Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 1 of 81
`
`
`
`
`
`
`
`
`
`
`
`
`Exhibit 17
`to the Declaration of Catherine Hartman
`Public Redacted Version
`
`
`
`
`Exhibit 17
`
`to the Declaration of Catherine Hartman
`
`Public Redacted Version
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 2 of 81
`
`
`
`UNITED STATES DISTRICT COURT
`NORTHERN DISTRICT OF CALIFORNIA
`
`
`
`MARIA SCHNEIDER, UNIGLOBE
`ENTERTAINMENT, LLC and AST
`PUBLISHING, LTD, individually and on behalf
`of all others similarly situated;
`Plaintiffs,
`
`
`vs.
`
`
`YOUTUBE, LLC; GOOGLE LLC;
`Defendants
`
`
`
`
`
`
`
`
`
`
`CASE NO. 3:20-cv-04423-JD
`
`
`
`
`EXPERT REPORT OF
`HAL J. SINGER, PH.D.
`FOR PLAINTIFFS
`
`
`
`November 17, 2022
`
`
`
`
`
`
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 3 of 81
`
`
`
` -2-
`
`
`
`Table of Contents
`
`Introduction and Assignment .......................................................................................................... 3
`
`Limited Data Availability ............................................................................................................... 6
`
`Summary of Opinions ................................................................................................................... 10
`
`Qualifications ................................................................................................................................ 11
`
`I.
`
`II.
`
`Industry Background and the Nature of the Challenged Conduct ........................................ 13
`A.
`Overview of YouTube .......................................................................................... 14
`1. Advertising Revenue ...................................................................................... 15
`2. Subscription Revenue .................................................................................... 22
`YouTube Is a Multi-Sided Platform ..................................................................... 23
`1. Direct and Indirect Network Effects .............................................................. 24
`2. Advertisers as Subsidizers of the Provider-Consumer Interaction ................ 29
`
`B.
`
`Impact and Disgorgement of Revenues from Infringing Content ......................................... 33
`A.
`Theoretical Framework for My Econometric Models .......................................... 35
`B. Empirical Estimation of Econometric Models ............................................................. 39
`1. Effect of Infringing Content on Overall YouTube Revenues ........................ 43
`2. Effect of Infringing Content on User Engagement with YouTube ................ 47
`Econometric Estimate of YouTube Advertising Revenues Subject to
`Disgorgement ........................................................................................................ 49
`
`B.
`
`Conclusion .................................................................................................................................... 58
`
`Appendix A: Materials Relied Upon ............................................................................................ 60
`
`Appendix B: Curriculum Vitae ..................................................................................................... 70
`
`
`
`
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 4 of 81
`
`
`
` -3-
`
`
`
`INTRODUCTION AND ASSIGNMENT
`
`1.
`
`I understand that Plaintiffs allege copyright infringement on the part of Defendants
`
`YouTube, LLC (“YouTube”) and its parent company Google LLC (“Google”) (collectively,
`
`“Defendants”). Specifically, Plaintiffs allege that Defendants engaged in acts causing videos that
`
`infringed the works of Plaintiffs and the putative Classes (“Infringing Content”) to be reproduced,
`
`distributed, displayed, and publicly performed on the internet (“the Challenged Conduct”).1
`
`2.
`
`YouTube offers a video-streaming service as well as a music service titled
`
`“YouTube Music” that offers similar services to Apple Music and Spotify. I understand that the
`
`Challenged Conduct at issue deals only with YouTube’s video-streaming service. Thus, for
`
`purposes of this report, I use the term “YouTube” to refer only to the video-streaming service.
`
`3.
`
`I also understand that, inter alia, Plaintiffs seek disgorgement of any profits that
`
`Defendants derived from the Challenged Conduct.2 Such profits fall into two main categories: (1)
`
`advertising or subscription revenues that YouTube obtained from the playback of videos that
`
`included the Infringing Content; and (2) incremental YouTube profits from the presence of non-
`
`infringing content resulting from the spillover effects to which the Infringing Content contributed,
`
`allowing YouTube to increase viewership and thus attract additional non-infringing content.
`
`4.
`
`Profits in the second category flow from the display of the subset of non-infringing
`
`content that is amplified by the display of Infringing Content—profits that would not have
`
`occurred absent the spillover effects that Infringing Content created. For example, such spillover
`
`
`1. First Amended Class Action Complaint (“Complaint”), Dkt. 99, ¶114.
`Id. ¶156.
`2.
`
`CONFIDENTIAL
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 5 of 81
`
`
`
` -4-
`
`
`
`effects can occur through the use of YouTube’s recommendation engine.3 As Google researchers
`
`observe:
`
`Many hours’ worth of videos are uploaded each second to YouTube.
`Recommending this recently uploaded (“fresh”) content is extremely important for
`YouTube as a product. We consistently observe that users prefer fresh content,
`though not at the expense of relevance.4
`
`Thus, spillover effects could occur when a content provider uploads a video that contains
`
`Infringing Content, a user views the video, then receives recommendations from YouTube for non-
`
`infringing content. In other words, videos that contain Infringing Content can serve as the basis
`
`for YouTube’s recommendation of non-infringing content. As Cristos Goodrow, Vice President
`
`of Engineering at YouTube, explains, “Recommendations drive a significant amount of the overall
`
`viewership on YouTube, even more than channel subscriptions or search.”5 YouTube benefits
`
`from such spillover effects by serving ads on non-infringing content and gathering user behavior
`
`information that can inform its ad serving throughout the Google ad platform.
`
`5.
`
`Defendants’ revenues and profits subject to disgorgement that I estimate in this
`
`report cover both categories (1) and (2) described above. Google owns YouTube, and has done so
`
`since 2006.6 In this report, I focus on YouTube, as the alleged infringement occurred either within
`
`
`3. Paul Covington, Jay Adams, Emre Sargin, Deep Neural Networks for YouTube Recommendations, Google
`Research, presentation paper at The ACM Conference Series on Recommender Systems (‘RecSys’), 2016, available
`at https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/45530.pdf (“YouTube is the
`world's largest platform for creating, sharing and discovering video content. YouTube recommendations are
`responsible for helping more than a billion users discover personalized content from an ever-growing corpus of
`videos.… During candidate generation, the enormous YouTube corpus is winnowed down to hundreds of videos that
`may be relevant to the user.”)
`Id.
`4.
`5. Cristos Goodrow, On YouTube’s recommendation system, YOUTUBE BLOG, Sept. 15, 2021, [hereafter
`“Goodrow 2021”] available at https://blog.youtube/inside-youtube/on-youtubes-recommendation-system/ (“You can
`find recommendations at work in two main places: your homepage and the ‘Up Next’ panel. Your homepage is what
`you see when you first open YouTube—it displays a mixture of personalized recommendations, subscriptions, and
`the latest news and information. The Up Next panel appears when you’re watching a video and suggests additional
`content based on what you’re currently watching, alongside other videos that we think you may be interested in.”).
`6. Nicholas Jackson, Infographic: The History of Video Advertising on YouTube, The Atlantic, August 3, 2011,
`available at https://www.theatlantic.com/technology/archive/2011/08/infographic-the-history-of-video-advertising-
`on-youtube/242836/.
`
`CONFIDENTIAL
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 6 of 81
`
`
`
` -5-
`
`
`
`its website or through the embedding of content uploaded to YouTube on third-party sites.7 By
`
`focusing on YouTube, I do not intend to imply that Plaintiffs do not seek damages from Google
`
`itself. Google serves the ads that occur on YouTube.8
`
`6.
`
`Plaintiffs’ motion for class certification identifies four putative classes: two
`
`Copyright Infringement (“CI”) Classes and two Copyright Management Information (“CMI”)
`
`Classes.9 My report focuses on disgorgement of profits from the effects of the Challenged Conduct
`
`on the CI Classes in the first instance. I understand from counsel that the CI Classes will include
`
`U.S. registered works (“the Registered Works Infringement Class”) and foreign unregistered
`
`works (“the Foreign Unregistered Works Infringement Class”). I further understand that the
`
`compensation to members of the Registered Works Infringement Class and CMI Class members
`
`may take the form of statutory damages. I also understand that one factor in assessing statutory
`
`damages includes the profits gained by the Defendants from the entirety of their course of conduct,
`
`and the amounts needed to deter Defendants from future infractions. My report quantifies the
`
`profits that Defendants have earned as a result of the Challenged Conduct, and so may also be
`
`relevant to the question of statutory damages. I also summarize the overall profitability of
`
`Defendants’ businesses over time, and the revenues generated by YouTube, as obtained through
`
`production in this litigation.
`
`7.
`
`I offer no opinion on whether infringement occurred, or on whether record
`
`documents support an inference of infringement. For purposes of my assignment, I assume
`
`Plaintiffs’ allegations regarding infringement are true. Counsel have asked me to propose a
`
`
`7. Embedding refers to adding a YouTube video or playlist to a website or blog by including the relevant HTML
`code. See, e.g., https://support.google.com/youtube/answer/171780?hl=en. I understand that the embedding feature is
`available for any video that appears on YouTube.
`8. https://ads.google.com/home/campaigns/video-ads/.
`9. ECF No. 190.
`
`CONFIDENTIAL
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 7 of 81
`
`
`
` -6-
`
`
`
`common and reliable methodology by which, if the factfinder determines that infringement
`
`occurred, then impact to Class Members may be reliably evaluated with common methods and
`
`evidence. Plaintiffs’ counsel have also asked me to calculate damages in this matter in the form of
`
`disgorgement of YouTube’s revenues and profits that resulted from the Challenged Conduct.
`
`8.
`
`The opinions expressed in this report reflect my review of the evidence, data,
`
`testimony, and other relevant materials to date. The materials upon which I rely for this report
`
`appear in Appendix A. I reserve the right to supplement or amend my opinions should new
`
`materials or information become available that warrant my doing so.
`
`LIMITED DATA AVAILABILITY
`
`9.
`
`I understand that Defendants maintain the requisite data I need to perform my
`
`assignment. Because Defendants have only produced a very limited portion of such data, however,
`
`I supplemented my analysis with publicly-available data where possible. I understand that Counsel
`
`for Plaintiffs have requested various data from Defendants that could inform the existence of
`
`network effects and quantify their financial impacts on the company. In the event that Defendants
`
`produce additional data, I reserve the right to supplement my report accordingly.
`
`10.
`
`For purposes of this report and the calculations contained herein, the most salient
`
`data that Defendants initially produced reflect monthly advertising revenues, costs, and profits for
`
`the four-year period from January 2017 through December 2020. Pursuant to a Court order,
`
`Defendants also produced additional limited data on October 7, 2022. I understand that the Court
`
`directed Defendants to produce the following: (1) Plaintiff-requested takedown notice data from a
`
`random sampling of 90 days; (2) the percentage of disputed Content ID claims for the year 2021
`
`where the user selected fair use as the ground for the dispute; (3) Plaintiff-requested Content ID
`
`statistics for the United States on August 1 of 2017, 2018, 2019, and 2020; and (4) Plaintiff-
`
`CONFIDENTIAL
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 8 of 81
`
`
`
` -7-
`
`requested CMI data from a random sampling of 30 days.10 Of these additional data that Defendants
`
`produced, only the takedown notice data and Content ID statistics are relevant to the issues I
`
`address in this report.11 As I explain subsequently, the results indicate that takedown notices have
`
`increased along with YouTube’s visits and revenues. The subsequent regressions that I estimate
`
`support the inference that Infringing Content, as proxied by takedown notices, has a causal impact
`
`on YouTube revenues via network effects.
`
`11.
`
`To my knowledge, Defendants have not produced any YouTube traffic data such
`
`as viewership and content metrics (e.g., number of content views, ad impressions, ad revenue by
`
`channel, number of channels, number of videos uploaded, total video minutes uploaded) by time
`
`period (e.g., weekly, monthly, etc.) that correspond to the YouTube monthly revenue and cost data
`
`noted above.12
`
`12.
`
`Such traffic data function, inter alia, as explanatory factors, or economic drivers,
`
`of YouTube revenues. The relationship between viewers and content providers also informs the
`
`indirect network effects that characterize the YouTube platform: content (on one side of the
`
`10. ECF No. 165. See also, letter from Lauren Gallo White to Carol O’Keefe dated October 7, 2022.
`11.
`In the event that Defendants are ordered to produce both the numerator and denominator values for the
`Content ID statistics, I would use both the total number of videos on the platform, and the number of publicly-
`displayed videos on the platform in my analysis.
`12. In light of Defendants’ limited production, my staff and I pursued potential sources of reliable, publicly
`available information that could function as a provisional substitute for YouTube’s data. My staff and I identified
`Comscore as one such possibility. I understand that Defendants have entered deals to provide Comscore with YouTube
`data. On August 22, 2011, Comscore introduced YouTube Partner Reporting as part of its Video Metrix product,
`explaining that, “This new capability allows Comscore to report on the video-viewing audiences cultivated by specific
`partners within the YouTube universe – an increasingly important component of the online video-viewing landscape.”
`Comscore also detailed various benefits of its YouTube Partner Reporting feature that provides substantial data on
`viewership and demographics overall by month and by channel. See, e.g., Comscore, Comscore Introduces YouTube
`Partner
`Reporting
`in
`Video Metrix, Comscore Blog, August
`available
`at
`22,
`2011,
`https://www.comscore.com/Insights/Blog/comScore-Introduces-YouTube-Partner-Reporting-in-Video-Metrix. At
`the request of Plaintiffs’ Counsel, I attempted to purchase such data and was informed that while Comscore sells such
`information for research purposes (e.g., a university subscription), a third-party purchase raises additional questions
`that would involve approval from its legal department. After an approximate two-week delay, the legal department
`denied approval for our requested subscription. Thus, neither YouTube nor the third-party data collection firm to
`which it provides the same data that I understand Plaintiffs’ counsel has requested would make the necessary
`information available, not even for a fee.
`
`HIGHLY CONFIDENTIAL - ATTORNEYS' EYES ONLY
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 9 of 81
`
`
`
` -8-
`
`
`
`platform) draws viewers (on the other side of the platform), who in turn draw more content, and
`
`so on, illuminating the self-reinforcing nature of such effects.13 The literature has recognized the
`
`potential influence of network effects on the economics of intellectual property since before the
`
`advent of YouTube, Facebook, and other digital media platforms.14
`
`13.
`
`In pursuing other available avenues of data, I located similar data through the online
`
`visibility management and content marketing Software as a Service (SaaS) platform, Semrush.15 I
`
`understand that technology firms work with Semrush, a publicly-traded firm, to evaluate online
`
`
`13. See, e.g., Lianfei Qui, Qian Tang, and Andrew Whinston, Two Formulas for Success in Social Media:
`Learning and Network Effects, 32(4) JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 78-108, 78 (2015) [hereafter
`Qui et al. 2015] (“Using a unique data set from YouTube, we empirically identify learning and network effects
`separately, and find that both mechanisms have statistically and economically significant effects on video views;
`furthermore, the mechanism that dominates depends on the video type.”); Anjana Susarla, Jeong-Ha Oh, and Yong
`Tan, Social Networks and the Diffusion of User-Generated Content: Evidence from YouTube, INFORMATION SYSTEMS
`RESEARCH 1-19, 1 (2011) (“The networked structure of interactions on YouTube and the tremendous variation in the
`success of videos posted online lends itself to an inquiry of the role of social influence. Using a unique data set of
`video information and user information collected from YouTube, we find that social interactions are influential not
`only
`in determining which videos become successful but also on
`the magnitude of
`that
`impact.”)
`https://pubsonline.informs.org/doi/epdf/10.1287/isre.1100.0339; Peter Menell, Economic Analysis of Network Effects
`and Intellectual Property, 34 BERKELEY TECHNOLOGY LAW JOURNAL 219-322, 232-3 (2019) (“Network effects have
`allowed one or a few firms to dominate many Internet markets, including search (Google), social networks (Facebook),
`mobile (iOS, Android), commerce (Amazon, eBay), content streaming (YouTube, Netflix, Spotify), payment systems
`(PayPal), and sharing networks (Airbnb, Uber).”). Id. at 322 (“Perhaps most significantly, social media platforms,
`such as Facebook, Instagram, YouTube, and Twitter, are increasingly important not only to economic activity but also
`to social mobilization, electoral processes, and the functioning of democracy. These platforms are driven by network
`effects but are also notable for their polarizing tendencies.”).
`14. National Research Council. The Digital Dilemma: Intellectual Property in the Information Age. Washington,
`DC: The National Academies Press (2000) at 17 (“Research should be conducted to characterize the economic impacts
`of copyright. Such research might consider, among other things, the impact of network effects in information industries
`available
`at
`and
`how
`digital
`networks
`are
`changing
`transaction
`costs.”),
`https://nap.nationalacademies.org/catalog/9601/the-digital-dilemma-intellectual-property-in-the-information-age.
`15. Semrush, Our message, available at https://www.semrush.com/company/. Semrush explains that “Semrush
`Traffic Analytics is based on petabytes of clickstream data that comes from multiple proprietary and 3rd party data
`sources, Semrush’s proprietary AI and machine learning algorithms and Big Data technologies. The data is
`accumulated and approximated from the user behavior of over 200 million real internet users and over a hundred
`different apps and browser extensions. We offer our ‘estimated accuracy’ metric at the top of the report so you can
`gauge how accurate the reported numbers are.” For a review of Semrush’s internal testing of its data accuracy, see
`https://www.semrush.com/blog/us-search-volume-update/.
`
`CONFIDENTIAL
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 10 of 81
`
`
`
` -9-
`
`
`
`marketing goals.16 From Semrush, I purchased a subscription to its Traffic Analytics Report17,
`
`through which I obtained monthly clickstream data for YouTube as well as other video platforms
`
`over the period from January 2017, the earliest month available, through September 2022.18 Such
`
`clickstream data reflect user behavior but do not capture server-side information such as the total
`
`number of pages available, number of channels, the number of subscriptions on each channel, and
`
`so on. Such limitations notwithstanding, clickstream data are widely employed in the industry to
`
`support marketing and analytic efforts.19 A sample of these data for April 2017 appears below.20
`
`TABLE 1: SEMRUSH TRAFFIC ANALYTICS DATA FOR YOUTUBE AND SELECTED VIDEO SITES, APRIL 2017
`Unique
`Pages /
`Avg. Visit
`Bounce
`Visitors
`Visit
`Duration
`Rate
`
`Target
`
`Target type
`
`Visits
`
`youtube.com
`
`root domain
`
`8,685,745,225
`
`691,441,692
`
`facebook.com
`
`root domain
`
`4,528,701,672
`
`386,402,252
`
`vimeo.com
`
`root domain
`
`90,729,711
`
`50,929,237
`
`tiktok.com
`
`root domain
`
`twitch.com
`
`
`root domain
`
`n/a
`
`9,164
`
`n/a
`
`8,616
`
`9.73
`
`9.87
`
`4.88
`
`n/a
`
`2.08
`
`41:50
`
`35:56
`
`11:09
`
`n/a
`
`03:45
`
`25.32%
`
`30.45%
`
`45.36%
`
`n/a
`
`54.41%
`
`14.
`
`Relying on these data, I demonstrate the existence of indirect network effects and
`
`quantify their impact on YouTube’s advertising revenues. My proposed methodology employs a
`
`
`16. See, e.g., Domenica D’Ottavio, 2021 Data-Backed Digital Marketing Predictions, available at
`https://www.frac.tl/2021-data-backed-digital-marketing-predictions/. See also Amanda Milligan, Three strategies for
`elevating brand authority in 2021, TECHCRUNCH, Feb. 19, 2021, available at https://techcrunch.com/2021/02/19/how-
`to-elevate-brand-authority-in-2021/.
`17. Traffic Analytics Overview Report, Semrush, available at https://www.semrush.com/kb/895-traffic-analytics-
`overview-report.
`18. Semrush indicates that it collects clickstream data from various providers. Clickstream data represent a record
`of an individual’s clicks while navigating the internet. See, Alexander S. Gillis, Definition: clickstream data
`(clickstream analytics)(“Clickstream data and clickstream analytics are the processes involved in collecting, analyzing
`and reporting aggregate data about which pages a website visitor visits -- and in what order. The path the visitor takes
`available
`at
`through
`a
`website
`is
`called
`the
`clickstream.”),
`https://www.techtarget.com/searchcustomerexperience/definition/clickstream-analysis-clickstream-analytics.
`Semrush explains that “we leverage petabytes of clickstream data received by 3rd party providers, along with machine
`learning algorithms and Big Data technologies.” Available at https://www.semrush.com/blog/what-is-clickstream-
`data/.
`19. For an overview of clickstream data, see BillAlbert, TomTullis, DonnaTedesco, BEYOND THE USABILITY LAB –
`(Elsevier 2010), available at
`CONDUCTING LARGE-SCALE ONLINE USER EXPERIENCE STUDIES 107-119
`https://www.sciencedirect.com/topics/computer-science/clickstream-data.
`20. I understand that monthly data represent the most granular data available through the Semrush Traffic
`Analytics tool.
`
`CONFIDENTIAL
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 11 of 81
`
`
`
` -10-
`
`
`
`similar approach to that used in the research literature, including work that investigates advertising
`
`effectiveness published by Google’s own research team.21
`
`SUMMARY OF OPINIONS
`
`15.
`
`In this report, I rely on common economic evidence and methods for the
`
`preliminary conclusions I reach.
`
`16.
`
`First, my analysis empirically confirms the observation ubiquitous in the research
`
`literature that YouTube benefits from indirect network effects. Such effects allow YouTube to
`
`earn higher advertising revenues on non-infringing content because the presence of Infringing
`
`Content attracts more consumer attention and increases the amount of user-specific and cohort-
`
`based preference and interest data generated by the platform as a whole.
`
`17.
`
`Second, I conclude that a formulaic methodology common to all members of each
`
`class can be utilized to estimate the YouTube revenues generated by the Challenged Conduct,
`
`including: (1) advertising or subscription revenues that YouTube obtained from the playback of
`
`videos that included the Infringing Content; and (2) incremental YouTube profits from the
`
`presence of non-infringing content resulting from the spillover effects to which the Infringing
`
`Content contributed, allowing YouTube to increase viewership and thus attract additional non-
`
`infringing content.
`
`18.
`
`Third, I have performed the requisite calculations and applied the methodologies
`
`detailed herein to Defendants’ data (or estimated using a combination of Defendants’ and third-
`
`party data) as I describe in this report. Such calculations do not require individualized inquiry.
`
`
`21. Jouni Kerman, Peng Wang, and Jon Vaver, Estimating Ad Effectiveness using Geo Experiments in a Time-
` available at
`Based Regression Framework, Google, March 2017,
`[hereafter Kerman et al],
`https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/45950.pdf.
`
`HIGHLY CONFIDENTIAL - ATTORNEYS' EYES ONLY
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 12 of 81
`
`
`
` -11-
`
`
`
`YouTube maintains the critical data needed for the application of the common methodology that I
`
`propose.
`
`19.
`
`Fourth, based on my performance of the requisite calculations and the applied
`
`methodologies, I have estimated the profits garnered by Defendants from the Challenged Conduct
`
`for the period from July 2017 through approximately September 2022 using three discrete inputs:
`
`• Utilizing the total number of unique infringing URLs included in the YouTube takedown
`
`data sample as a proxy for the total Infringing Content (videos that infringed the works of
`
`Plaintiffs and the putative Classes), I estimate that Infringing Content accounts for
`
`of time spent on the site and has generated profits of approximately
`
`million;
`
`• Utilizing a publicly available estimate that 2% of videos on the platform are infringing, I
`
`estimate that all infringing videos account for 2% of views and generated profits of
`
`approximately $1.49 billion;
`
`• Utilizing a publicly available estimate that 6% of videos on the platform are infringing, I
`
`estimate that all infringing videos account for 6% of views and generated profits of
`
`approximately $4.48 billion.
`
`20.
`
`The methodological framework and analysis in this report can be applied to a
`
`broader scope of data should YouTube make such data available. Once Plaintiffs’ expert Dr.
`
`Charles Cowan has calculated the number of unique URLs subject to recovery for each of the CI
`
`Classes, I can apply this same methodological framework and analysis to estimate the profits
`
`generated by the unique URLs subject to recovery for each of the CI Classes.
`
`QUALIFICATIONS
`
`21.
`
`I am a managing director at Econ One, an economic consulting firm that provides
`
`expert economic and econometric analysis for antitrust cases. I am also an adjunct professor at the
`
`HIGHLY CONFIDENTIAL - ATTORNEYS' EYES ONLY
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 13 of 81
`
`
`
` -12-
`
`
`
`McDonough School of Business at Georgetown University, where I teach advanced pricing to
`
`MBA candidates. I am also an Adjunct Professor of Economics at the University of Utah and the
`
`director of the Utah Project, an inter-disciplinary center devoted to the study of antitrust and
`
`consumer protection issues.
`
`22.
`
`I am an applied microeconomist with an emphasis on industrial organization and
`
`regulation. In an academic capacity, I have published several books and book chapters, spanning
`
`a range of industries and topics, and my articles have appeared in dozens of legal and economic
`
`journals. My competition-related articles have appeared in multiple American Bar Association
`
`(ABA) Antitrust Section journals, and I have been a panelist at several ABA Antitrust events. In a
`
`consulting capacity, I have been nominated for antitrust practitioner of the year among economists
`
`by the American Antitrust Institute (AAI) for my work in Tennis Channel v. Comcast, and AAI
`
`named me as co-Honoree in the same category in 2018 for my work In Re Lidoderm Antitrust
`
`Litigation. I have specific experience and expertise in technology markets, such as the one here.
`
`For example, I am currently the expert for Plaintiffs in the Carr v. Google litigation. And I am the
`
`Federal Trade Commission’s expert in its challenge of Meta’s acquisition of fitness app maker
`
`Within.
`
`23.
`
`I have testified as an economic expert in state and federal courts, as well as before
`
`regulatory agencies. I also have testified before the House Judiciary Subcommittee on Antitrust
`
`and the Senate Judiciary Subcommittee on Competition Policy, Antitrust, and Consumer Rights
`
`on the interplay between antitrust and sector-specific regulation. Federal courts have relied on my
`
`CONFIDENTIAL
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 14 of 81
`
`
`
` -13-
`
`
`
`work in certifying seven classes in antitrust matters,22 and two classes in consumer protection
`
`matters.23
`
`24.
`
`I earned M.A. and Ph.D. degrees in economics from Johns Hopkins University and
`
`a B.S., magna cum laude, in economics from Tulane University. My full curriculum vitae appears
`
`as Appendix B to this report and reflects a full list of the cases in which I have served as a testifying
`
`expert since 2014 and a list of publications I have authored in the last ten years.
`
`25.
`
`Neither Econ One nor I have a financial stake in the outcome of this case. Econ
`
`One is being compensated for my work in this matter at my standard rate of $885 per hour,
`
`regardless of whether Plaintiffs prevail or not.
`
`I. INDUSTRY BACKGROUND AND THE NATURE OF THE CHALLENGED CONDUCT
`
`26.
`
`In this section, I summarize my understanding of YouTube’s video platform, the
`
`methods that YouTube uses to monetize traffic, and how the alleged infringement generated
`
`revenues that YouTube would not have otherwise enjoyed. I do not intend this section to represent
`
`a comprehensive discussion or testimony regarding the factual background of this case but rather
`
`to serve as a frame of reference for the economic opinions that follow. I am aware of the following
`
`
`22. See Meijer, Inc. v. Abbott Laboratories, No. C 07-5985 CW, 2008 WL 4065839 (N.D. Cal. Aug. 27, 2008)
`(Order Granting Plaintiffs’ Motion for Class Certification); Natchitoches Parish Hosp. Serv. Dist. v. Tyco Intl., Ltd.,
`262 F.R.D. 58 (D. Mass. 2008) (Granting Motion to Certify Class); Southeast Missouri Hospital and St. Francis
`Medical Center v. C.R. Bard, No. 1:07cv0031 TCM, 2008 WL 4372741 (E.D. Mo. Sept. 22, 2008) (Granting in Part
`Motion for Class Certification); Johnson v. Arizona Hosp. and Healthcare Assoc. No. CV 07-1292-PHX-SRB, 2009
`WL 5031334 (D. Ariz. July 14, 2009) (Granting in Part Motion for Class Certification); In re Delta/AirTran Baggage
`Fee Antitrust Litig., 317 F.R.D. 665 (N.D. Ga. 2016) (Granting Motion to Certify Class); and In re Lidoderm Antitrust
`Litig., No. 12-md-02521, 2017 WL 679367 (N.D. Cal. Feb. 21, 2017) (Order Granting Motions for Class Certifications
`and Denying Daubert Motions); Cung Le, et al. v. Zuffa, LLC d/b/a Ultimate Fighting Championship, Minute Entry,
`2:15-cv-01045-RFB-BNW (D. Nev. Dec. 10, 2020), ECF No. 781 (announcing the court’s intention to grant the
`Plaintiffs’ Motion for Class Certification). As of the time of this report, the court has not issued the written opinion
`certifying the class in Zuffa.
`23. See In Re: MacBook Keyboard Litigation, Case No. 5:18-cv-02813-EJD, 2021 WL 1250378 (N.D. Cal.,
`Mar. 8, 2021) (Order Granting Motion to Certify Class); In Re: JUUL Labs, Inc., Marketing, Sales Practices, and
`Products Liability Litigation, Case No. 19-md-02913-WHO (N.D. Cal., Dec. 5, 2021) (Tentative Opinions on
`Motion for Class Certification, Daubert Motions, and Motion to Dismiss Bellwether Claims).
`
`CONFIDENTIAL
`
`

`

`Case 3:20-cv-04423-JD Document 209-15 Filed 12/08/22 Page 15 of 81
`
`
`
` -14-
`
`
`
`facts and allegations regarding this matter based on documents related to this case and publicly
`
`availab