Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 1 of 28 1 2 3 4 5 6 7 Eric H. Gibbs (SBN 178658) Dylan Hughes (SBN 209113) David Stein (SBN 257465) Aaron Blumenthal (SBN 310605) GIBBS LAW GROUP LLP 505 14th Street, Suite 1110 Oakland, CA 94612 Telephone: (510) 350-9700 Facsimile: (510) 350-9701 ehg@classlawgroup.com dsh@classlawgroup.com ab@classlawgroup.com Robert T. Eglet (pro hac vice) Robert M. Adams (pro hac vice) Erica D. Entsminger (pro hac vice) Artemus W. Ham (pro hac vice) EGLET PRINCE 400 South Seventh Street, Suite 400 Las Vegas, NV 89101 Telephone: (702) 450-5400 Facsimile: (702) 450-5451 eservice@egletlaw.com 8 9 10 11 12 13 14 Andrew N. Friedman (pro hac vice) Geoffrey Graber (SBN 211547) Eric Kafka (pro hac vice) COHEN MILSTEIN SELLERS & TOLL PLLC 1100 New York Ave. NW, Fifth Floor Washington, DC 20005 Telephone: (202) 408-4600 Facsimile: (202) 408-4699 afriedman@cohenmilstein.com ggraber@cohenmilstein.com ekafka@cohenmilstein.com 15 [Additional counsel on signature page] 16 Counsel for Plaintiffs and Proposed Class 17 18 19 UNITED STATES DISTRICT COURT FOR THE NORTHERN DISTRICT OF CALIFORNIA OAKLAND DIVISION 20 LLE ONE, LLC, d/b/a Crowd Siren and d/b/a 21 Social Media Models, and JONATHAN MURDOUGH, on behalf of themselves and all 22 others similarly situated, Case Nos.: 4:16-cv-06232-JSW 23 DEMAND FOR JURY TRIAL Plaintiffs, FOURTH AMENDED CLASS ACTION COMPLAINT 24 v. 25 FACEBOOK, INC., 26 27 Defendant. 28 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 2 of 28 1 2 INTRODUCTION 1. In September 2016, the Wall Street Journal revealed that, for the past two years, 3 Facebook had been overstating the average time its users spent watching paid video advertisements. 4 Based on information from advertising agencies who had spoken with Facebook, the Wall Street 5 Journal reported that Facebook’s metrics had been overstated by between 60 and 80%. 6 2. In response to the media attention, Facebook admitted it made a mistake, but 7 emphasized that it had only discovered the mistake “about a month ago,” and that “as soon as we 8 discovered [it], we fixed it.” When Plaintiffs filed this class action on behalf of advertisers, they had 9 no reason to believe Facebook was not telling the truth, and limited their claims to contract and 10 11 statutory claims, which have already been tested and found viable. 3. Internal records recently produced in this litigation suggest, however, that Facebook’s 12 action rises to the level of fraud and may warrant punitive damages. Facebook did not discover its 13 mistake one month before its public announcement. Facebook engineers knew for over a year, and 14 multiple advertisers had reported aberrant results caused by the miscalculation (such as 100% 15 average watch times for their video ads). Yet Facebook did nothing to stop its dissemination of false 16 metrics. 17 4. In addition to Facebook knowing about the problem far longer than previously 18 acknowledged, Facebook’s records also show that the impact of its miscalculation was much more 19 severe than reported. The average viewership metrics were not inflated by only 60%-80%; they were 20 inflated by some 150 to 900%. 21 5. To the extent anyone at Facebook did not fully realize how it was calculating average 22 viewership from the beginning (as Facebook’s engineers did), it was only because of Facebook’s 23 longstanding reckless indifference to the metrics’ accuracy. The wide disparity between the actual 24 average viewership and Facebook’s reported metrics should have been corrected immediately. But 25 Facebook severely understaffed the engineering team in charge of fixing errors—employing only as 26 few as two engineers—and this led to long delays before errors were fully investigated and 27 corrected. As the documents produced in this litigation show, Facebook did not take verification of 28 its metrics seriously, did not timely address known problems, and had never conducted an internal 1 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 3 of 28 1 2 audit of its metrics. 6. Even once Facebook decided to correct the false metrics, it chose not to do so 3 immediately. Instead, Facebook chose to continue disseminating false metrics for several more 4 months while it developed and deployed a “no PR” strategy designed to “obfuscate the fact that we 5 screwed up the math.” All the while, Facebook continued to reap the benefits from the inflated 6 numbers. 7 7. Based on this newly discovered information, Plaintiffs are amending their complaint 8 to add a claim for fraud and are now seeking punitive damages. Plaintiffs’ additional allegations 9 appear in this Introduction, in a New Allegations section set that begins on page 14, and in their 10 Third Cause of Action and revised Prayer, which begins on page 24. (They have also removed the 11 allegations pertaining to former plaintiff, Tyler Barnett PR, LLC.) In all other respects this Fourth 12 Amended Complaint remains unchanged from the Third Amended Complaint, and Plaintiffs are 13 continuing to pursue their previously alleged claims for violation of California’s Unfair Competition 14 Law and breach of Facebook’s contractual duty to perform with reasonable care. 15 16 JURISDICTION 8. This Court has subject matter jurisdiction over this action under 28 U.S.C. § 17 1332(d)(2) because this is a class action wherein the amount in controversy exceeds the sum or value 18 of $5,000,000, exclusive of interest and costs, there are more than 100 members in the proposed 19 class, and at least one member of the class of Plaintiffs is a citizen of a state different from a 20 Defendant. 21 22 23 9. This Court has personal jurisdiction over Defendant Facebook, Inc., because Facebook, Inc., is headquartered in California, and conducts business in the state of California. 10. Venue is proper in this Court pursuant to 28 U.S.C. §1391(b) because a substantial 24 part of the events or omissions giving rise to the claims occurred in, were directed to, and/or 25 emanated from this District. Venue is also proper because Facebook’s terms of service require that 26 claims are resolved “exclusively in the U.S. District Court for the Northern District of California or a 27 28 2 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 4 of 28 1 state court located in San Mateo County….”1 2 PARTIES 3 4 11. Plaintiff LLE One, LLC, d/b/a Crowd Siren and d/b/a Social Media Models is a Nevada limited liability company. 5 12. Plaintiff Jonathan Murdough is a citizen and resident of Pennsylvania. 6 13. Defendant Facebook, Inc., is incorporated in Delaware, and its principal place of 7 business is 1 Hacker Way, Menlo Park, CA 94025. 8 FACTUAL ALLEGATIONS 9 10 14. Facebook is a Fortune 500 company that operates social media services, including the www.facebook.com and www.instagram.com websites and the WhatsApp Messenger application. 11 15. Facebook.com has 1.79 billion monthly active users.2 Users do not pay Facebook to 12 create a facebook.com account. Once a user opens a Facebook account, the account holder can, at no 13 cost, create a profile page, post content (such as photographs, videos, and links to articles), make 14 friends with other users, and view content posted by other users. 15 16. Instead of charging account holders to access facebook.com, Facebook earns most of 16 its revenue by selling advertising services. For example, in 2015, Facebook’s advertising revenue 17 was over $17.0 billion, more than 95% of its overall revenue.3 18 19 17. One type of advertising service Facebook sells is video advertisements, where advertisers can pay money to have video displayed to Facebook users. 20 18. Facebook videos (including video ad products) autoplay by default though the audio 21 remains off unless users actively turn it on.4 Facebook allows users to scroll past autoplaying videos 22 (including video advertisements) without ever turning on the sound or watching more than a few 23 seconds of the video. 24 25 26 27 28 1 Facebook, Statement of Rights and Responsibilities, https://www.facebook.com/terms (last accessed: Jan. 9, 2017). 2 Facebook, Company Info, http://newsroom.fb.com/company-info/ (last accessed: Jan. 9, 2017). 3 Facebook, Inc., Annual Report (Form 10-K) at 42 (Jan. 28, 2016). 4 Eric Blattberg, Like it or not, autoplay video won (Apr. 21, 2015), http://digiday.com/publishers/autoplay-video-beat-regular-video-sorry-guys/. 3 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 5 of 28 1 19. As part of Facebook’s video advertising services, advertisers can choose to purchase 2 advertisements with specific objectives, including maximizing “Brand awareness,” “Video views,” 3 or “Conversions.” Facebook says that a “Brand awareness” campaign will “[i]ncrease awareness for 4 your brand,” whereas a “Video views” campaign will “[g]et more people to view your video 5 content,” and a “Conversions” campaign will “[d]rive valuable actions on your website or app.” 6 20. Facebook video advertising services include marketing analytics, which enable 7 purchasers to monitor and evaluate their video advertisements’ performance. Marketing analytics 8 refers to the practice of measuring and analyzing the performance of an advertising or marketing 9 campaign using data from a variety of metrics. One of the main selling points of online advertising is 10 that it offers more detailed and closer to real time marketing analytics than traditional media (such as 11 television or radio), and online advertisement sellers such as Facebook have promoted their 12 marketing analytics as a prime reason that advertisers should purchase advertisements on their 13 platforms.5 Online advertisers do so because advertisers use these analytics to determine where to 14 spend advertising dollars and the effectiveness of the dollars spent. As an article in the Harvard 15 Business Review explains, analytics allows companies to make informed decisions about how to 16 allocate their limited marketing budgets across different mediums (such as television, YouTube, or 17 Facebook), to determine which advertising campaigns to “expand” and which to “kill,” and to 18 “readily adjust or allocate advertising in different markets on a monthly, weekly, or daily basis— 19 and, online, even from one fraction of a second to the next.”6 Accordingly, it has become a standard 20 practice in the industry for online advertisers like Facebook to include marketing analytics as part of 21 their advertising services, and online advertising purchasers expect that any advertising they 22 purchase will include marketing analytics to evaluate the advertising’s performance. 23 21. The importance of analytics is apparent in the competitive online video advertising 24 market. YouTube, LinkedIn, Twitter, and Facebook all have online video advertising offerings and 25 all emphasize the value of their analytics platforms. When Facebook was first entering the online 26 27 28 See, e.g., Yahoo, Yahoo! Gemini, https://gemini.yahoo.com/advertiser/home (“Measure the impact of your campaigns and drive better results by accessing actionable insights . . .”). 6 Id. 5 4 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 6 of 28 1 video advertising space, it knew that its analytics would be the key to its success. In a November 2 2013 private presentation to its advertising partners (“November 2013 Presentation”), Facebook 3 provided talking points on how to convince potential customers to purchase Facebook video 4 advertisements instead of other video advertisements, such as YouTube advertisements or television 5 advertisements. The presentation acknowledged that one of the weaknesses of Facebook’s video 6 advertising platform was the relatively basic level of metrics it provided to video purchasers. The 7 presentation stated, “Currently, we only report on video plays, which is a weakness compared to 8 YouTube, which reports on video views, completed views, and average duration of view. We are 9 working on building out our video insights to give advertisers a better sense for how videos are 10 performing. New video insights target launch: Q1 2014.”7 11 22. In May 2014, Facebook began providing video advertising purchasers with more 12 analytics, including video views, completed views, and average duration of view, as part of 13 Facebook’s video advertising services. Facebook told its users that the purpose of the new analytics 14 was to help video advertising purchasers “learn what’s resonating with people and determine how to 15 more effectively create and promote your videos on Facebook.”8 Facebook said its Audience 16 Retention analytics in particular, such as Average Duration of Video Viewed, would assist video 17 advertising purchasers in identifying underperforming videos and finding “the precise moment when 18 most people lost interest and stopped watching.”9 Facebook knew and admitted that “having access 19 to reliable metrics is important to the millions of partners who use our services to grow their 20 businesses.” 21 23. “Average Duration of Video Viewed,” which is the average amount of time that users 22 watched a video, is one of the most important analytics used in evaluating video advertisements’ 23 performance. Average Duration of Video Viewed is a measure of “retention,” which advertisers 24 7 25 26 27 28 Josh Constine, Leaked Facebook Video Ad Pitch Deck Reveals Plans To Steal TV and YouTube Dollars, TechCrunch (Dec. 13, 2013), https://techcrunch.com/2013/12/13/facebook-vs-tv-andyoutube. 8 Facebook, Introducing Video Metrics (May 5, 2014), https://www.facebook.com/business/news/Coming-Soon-Video-Metrics. 9 Id. 5 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 7 of 28 1 care about because the longer people watch an advertisement, the greater the advertisement’s impact 2 on the viewer.10 One study of video advertising campaigns on Facebook found that increasing 3 retention of a viewer from the 3-second mark to the 10-second mark in a video resulted in a 57 4 percent increase in ad recall, a 103 percent increase in brand awareness, and a 64 percent increase in 5 “purchase intent” (the intent to the purchase the advertised product).11 And because advertisers place 6 higher value upon video advertisements that are viewed for longer periods, they are willing to pay 7 more for such advertisements. 8 24. After Facebook announced and released its new video analytics platform in May 9 2014, Plaintiffs and putative class members purchased video advertising services from Facebook 10 with the understanding that video advertising analytics were included in the purchased advertising 11 services. 12 25. Upon introducing the new video metrics in their analytics platform, Facebook didn’t 13 disclose that its new video metrics were not audited or accredited by the Media Rating Council, the 14 marketing industry’s standard-bearer for accurate measurements. 15 26. As acknowledged by Facebook in its private November 2013 Presentation and public 16 statements, Facebook created and disseminated the new video analytics platform and its video 17 metrics to induce users to purchase Facebook’s video advertising services. 18 27. Among the new video metrics provided by Facebook were “Average Duration of 19 Video Viewed” and the “Average Percentage of Video Viewed.” Below is example of how these 20 video advertising metrics would appear on a user’s screen: 21 22 23 24 25 26 27 28 10 Facebook Business, The Value of Video for Brands (Mar. 17, 2015), https://www.facebook.com/business/news/value-of-video. 11 Id. 6 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 8 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 7 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 9 of 28 1 28. Facebook told advertising purchasers that the “Average Duration of Video Viewed” 2 was the “total time spent watching a video divided by the total number of people who have played 3 the video”12 (as a reasonable advertising purchaser would expect). Figure 1, below, depicts how 4 Facebook defined “Average Duration of Video Viewed.” 5 Figure 1: Formula for Facebook’s “Definition” of “Average Duration of Video Viewed” 6 7 8 9 29. In August 2016, Facebook disclosed in a post in its “Advertising Help Center” that its 10 “Average Duration of Video Viewed” and “Average Percentage of Video Viewed” metrics had been 11 improperly calculated. Facebook admitted that it had erroneously “*calculated* the Average 12 Duration of Video Viewed as ‘the total time spent watching a video divided by *only* the number of 13 people who have viewed a video for three or more seconds’”13 The “Average % of Video Viewed” 14 was also erroneous because Facebook calculated that metric by using the “Average Duration of 15 Video Viewed” as a calculation input. Figure 2 below, depicts how Facebook erroneously calculated 16 “Average Duration of Video Viewed.”14 17 Figure 2: Formula for How Facebook Actually Calculated “Average Duration of Video 18 Viewed” from May 2014 to September 2016 19 20 21 22 23 30. On or about September 23, 2016, David Fischer, Facebook’s vice president of business and marketing partnerships, admitted that, due to Facebook’s miscalculation, Facebook had 24 25 26 27 28 Facebook, How Is the “Average Duration of Video Viewed” Calculated?, https://www.facebook.com/business/help/community/question/?id=10104227902985423 (late accessed: Jan. 9, 2017). 13 Id. 14 Id. 12 8 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 10 of 28 1 misrepresented its “Average Duration of Video Viewed” metric. He wrote that the “average duration 2 of video viewed … metric should have reflected the total time spent watching a video divided by the 3 total number of people who played the video. But it didn’t – it reflected the total time spent watching 4 a video divided by only the number of ‘views’ of a video (that is, when the video was watched for 5 three or more seconds). And so the miscalculation overstated this metric.”15 6 31. Facebook’s “Average Duration of Video Viewed” metric was inflated because when 7 calculating the average, Facebook included the total time spent by under-3-second viewers in 8 watching the video in the numerator of the fraction, but excluded these under-3-second views from 9 the denominator of the fraction. Thus, Facebook’s “Average Duration of Video Viewed” metric 10 failed to align its criteria for eligible viewing time (the numerator) and its criteria for an eligible 11 view (the denominator). 12 32. The impact of the numerator-denominator mismatch can be seen through the 13 following illustration. 14 Figure 3: Example - How Duration Metric Should Have Been Calculated, According to 15 Facebook’s Definition, For Video Viewed By Four Users for 1, 2, 4, and 9 seconds respectively. 16 17 18 Figure 4: Example - How Duration Metric Was Actually Calculated, From May 2014 to 19 September 2016, For Video Viewed By Four Users for 1, 2, 4, and 9 seconds respectively. 20 21 22 33. Figure 3 demonstrates how Facebook should have measured average duration of 23 video viewed according to its definition. By using the correct method, all four viewers are included 24 in the denominator, yielding an average duration of video viewed of 4 seconds in this hypothetical. 25 Figure 4 demonstrates what Facebook actually did by excluding the video-watches under 3 seconds 26 from the denominator. In this hypothetical, 2 of the 4 video-watches are dropped from the 27 15 28 David Fischer, Facebook Video Metrics Update, Facebook.com (Sept. 23, 2016) https://www.facebook.com/business/news/facebook-video-metrics-update. 9 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 11 of 28 1 denominator, causing the average duration of video viewed to incorrectly rise to 8 seconds. 2 34. By including the time spent watching the videos by users who viewed the videos for 3 less than 3 seconds, while excluding those “under 3-second” users from the denominator, the 4 Average Duration of Video Viewed metric not only failed to reflect its stated definition, it also 5 created a highly misleading result that favored Facebook. 6 35. Facebook informed some of its advertisers that it inflated the “Average Duration of 7 Video Viewed” by between 60 and 80%.16 The inflated metric thus made video advertisements 8 appear as if they were performing much better on Facebook than they actually were. 9 36. The 60 to 80% inflation in the “Average Duration of Video Viewed” metric reveals 10 that many users watch Facebook video advertisements for less than 3 seconds. When users scroll 11 right past muted video advertisements, the users have, under Facebook’s reported metric for 12 “Average Duration of Video Viewed,” technically spent time watching the video advertisement. 13 Thus, many of the 1-2 second video-watches (that contributed to the numerator, but not the 14 denominator, of Average Duration of Video Viewed) reflect users quickly scrolling past muted auto- 15 playing video advertisements. 16 37. Video advertising purchasers, including Plaintiffs, viewed the “Average Duration of 17 Video Viewed” and “Average % of Video Viewed” as important metrics because users are more 18 likely to remember a video advertisement and be affected by it if they watch a longer portion of the 19 advertisement. 20 38. Facebook’s misrepresentations induced video advertising purchasers, including 21 Plaintiffs, to purchase video advertisements because purchasers wanted accurate video advertising 22 metrics regarding “Average Duration of Video Viewed” and “Average % of Video Viewed” so that 23 they could monitor their video advertisements’ performance. 24 25 39. Facebook’s misrepresentations induced video advertising purchasers, including Plaintiffs, to continue purchasing video advertisements, and to purchase additional video 26 27 28 16 Suzanne Vranica and Jack Marshall, Facebook Overestimated Key Video Metric for Two Years, Wall Street Journal (Sept. 22, 2016, 7:29 PM), http://www.wsj.com/articles/facebook-overestimatedkey-video-metric-for-two-years-1474586951 10 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 12 of 28 1 advertisements, because purchasers believed that users were watching their videos, on average, for 2 longer than users were actually watching their videos. 3 40. Facebook’s misrepresentations induced video advertising purchasers, including 4 Plaintiffs, to pay more for Facebook video advertising than they otherwise would have been willing 5 to pay. 6 41. Facebook’s misrepresentations thereby distorted the market price for its video 7 advertising by artificially increasing the price of Facebook video advertising, causing video 8 advertising purchasers, including Plaintiffs, to pay more than they otherwise would have paid. 9 10 42. Facebook’s misrepresentations provided Facebook with an unfair competitive advantage over other online video advertising platforms, such as YouTube, LinkedIn, and Twitter. 11 43. Facebook’s misrepresentations interfered with Plaintiffs and other putative class 12 members’ attempts to utilize Facebook’s video advertising analytics, and to run effective video 13 advertising campaigns. 14 44. Facebook’s overstated audience retention metrics resulted directly from Facebook’s 15 failure to properly audit and verify the accuracy of its video-advertising metrics before disseminating 16 them. 17 45. Any reasonable audit or verification process would have alerted Facebook earlier that 18 its metrics for “Average Duration of Video Viewed” and “Average % of Video Viewed” were 19 inaccurate and inflated. 20 46. Facebook’s audit and verification failures have been systemic and continuous, and 21 Facebook has continually refused to reform its conduct. By failing to adopt reasonably reliable 22 auditing or third party verification measures, Facebook has put Plaintiffs and other advertisers at 23 continued risk of being subjected to inaccurate analytics in the future and also made it so they will be 24 unable to rely on the accuracy of Facebook’s analytics. 25 47. In the months surrounding the announcement of the inflated and inaccurate “Average 26 Duration of Video Viewed” and “Average % of Video Viewed” metrics, Facebook’s inadequate 27 audit and verification measures caused approximately eleven additional metrics errors. The errors 28 have been so numerous and unabated that a major marketing publication posted a news article last 11 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 13 of 28 1 year to keep a running tab on the errors.17 The article begins: 2 To paraphrase J. Cole, Facebook messes up on its math one time; shame on Facebook. Facebook messes up on its math a second time; shame on those of us blindly trusting Facebook’s math. Facebook messes up on its math a third time; time to make a list. 3 4 To help myself and anyone else keep track of Facebook’s measurement errors, here’s an itemized list of each error Facebook has announced to date…. 5 6 Each of the 12 reported errors was discovered within a period of just over one year (from September 7 2016 through November 2017). The errors all pertain to the data and analytics that Facebook 8 provides to advertisers about the efficacy of advertising on Facebook, including the reach of 9 Facebook Page posts, video advertisement completion rate (which is a separate error from the one 10 affecting the two video advertising metrics that are the subject of this Complaint), average time spent 11 on Facebook Instant Articles, Facebook’s “referrals” metric, and video carousel ad link click 12 quantities. 13 48. Many individuals with considerable experience and knowledge about advertising on 14 Facebook have attributed the inflated and inaccurate “Average Duration of Video Viewed” and 15 “Average % of Video Viewed” metrics or the pattern of errors to Facebook’s failure to conduct 16 audits and verifications of its data and its advertising metrics. 17 49. Keith Weed, chief marketing officer of Unilever, said in an interview in 2015, that 18 allowing tech companies to block third parties from measuring their platforms is equivalent to 19 “letting them mark their own homework.”18 20 50. After the news first broke about the inflated and inaccurate “Average Duration of 21 Video Viewed” and “Average % of Video Viewed” metrics, Ad buying agency Publicis Media 22 wrote: “This once again illuminates the absolute need to have 3rd party tagging and verification on 23 24 25 17 Tim Peterson, FAQ: Everything Facebook has admitted about its measurement errors, Marketing Land (May 17, 2017 12:44 p.m.) https://marketingland.com/heres-itemized-listfacebooks-measurement-errors-date-200663 26 18 27 28 Suzanne Vranica and Jack Marshall, Facebook Overestimated Key Video Metric for Two Years, Wall Street Journal (Sept. 22, 2016, 7:29 PM), http://www.wsj.com/articles/facebookoverestimated-key-video-metric-for-two-years-1474586951 12 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 14 of 28 1 Facebook’s platform. Two years of reporting inflated performance numbers is unacceptable.”19 2 51. Also in reaction to the inflated and inaccurate “Average Duration of Video Viewed” 3 and “Average % of Video Viewed” metrics, Sir Martin Sorrell, the CEO of the world’s biggest 4 advertising group, WPP, said that the error underscored the need for a third party to oversee such 5 metrics: “We have also been calling for a long time for media owners like Facebook … not to mark 6 their own homework and release data to ComScore to enable independent evaluation. The referee 7 and player cannot be the same person.”20 8 9 10 52. The New York Times also reacted to the news of the error by writing, “Above all, it reinforces the importance of outside measurement when it comes to data from Facebook and other media companies.”21 11 53. Omnicom, another major advertising agency, had a similar response, saying: “This is 12 why we and many others have been repeatedly saying that third-party verification is a fundamental 13 requirement. Publishers should not ‘grade their own homework.’”22 14 54. Sarah Wood, co-chief executive of Unruly, the social marketing agency to video ad 15 tech company that is owned by News Corp, said that the error would not only “raise questions about 16 the effectiveness of Facebook as a video platform [but] it will raise more fundamental questions 17 around trustworthiness and highlights just why third-party verification is so critical . . . it’s time the 18 walled gardens came crashing down.”23 19 55. Also in response to news of the error, Fortune quoted Shafqat Islam, co-founder of 20 21 19 22 23 24 25 26 27 28 Id. Eric Oster, Sir Martin Sorrell Has Harsh Words for Facebook’s Fake Data in ‘Overstategate’, AgencySpy (Sept. 23, 2016 11:32 a.m.) http://www.adweek.com/agencyspy/sirmartin-sorrell-has-harsh-words-for-facebooks-fake-data-in-overstategate/117517 21 John Herrman & Sapna Maheshwari, Facebook Apologizes for Overstating Video Metrics (Sept. 23, 2016) https://www.nytimes.com/2016/09/24/business/media/facebook-apologizes-foroverstating-video-metrics.html 22 Id. 23 Matthew Garrahan & Lauren Fedor, Ad groups seize on Facebook’s video measuring error, Financial Times (Sept. 23, 2016) https://www.ft.com/content/230ec388-8169-11e6-8e508ec15fb462f4 20 13 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 15 of 28 1 NewsCred: “Need 3rd party verification everywhere.”24 2 56. After Facebook continued to disclose additional errors, The Wall Street Journal 3 reported that “The lack of fully independent third-party ad verification was a major sticking point 4 following Facebook’s recent video counting mishap…”25 In the same article, it was noted that 5 Facebook was resisting letting third parties use technology to evaluate advertising efficacy. GroupM, 6 which is the ad buying unit of WPP, was quoted as saying that a transition to “full third-party 7 measurement” remained “crucial.” And the article reports that the Association of National 8 Advertisers publicly advocated for an MRC audit. 9 NEW ALLEGATIONS 10 11 57. Facebook’s internal documents show that Facebook personnel had been alerted to the audience retention metrics mismatch long before news broke in the Wall Street Journal. 12 58. In 2015, several advertisers had inquired about “Average % of Video Viewed” 13 numbers that appeared exorbitant—some were displayed at 100%. As Facebook would put it later, 14 complaints about the inflated metrics had “been filed for a long time and we didn’t fully 15 investigate.” Other Facebook personnel acknowledged internally that one of the things they learned 16 from the fallout was to “[b]e more serious on publishers/advertisers’ reports.” 17 59. In July 2015, in an internal response to one such inquiry, a Facebook engineer 18 discussed the numerator/denominator mismatch: “I remember [another Facebook engineer] 19 mentioned when computing the average, we only consider views greater than 3second [sic], but use 20 the total watch time (including those under 3s).” This mismatch caused the inflation to the metrics 21 that Facebook ultimately disclosed in the fall of 2016. Still, Facebook did not correct the inflated 22 metrics. 23 24 60. In addition to overt discussion of the discrepancy, Facebook repeatedly and recklessly failed to check its own work. It has long been known that Facebook resisted outside help—such as 25 26 27 28 Michelle Toh, Why Facebook’s Ad Buyers Are Not Happy Right Now (Sept. 23, 2016) http://fortune.com/2016/09/23/facebook-video-view-metric/ 25 Mike Shields, Facebook Says It Found More Miscalculated Metrics (Nov. 16, 2016 10:03 a.m.) https://www.wsj.com/articles/facebook-says-it-found-more-miscalculated-metrics-1479303984 24 14 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 16 of 28 1 through third party verification or auditing—to ensure the accuracy of the metrics, but Facebook’s 2 internal efforts behind the scenes reflect a company mentality of reckless indifference toward the 3 accuracy of its metrics. Facebook’s internal documents show the company had never performed a full audit of 4 61. 5 its video metrics. 6 62. Facebook’s management also severely understaffed the Video Insights team, which 7 was responsible for investigating and fixing errors. As of September 2016, that team had dwindled 8 to only two engineers. 9 63. In June 2016, a Facebook engineering manager finally followed up on advertiser 10 complaints dating back to early 2015, writing that “[s]omehow there was no progress on the task for 11 a year.” But even once it was decided to take action on the metrics, Facebook did not promptly fix 12 its calculation or disclose that the calculation was wrong. Instead, it continued reporting 13 miscalculated viewership metrics for another several months, as it developed a “no PR” strategy to 14 avoid drawing attention to the error. The company decided to “obfuscate the fact that we screwed up 15 the math” by quietly retiring the erroneous metrics and replacing them with corrected metrics under 16 a new name. For instance, Average Duration of Video Viewed would be replaced with Average 17 Watch Time. 18 64. In August 2016, Facebook began reaching out privately to select, large advertisers, 19 telling them that Facebook had “recently discovered a discrepancy” in the video ad average view 20 metrics. Facebook pushed that message even as personnel internally emphasized that “we didn’t 21 recently discover a discrepancy.” (emphasis added). 22 65. Facebook had successfully used a “no PR” approach before with these same metrics. 23 In January 2015, Facebook discovered it had been over-reporting the Average Duration of Video 24 Viewed and Average % Percentage of Video Viewed metrics by 40% (due to reasons separate from 25 the numerator/denominator mismatch). But rather than correct the metrics immediately in January 26 2015 and tell advertisers, Facebook recognized that “a 40% drop will be too significant and might 27 hurt user trust.” It therefore decided to do a “slow launch to fix” the metric without telling anyone: 28 “so advertisers … won’t notice significant changes.” 15 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 17 of 28 1 66. This time, the impact of Facebook’s error was much larger. Facebook chose not to 2 correct the Wall Street Journal’s report that its miscalculation had only inflated average viewership 3 metrics by between 60 and 80%. In fact, they had inflated the metrics by some 150 to 900%. For 4 example, when Facebook summarized the impact of the miscalculation internally it presented two 5 typical cases: in one, Facebook had inflated Average Duration of Video Viewed from 2.0 seconds to 6 17.5 seconds (an increase of 775%); in the other, Facebook had inflated Average Duration of Video 7 Viewed from 2.4 seconds to 17.3 seconds (an increase of 621%). 8 9 67. If Facebook had immediately corrected its miscalculation in a straightforward manner, advertisers would have seen a sudden and precipitous drop in their viewership metrics. 10 They also would have seen that they had spent considerable money on past advertising campaigns 11 that were not enjoying anywhere near the average viewership that Facebook had represented. 12 68. Facebook did not wish to draw scrutiny to its viewership figures because it knows 13 that the majority of video ads on its platform are viewed for very short periods of time — users 14 scroll right past. If advertisers were more widely aware of this fact, and in particular, if they knew 15 that their advertisements were among those that were not drawing viewers’ attention, they would be 16 less likely to continue buying video advertising from Facebook. 17 69. Facebook’s “no PR” approach foundered in September 2016, when Facebook learned 18 that the Wall Street Journal would be publishing an article about its inflated metrics. Facebook then 19 continued to support a misleading narrative. The Journal quoted a Facebook statement claiming that 20 Facebook had only “recently discovered an error in the way we calculate one of our video metrics.” 21 Facebook reiterated that notion in Mr. Fischer’s September 23rd blog post, which stated that 22 Facebook had discovered the error “[a]bout a month ago,” and that “[a]s soon as we discovered the 23 discrepancy, we fixed it.” And of course, Facebook allowed the 60 – 80% viewership inflation 24 range to go uncorrected even though the true level of inflation had been some 10 times higher. PLAINTIFFS’ EXPERIENCES 25 26 a. LLE One, LLC 27 70. LLE One, LLC, is a social media marketing agency. 28 71. During the period between May 4, 2014, and September 23, 2016, when Facebook 16 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 18 of 28 1 was providing inflated “Average Duration of Video Viewed” and “Average Percentage of Video 2 Viewed” metrics, LLE One, LLC, purchased video advertising services from Facebook. 3 72. Before LLE One, LLC, purchased video advertising services, LLE One, LLC, knew 4 and expected that as part of the video advertising services, Facebook provides metrics designed to 5 enable advertisers to monitor their video advertisements’ performance. 6 73. After first purchasing video advertising services, LLE One, LLC, began viewing the 7 video metrics provided by Facebook, including the “Average Duration of Video Viewed” and 8 “Average % of Video Viewed” metrics. 9 74. The inflated “Average Duration of Video Viewed” and “Average % of Video 10 Viewed” metrics were material to LLE One, LLC, and LLE One, LLC, relied on the metrics, 11 purchasing more video advertising services from Facebook than it otherwise would have. 12 13 75. Since September 23, 2016, LLE One, LLC, has continued to purchase Facebook video advertising services and it intends to continue purchasing Facebook video advertising services. 14 b. Jonathan Murdough 15 76. Jonathan Murdough (“Murdough”) is a citizen and resident of Pennsylvania. 16 77. During the period between May 4, 2014, and September 23, 2016, when Facebook 17 was providing inflated “Average Duration of Video Viewed” and “Average Percentage of Video 18 Viewed” metrics, Murdough purchased video advertising services from Facebook. 19 78. Before Murdough purchased video advertising services for himself, he viewed video 20 metrics provided by Facebook to a company that he worked for, including the “Average Duration of 21 Video Viewed” and “Average % of Video Viewed” metrics. 22 79. Before Murdough purchased video advertising services, Murdough knew and 23 expected that as part of the video advertising services, Facebook provides metrics designed to enable 24 advertisers to monitor their video advertisements’ performance. 25 80. After purchasing video advertising services for himself, Murdough viewed video 26 metrics provided by Facebook, including the “Average Duration of Video Viewed” and “Average % 27 of Video Viewed” metrics. 28 81. The inflated “Average Duration of Video Viewed” and “Average % of Video 17 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 19 of 28 1 Viewed” metrics were material to Murdough, and Murdough relied on the metrics, purchasing more 2 video advertising services from Facebook than he otherwise would have. 3 82. Murdough intends to continue purchasing Facebook video advertising services. 4 5 CLASS ALLEGATIONS 6 7 8 9 83. Plaintiffs re-allege and incorporate by reference herein all of the allegations contained 84. Pursuant to the Federal Rule of Civil P. 23(b)(3), Plaintiffs assert claims on behalf of above. the following “Class”: All persons or entities who, from May 4, 2014, to September 23, 2016 (“Class 10 Period”), had an account with Facebook, Inc., and who paid for placement of video advertisements 11 on a Facebook-owned website. Excluded from the Class are Defendant, any entity in which 12 Defendant has a controlling interest, and Defendant’s officers, directors, legal representatives, 13 successors, subsidiaries, and assigns. Also excluded from the Class is any judge, justice, or judicial 14 officer presiding over this matter and the members of their immediate families and judicial staff. 15 85. This action has been brought and may properly be maintained as a class action as it 16 satisfies the numerosity, commonality, typicality, adequacy, and superiority requirements of Rule 17 23(b)(3). Plaintiffs seek to represent an ascertainable Class, as determining inclusion in the class can 18 be done through Facebook’s own records. 19 86. Plaintiffs reserve the right to amend the Class definition if discovery and further 20 investigation reveal that the Class should be expanded, divided into subclasses, or modified in any 21 other way. 22 87. Although the precise number of Class members is unknown and can only be 23 determined through appropriate discovery, Plaintiffs believe, and on that basis allege, that the 24 proposed Class is so numerous that joinder of all members would be impracticable as Facebook sells 25 millions of advertisements annually, and a significant portion of those advertisements are video 26 advertisements. 27 28 88. Questions of law and fact common to the putative Class exist that predominate over questions affecting only individual members, including inter alia: 18 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 20 of 28 1 a. Whether Facebook made material misrepresentations about its video advertising 2 services, including misrepresenting the “Average Duration of Video Viewed” and 3 “Average % of Video Viewed” metrics; b. Whether Facebook’s use of inaccurate, unaudited, and unverified video metrics was 4 5 likely to deceive members of the public and thus constituted a fraudulent business 6 practice under California’s Unfair Competition Law (Cal. Bus. & Prof. Code § 7 17200); c. Whether Facebook’s failure to properly audit and verify its video metrics was 8 9 unethical, unscrupulous, or substantially injurious to video advertising purchasers and 10 thus constituted an unfair business practice under California’s Unfair Competition 11 Law; and 12 d. Whether Facebook breached its contractual duty to perform competently and with 13 14 reasonable care by reporting erroneous video metrics. 89. Plaintiffs are members of the putative Class. The claims asserted by the Plaintiffs in 15 this action are typical of the claims of the members of the putative Class, as the claims arise from the 16 same course of conduct by the Defendant and the relief sought is common. 17 90. Plaintiffs will fairly and adequately represent and protect the interests of the members 18 of the putative Class, as their interests are coincident with, not antagonistic to, the other members of 19 the Class. 20 91. Plaintiffs have retained counsel competent and experienced in both consumer 21 protection and class action litigation. Plaintiffs’ counsel specifically has experience litigating some 22 of the largest and most complex consumer class actions, including numerous consumer class actions 23 in the Northern District of California. 24 92. Certification of the Class is appropriate pursuant to Fed. R. C. P. 23(b)(3) because 25 questions of law or fact common to the respective members of the Class predominate over questions 26 of law or fact affecting only individual members. This predominance makes class litigation superior 27 to any other method available for the fair and efficient adjudication of these claims including 28 consistency of adjudications. Absent a class action it would be highly unlikely that the members of 19 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 21 of 28 1 the Class would be able to protect their own interests because the cost of litigation through 2 individual lawsuits might exceed the expected recovery. 3 93. A class action is a superior method for the adjudication of the controversy in that it 4 will permit a large number of claims to be resolved in a single forum simultaneously, efficiently, and 5 without the unnecessary hardship that would result from the prosecution of numerous individual 6 actions and the duplication of discovery, effort, expense, and the burden of the courts that individual 7 actions would create. 8 94. In the alternative, the Class should be certified because: 9 a. The prosecution of separate actions by the individual members of the proposed class 10 would create a risk of inconsistent adjudications, which could establish incompatible 11 standards of conduct for Facebook; 12 b. The prosecution of individual actions could result in adjudications, which as a 13 practical matter, would be dispositive of the interests of non-party class members or 14 which would substantially impair their ability to protect their interests; and 15 c. Facebook has acted or refused to act on grounds generally applicable to the proposed 16 Class, thereby making appropriate final and injunctive relief with respect to the 17 members of the proposed Class as a whole. 18 FIRST CAUSE OF ACTION (On behalf of the Class)26 CALIFORNIA UNFAIR COMPETITION LAW, CAL. BUS. & PROF. CODE § 17200, et seq. 19 20 21 22 95. Plaintiffs re-allege and incorporate by reference herein all of the allegations contained 96. Facebook violated California’s Unfair Competition Law (UCL), Cal. Bus. & Prof. above. 23 24 Code §17200 et seq., by engaging in the fraudulent and unfair business acts and practices alleged 25 previously, and as further specified below. 26 27 28 Pursuant to Facebook’s terms of service, the laws of the State of California govern “any claim” that arises between Facebook and its users, “without regard to conflict of law provisions.” Facebook, Statement of Rights and Responsibilities, https://www.facebook.com/terms (last accessed: Jan. 9, 2017). 20 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW 26 Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 22 of 28 1 97. Facebook’s dissemination of inaccurate and inflated video-advertising metrics 2 constitutes a fraudulent practice under the UCL, as it is likely to deceive Class members into 3 believing that their video advertisements were viewed, on an average, for a longer duration than they 4 actually were viewed. 5 98. Facebook’s failure to properly audit and verify the accuracy of its video-advertising 6 metrics before disseminating them to Class members is unethical, unscrupulous, and substantially 7 injurious to video-advertising purchasers, and thus constitutes an unfair practice under the UCL. 8 Facebook’s practice was also contrary to legislatively declared and public policies that seek to 9 protect consumers from misleading statements, as reflected by laws like the Federal Trade 10 Commission Act (15 U.S.C. § 45), Consumers Legal Remedies Act (Cal. Civ. Code § 1750 et seq.), 11 and California False Advertising Law (Cal Bus. & Prof. Code § 17500). The harm these practices 12 caused to Plaintiffs and the Class members outweigh their utility, if any. 13 99. Facebook should have known that its metrics for “Average Duration of Video 14 Viewed” and “Average % of Video Viewed” were inaccurate and inflated, and had Facebook 15 properly audited and verified its video-advertising metrics it would have known that those metrics 16 were inaccurate and inflated. The calculation errors that Facebook allowed to persist for over two 17 years were obvious errors that would have been discovered by a reasonable auditing and verification 18 process. 19 100. Facebook’s failure to employ reasonable auditing and verification procedures gave it 20 an unfair competitive advantage, as it allowed Facebook to provide video-advertising services at a 21 lower cost and made those advertising services appear to be more effective than they were. 22 101. Plaintiffs have standing to bring these claims under the UCL because they were 23 injured and lost money or property, including but not limited to money paid for Facebook video 24 advertisements, as a result of Facebook’s fraudulent and unfair business practices. Among other 25 things, Plaintiffs would not have bought as much video-advertising services if Facebook had not 26 disseminated inflated metrics and would have paid a lower price for the video-advertising services 27 they did purchase. 28 102. Pursuant to Cal. Bus. & Prof. Code § 17203, Plaintiffs seek equitable relief to prevent 21 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 23 of 28 1 the continued use of Facebook’s unfair and fraudulent practices and to restore to the Class all money 2 Facebook may have acquired by means of its fraudulent and unfair business practices. 3 5 SECOND CAUSE OF ACTION (On behalf of the Class) BREACH OF IMPLIED DUTY TO PERFORM WITH REASONABLE CARE 6 103. Plaintiffs re-allege and incorporate by reference herein all of the allegations contained 104. Plaintiffs and Class members contracted with Facebook to provide them with video- 4 7 8 9 above. advertising services. They did so through one or more of Facebook’s advertising interfaces, where 10 Class members can submit video advertisements for approval, set a daily advertising budget that can 11 be adjusted at any time, set or adjust their target audience, and set or adjust the placement of the 12 advertising. 13 105. Plaintiffs and Class members met all or substantially all of their contractual 14 obligations, including submitting their advertising for Facebook’s approval and paying for 15 Facebook’s advertising services. 16 106. There is not one integrated contract that spells out Facebook’s obligations to the 17 Class, but those obligations can be determined by reference to Facebook’s course of dealing with the 18 Class, industry practice, and from various webpages created by Facebook, including Facebook’s 19 Statement of Rights and Responsibilities (https://www.facebook.com/legal/terms); Facebook’s Self- 20 Serve Ad Terms (https://www.facebook.com/legal/self_service_ads_terms); Facebook’s Advertising 21 Guidelines (https://www.facebook.com/policies/ads); Facebook’s Advertiser Help Center 22 (https://www.facebook.com/business/help); Facebook’s Community Payments Terms 23 (https://www.facebook.com/payments_terms); and Facebook’s ad-creation, editing, and management 24 interfaces, which may be accessed through Facebook’s Ads Manager 25 (https://www.facebook.com/ads/manager/creation/), Facebook’s Power Editor 26 (https://www.facebook.com/ads/manage/powereditor), Facebook’s Ads Manager App, Facebook 27 Pages, and Facebook Ads API. 28 107. One of Facebook’s obligations is to provide Class members with an advertising 22 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 24 of 28 1 interface: a central portal where Class members can create, edit, and monitor advertisements for 2 placement on Facebook’s websites. This obligation is implied by Facebook’s course of dealing with 3 the Class, industry practice, Facebook’s Self-Serve Ad Terms, and Facebook’s Advertising Help 4 Center. For example, the Self-Serve Ad Terms were drafted by Facebook to “apply to [Class 5 members’] use of the self-service advertising interfaces and APIs … (collectively, the ‘Self-Serve 6 Ad Interfaces’),” and those terms state that “you can cancel your Order at any time through our 7 online portal”—the implication being that there would in fact be advertising interfaces for Class 8 members to use. Facebook’s Advertiser Help Center also refers to Facebook’s advertising interfaces 9 on several occasions. 10 108. Another of Facebook’s obligations is to track the advertising it places for Plaintiffs 11 and Class members and report metrics by which Plaintiffs and Class members can evaluate the 12 performance of their advertisements. This obligation is implied by Facebook’s course of dealing 13 with the Class, industry practice, statements made by Facebook in the Advertiser Help Center and in 14 standardized communications with the Class. For example, Facebook has stated in its Advertiser 15 Help Center: “You can always go to Ads Manager to see how your ads are performing,” and “When 16 you create a Video Views campaign, we’ll measure both impressions and engagement with your 17 videos.” And when Facebook approves an advertisement, it sends a standardized email that states: 18 “Your ad is approved and should begin delivering shortly. Click the ad name below to manage it or 19 view its performance.” 20 109. Under California law, Facebook was required to perform its contractual obligations 21 competently and with reasonable care. Facebook breached that duty by incorrectly measuring 22 viewer engagement with the advertising it placed for Plaintiffs and Class members; by including 23 inaccurate data in the advertising interface that it provided to Plaintiffs and Class members; and by 24 reporting inaccurate advertising-performance metrics to Plaintiffs and Class members. 25 110. Had Facebook used reasonable care, it would have calculated the video-viewing 26 metrics correctly, it would have discovered any calculation errors through routine auditing and 27 verification, and it would not have used an incorrect formula for over two years. 28 111. As a result of Facebook’s failure to provide its agreed advertising services 23 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 25 of 28 1 competently and using reasonable care, Plaintiffs and Class members purchased advertising services 2 they would not otherwise have purchased and failed to receive the benefit of their bargain. They are 3 entitled to damages in an amount to be proven at trial. 4 5 112. and given it an opportunity to remedy the harm suffered by the Class, but Facebook has refused. 6 THIRD CAUSE OF ACTION (On behalf of the Class) FRAUD 7 8 9 10 On behalf of the proposed Class, Plaintiffs have notified Facebook that it is in breach 113. Plaintiffs re-allege and incorporate by reference herein all of the allegations contained 114. Facebook falsely represented the average time that Facebook users spent watching above. 11 Plaintiffs’ and Class members’ video advertisements. The Average Duration of Video Viewed and 12 Average Percentage of Video Viewed metrics that Facebook reported to Plaintiffs and Class 13 members was typically inflated by between 150% and 900%. 14 115. Facebook either knew that the average viewership metrics it was reporting to 15 Plaintiffs and Class members was false or reported those metrics recklessly and without regard for 16 their truth. Facebook engineers knew about the miscalculation by July 2015. Facebook received 17 numerous reports from advertisers (such as nonsensical 100% viewership rates) over the course of 18 many months, and still continued to report false metrics. And even after Facebook ultimately 19 decided to correct its metrics, it still knowingly disseminated false metrics for several additional 20 months. The persistence of Facebook’s false metrics was possible only because Facebook did not 21 take verification of its metrics seriously, severely understaffed the engineering team in charge of 22 fixing errors, did not fully investigate or correct errors that were reported to it, and refused to allow 23 third-party verification of its metrics. 24 116. Facebook intended that Plaintiffs and Class members rely on its average viewership 25 metrics. Facebook promoted the metrics to users as useful tools to identify which video 26 advertisements were effective; Facebook knew that the average viewership metrics were considered 27 key metrics by advertisers; and Facebook concealed the fact that its recklessness had caused the 28 metrics to be inflated because it knew that would hurt user trust and result in advertisers purchasing 24 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 26 of 28 1 2 fewer video ads. 117. Plaintiffs and Class members did rely on Facebook’s inflated viewership metrics 3 when deciding whether and how to purchase video advertising from Facebook. As a result of the 4 inflated metrics, Plaintiffs and Class members purchased more video advertising from Facebook 5 than they otherwise would have and paid a higher price than they otherwise would have. 6 118. Plaintiffs seek an award of compensatory and punitive damages. Facebook’s conduct 7 as previously described constitutes oppression, fraud, or malice, and was authorized or ratified by 8 Facebook’s officers. 9 10 11 12 PRAYER FOR RELIEF WHEREFORE, Plaintiffs LLE One, LLC, d/b/a Crowd Siren and d/b/a Social Media Models, and Jonathan Murdough, on behalf of themselves and the Class, seek the following relief: A. An order certifying this action as a class action under Fed. R. Civ. P. 23, defining the 13 Class as requested herein, appointing Cohen Milstein Sellers & Toll PLLC, Gibbs Law Group LLP, 14 and Eglet Prince as Class Counsel, and finding that Plaintiffs are proper representatives of the Class 15 requested herein. 16 B. Plaintiffs request injunctive relief. Awarding injunctive and other equitable relief as 17 is necessary to protect the interests of the Class, including: (i) an order prohibiting Facebook from 18 engaging in the wrongful acts described herein; (ii) requiring Facebook to engage third-party 19 auditors to conduct audits and evaluations of Facebook’s advertising metrics on a periodic basis and 20 ordering them to promptly correct any problems or issues detected by these auditors, and (iii) 21 requiring Facebook to disclose any further inaccurate advertising metrics in a timely and accurate 22 manner. 23 C. Plaintiffs also request damages (including punitive damages), restitution, attorneys’ 24 fees, statutory costs, and such other and further relief as is just and proper. Plaintiffs seek attorneys’ 25 fees under California Code of Civil Procedure 1021.5. 26 27 JURY TRIAL DEMAND Plaintiffs hereby request a jury trial for all issues so triable. 28 25 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 27 of 28 1 2 DATED: August 3, 2018 Respectfully submitted, By: /s/ Eric Gibbs 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Eric Gibbs Dylan Hughes David Stein Aaron Blumenthal GIBBS LAW GROUP LLP 505 14th Street, Suite 1110 Oakland, CA 94612 Telephone: (510) 350-9700 Facsimile: (510) 350-9701 ehg@classlawgroup.com dsh@classlawgroup.com ds@classlawgroup.com ab@classlawgroup.com Andrew N. Friedman Geoffrey Graber Eric Kafka COHEN MILSTEIN SELLERS & TOLL PLLC 1100 New York Ave. NW, Fifth Floor Washington, DC 20005 Telephone: (202) 408-4600 Facsimile: (202) 408-4699 afriedman@cohenmilstein.com ggraber@cohenmilstein.com ekafka@cohenmilstein.com Michael Eisenkraft COHEN MILSTEIN SELLERS & TOLL PLLC 88 Pine Street, 14th Floor New York, NY 10005 Telephone: (212) 838-7797 Facsimile: (212) 838-7745 meisenkraft@cohenmilstein.com 23 24 25 26 27 28 Robert T. Eglet Robert M. Adams Erica D. Entsminger Artemus W. Ham EGLET PRINCE 400 South Seventh Street, Suite 400 Las Vegs, NV 89101 Telephone: (702) 450-5400 Facsimile: (702) 450-5451 26 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW Case 4:16-cv-06232-JSW Document 165 Filed 10/16/18 Page 28 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 eservice@egletlaw.com Aisha Christian (pro hac vice forthcoming) 129 West 27th Street, 11th Floor New York, NY 10001 Telephone: (646) 285-2029 Charles Reichmann (SBN 206699) LAW OFFICES OF CHARLES REICHMANN 16 Yale Circle Kensington, CA 94708-1015 Telephone: (415) 373-8849 charles.reichmann@gmail.com Joseph A. Motta (SBN 133531) RUEB & MOTTA 1401 Willow Pass Road, Suite 880 Concord, CA 94520 Telephone: (925) 602-3400 Facsimile: (925) 602-0622 joe@rmmprolaw.com Counsel for Plaintiffs and Proposed Class 15 16 17 18 19 20 21 22 23 24 25 26 27 28 27 FOURTH AMENDED CLASS ACTION COMPLAINT Case No.: 4:16-cv-06232-JSW