will Louch · 14 Jul 2021 · 403pp · 105,550 words
by Adam Aleksic · 15 Jul 2025 · 278pp · 71,701 words
by Joseph Menn · 3 Jun 2019 · 302pp · 85,877 words
by Brad Smith and Carol Ann Browne · 9 Sep 2019 · 482pp · 121,173 words
by Nicole Perlroth · 9 Feb 2021 · 651pp · 186,130 words
by Natalie Berg and Miya Knights · 28 Jan 2019 · 404pp · 95,163 words
by Christine Lagorio-Chafkin · 1 Oct 2018
by Greg Lukianoff and Jonathan Haidt · 14 Jun 2018 · 531pp · 125,069 words
by Deborah E. Lipstadt · 29 Jan 2019 · 276pp · 71,950 words
by Valliappa Lakshmanan, Sara Robinson and Michael Munn · 31 Oct 2020
by Christian Davenport · 20 Mar 2018 · 390pp · 108,171 words
by Robert M. Sapolsky · 1 May 2017 · 1,261pp · 294,715 words
by Robert J. Gordon · 12 Jan 2016 · 1,104pp · 302,176 words
by Pete Dyson and Rory Sutherland · 15 Jan 2021 · 342pp · 72,927 words
by Jordan Ellenberg · 14 May 2021 · 665pp · 159,350 words
by Sarah Kendzior · 6 Apr 2020
by Vaclav Smil · 2 Mar 2021 · 1,324pp · 159,290 words
by Anthony Sattin · 25 May 2022 · 412pp · 121,164 words
by Brian Hare and Vanessa Woods · 13 Jul 2020
by Christian Wolmar · 5 Sep 2018 · 292pp · 85,381 words
by Tracee Stanley · 9 Mar 2021
by Steve Coogan · 1 Sep 2011
by Nick Harkaway · 18 Oct 2017 · 778pp · 239,744 words
by John Micklethwait and Adrian Wooldridge · 14 May 2014 · 372pp · 92,477 words
by Charles L. Marohn, Jr. · 24 Sep 2019 · 242pp · 71,943 words
by Morgan Housel · 7 Sep 2020 · 209pp · 53,175 words
by Howard G. Buffett · 2 Apr 2018 · 350pp · 109,521 words
by Elaine N. Aron · 1 Dec 2013 · 323pp · 94,683 words
by Emily Guendelsberger · 15 Jul 2019 · 382pp · 114,537 words
by Ben Stewart · 4 May 2015 · 347pp · 94,701 words