Epic Games, Inc. v. Google LLC et al

Filing 674


Download PDF
1 2 3 4 UNITED STATES DISTRICT COURT 5 NORTHERN DISTRICT OF CALIFORNIA 6 IN RE GOOGLE PLAY STORE ANTITRUST LITIGATION 7 8 MDL Case No. 21-md-02981-JD Member Case No. 20-cv-05671-JD ORDER RE GOOGLE’S RENEWED MOTION FOR JUDGMENT AS MATTER OF LAW OR FOR NEW TRIAL IN EPIC CASE 9 10 United States District Court Northern District of California 11 12 13 After 15 days of trial, a jury found in favor of plaintiff Epic Games Inc. on its antitrust 14 claims against Google. See Dkt. No. 866.1 Google had moved for judgment as a matter of law at 15 an appropriate stage of the trial, which the Court denied. Dkt. Nos. 825, 831. Google renewed the 16 motion post-verdict under Federal Rule of Civil Procedure 50(b), with a motion in the alternative 17 for a new trial under Rule 59. Dkt. No. 925. The Court denied both motions. Dkt. No. 951. This 18 order provides a detailed explanation for that decision. 19 BACKGROUND 20 The Court presented the background of this multidistrict antitrust litigation in other orders. 21 See, e.g., Dkt. Nos. 383, 588. In pertinent part, Epic is a well-known video game and software 22 developer, and its apps include Fortnite, a popular online game. Fortnite can be played on a 23 variety of consoles and devices, including smartphones running the Android mobile operating 24 system. 25 26 Epic distributed a Fortnite Android app through the Google Play Store for a handful of months starting in April 2020, until Epic’s relationship with Google broke down in August 2020. 27 28 1 All docket number references are to the ECF docket for In re Google Play Store Antitrust Litigation, Case No. 21-md-02981-JD. 1 A particular sticking point was Epic’s objection to Google’s requirement that Epic use Google’s 2 billing system and pay Google a 30% fee on all in-app purchases made by Fortnite users. Epic 3 wanted to use its own in-app payment solution and not pay Google a 30% cut, which Google 4 refused to allow. Epic then deployed a “hotfix,” which was in effect a covert app update that 5 allowed Fortnite users to use Epic’s payment system. Google responded by removing Fortnite 6 from the Google Play Store. United States District Court Northern District of California 7 On the day that Fortnite was removed from the Google Play Store, Epic filed this lawsuit 8 against Google LLC and certain of its affiliates alleging that Google had engaged in 9 anticompetitive conduct in violation of the antitrust laws in connection with the Google Play 10 Store. Dkt. No. 1. As alleged in its second amended complaint (SAC), Epic presented claims 11 under Sections 1 and 2 of the Sherman Act, 15 U.S.C. §§ 1, 2; the California Cartwright Act, Cal. 12 Bus. & Prof. Code § 16700 et seq.; and the California Unfair Competition Law (UCL), Cal. Bus. 13 & Prof. Code § 17200 et seq. Dkt. No. 378. Epic sought injunctive relief only, and no monetary 14 damages. Id. Google filed counterclaims against Epic, including for breach of the Google Play 15 Developer Distribution Agreement (DDA). Dkt. No. 386. 16 Epic’s lawsuit was consolidated into a multidistrict litigation action along with similar 17 antitrust complaints filed by Google Play Store users and developers, and the attorneys general of 18 many states. A substantial period of litigation ensued for all of the member cases, and several 19 important issues were resolved in the pretrial stage. One was a determination that Google had 20 willfully failed to preserve relevant, substantive business communications that were made by 21 employees on the Google Chat system. This determination required an extensive inquiry by the 22 Court that culminated in an evidentiary hearing featuring witness testimony and documents, and 23 extensive findings of fact. See Dkt. No. 469. Testimony at trial adduced even more troubling 24 evidence of improper assertions of the attorney-client privilege by Google’s employees, including 25 its CEO, to keep communications secret, and a widespread understanding within the company that 26 discussions of sensitive topics should be done in a way that evaded preservation. See, e.g., Trial 27 28 2 1 Tr. at 964:21-23, 991:16-992:8, 1075:20-1076:12, 1321:17-24.2 Another important pretrial 2 determination was whether Google could evade a jury altogether by asking for a bench trial at the 3 very last moment. The Court concluded, based on Google’s own conduct, that it had consented to 4 a jury trial. See id. at 6:13-7:16. In time, the other cases went into settlement proceedings. Epic’s case was tried by a jury United States District Court Northern District of California 5 6 of nine citizens in November and December 2023. The parties put on forty-five witnesses, 7 including nine expert witnesses, over the course of fifteen days of testimony. More than three 8 hundred documents were admitted into evidence. See Dkt. Nos. 622, 623, 624. The final jury 9 instructions totaled fifty-five pages. Dkt. No. 850. The instructions were based on extensive 10 discussions with, and submissions by, the parties. See, e.g., Dkt. Nos. 487, 528, 554, 564, 847, 11 848, 849. At the conclusion of deliberations, the jury returned a unanimous verdict in favor of Epic. 12 13 Dkt. No. 866. For the monopolization claim under Section 2 of the Sherman Act, the jury found 14 that Epic had proved two relevant product markets: a market for the distribution of Android apps, 15 and for Android in-app billing services for digital goods and services transactions. The jury also 16 found that Epic had proved for both of these markets that the geographic scope was worldwide 17 excluding China. The jury further concluded that Epic had proved that Google willfully acquired 18 or maintained monopoly power by engaging in anticompetitive conduct in each of the product 19 markets, and that Epic had proved it was injured as a result of Google’s violation of the antitrust 20 laws. Id. at 1-4. For the unlawful restraint of trade claim under Section 1 of the Sherman Act and 21 California state law, the jury found that Epic had proved that Google entered into one or more 22 agreements that unreasonably restrained trade in the same two product markets as for the 23 24 25 26 27 28 “Trial Tr.” references are to the trial transcript, which consists of 17 volumes with 3,442 total pages that are consecutively numbered. The transcript can be found at Dkt. No. 834 (Vol. 1; pages 1-116); Dkt. No. 835 (Vol. 2; pages 117-322); Dkt. No. 836 (Vol. 3; pages 323-578); Dkt. No. 837 (Vol. 4; pages 579-788); Dkt. No. 838 (Vol. 5; pages 789-1036); Dkt. No. 839 (Vol. 6; pages 1037-1302); Dkt. No. 840 (Vol. 7; pages 1303-1539); Dkt. No. 841 (Vol. 8; pages 1540-1785); Dkt. No. 842 (Vol. 9; pages 1786-1866); Dkt. No. 843 (Vol. 10; pages 1867-2103); Dkt. No. 844 (Vol. 11; pages 2104-2291); Dkt. No. 845 (Vol. 12; pages 2292-2518); Dkt. No. 846 (Vol. 13; pages 2519-2763); Dkt. No. 847 (Vol. 14; pages 2764-2854); Dkt. No. 848 (Vol. 15; pages 28553065); Dkt. No. 849 (Vol. 16; pages 3066-3293); and Dkt. No. 867 (Vol. 17; pages 3294-3442). 3 2 1 monopolization claim. The jury determined that the illegal agreements were Google’s DDA 2 agreements; agreements with alleged competitors or potential competitors under Project Hug and 3 the Games Velocity Program; and agreements with original equipment manufacturers (OEMs) that 4 sell mobile devices, including the MADA and RSA agreements. Epic was found to have proved 5 antitrust injury from these violations. Id. at 5-6. For the tying claim under Section 1 of the 6 Sherman Act and California law, the jury determined that Epic had proved that Google unlawfully 7 tied the use of the Google Play Store to the use of Google Play Billing, and that Epic again had 8 been injured by this conduct. Id. at 7. United States District Court Northern District of California 9 Epic’s UCL claim was not presented to the jury and was reserved for the Court’s decision. 10 Google’s breach of contract counterclaim also was not presented to the jury pursuant to the 11 parties’ agreement, and the Court will decide Epic’s illegality defense, with the parties’ stipulated 12 facts to be treated as proved. See Dkt. No. 850 at 1. Also reserved for the Court’s decision is the 13 issue of an injunctive remedy under the Sherman Act and Cartwright Act in light of the jury’s 14 verdict. The remedy proceedings are currently underway. See Dkt. No. 978. 15 Google has fired a barrage of objections and allegations of error in an effort to escape the 16 judgment of the jury. This approach has been Google’s modus operandi throughout the case, and 17 often results in headline-style arguments that lack useful development. The 90 pages of objections 18 that Google filed to Epic’s proposed injunction in the remedy proceedings are the latest 19 manifestation of this problem. See Dkt. No. 958. In the ensuing discussion, the Court addresses 20 Google’s attacks on the verdict even when Google’s argument was little more than a passing 21 comment or two. 22 As the Supreme Court has observed, the “[d]etermination of whether a new trial should be 23 granted or a judgment entered under Rule 50(b) calls for the judgment in the first instance of the 24 judge who saw and heard the witnesses and has the feel of the case which no appellate printed 25 transcript can impart.” Cone v. West Virginia Pulp & Paper Co., 330 U.S. 212, 216 (1947) 26 (citation omitted). It is on this basis, and the trial record as a whole, that the Court concludes 27 Google is not entitled to undo the jury’s verdict under Rule 50(b) or Rule 59. 28 4 LEGAL STANDARDS 1 United States District Court Northern District of California 2 Under Rule 50(b), a party that has previously made a motion for judgment as a matter of 3 law during a jury trial, as Google did, may “file a renewed motion for judgment as a matter of law 4 and may include an alternative or joint request for a new trial under Rule 59.” Fed. R. Civ. P. 5 50(b). Judgment as a matter of law is appropriate when “the evidence, construed in the light most 6 favorable to the nonmoving party, permits only one reasonable conclusion, and that conclusion is 7 contrary to that of the jury.” White v. Ford Motor Co., 312 F.3d 998, 1010 (9th Cir. 2002) 8 (citation omitted); see also Lakeside-Scott v. Multnomah Cnty., 556 F.3d 797, 803 (9th Cir. 2009) 9 (“JMOL is . . . appropriate when the jury could have relied only on speculation to reach its 10 verdict.”). The “district court must uphold the jury’s award if there was any ‘legally sufficient 11 basis’ to support it.” Experience Hendrix L.L.C. v. Hendrixlicensing.com Ltd, 762 F.3d 829, 842 12 (9th Cir. 2014) (citation omitted); see also Dunlap v. Liberty Nat. Prod., Inc., 878 F.3d 794, 797 13 (9th Cir. 2017) (“A jury’s verdict must be upheld if it is supported by substantial evidence that is 14 adequate to support the jury’s findings, even if contrary findings are also possible.”) (citation 15 omitted). In making this determination, the Court is to “consider[] all of the evidence in the 16 record, drawing all reasonable inferences in favor of the nonmoving party,” and it “may not make 17 any credibility determinations or reweigh the evidence.” Experience Hendrix, 762 F.3d at 842. 18 Put more plainly, the Court must “draw all inferences in favor of the verdict.” Id. at 845. 19 Rule 59 permits the Court to grant a new trial on all or some of the issues, and to any party, 20 “for any reason for which a new trial has heretofore been granted in an action at law in federal 21 court.” Fed. R. Civ. P. 59(a)(1)(A). Although for a Rule 59 motion, the Court is “not required to 22 view the trial evidence in the light most favorable to the verdict,” Experience Hendrix, 762 F.3d at 23 842, the Court “may not grant a new trial simply because it would have arrived at a different 24 verdict.” Silver Sage Partners, Ltd. v. City of Desert Hot Springs, 251 F.3d 814, 819 (9th Cir. 25 2001). Our circuit has stated that “[a] trial court may grant a new trial only if the verdict is against 26 the clear weight of the evidence.” Pavao v. Pagay, 307 F.3d 915, 918 (9th Cir. 2002) (citing 27 Silver Sage, 251 F.3d at 818-19). 28 5 Google presented its JMOL and new trial arguments in a single, interwoven fashion. See 1 2 Dkt. No. 925. The Court will follow suit, while being mindful of the different standards that 3 govern each rule. DISCUSSION 4 5 United States District Court Northern District of California 6 I. ANDROID-ONLY RELEVANT MARKETS For the monopolization claim, the jury found that Epic had proved the existence of two 7 relevant product markets: (1) an “Android app distribution market,” and (2) a market for 8 “Android in-app billing services for digital goods and services transactions.” Dkt. No. 866 at 3 9 (Question No. 2). The jury found the same two relevant product markets for Epic’s unlawful 10 restraint of trade claim. See id. at 6 (Question No. 8). Google proposes two reasons why, in its 11 view, Epic should not have been permitted to argue for these relevant markets that were “limited 12 to Android devices.” Dkt. No. 925 at 1-7. 13 A. 14 To start, Google says that it is entitled to judgment as a matter of law on all claims Issue Preclusion 15 submitted to the jury because of the preclusive effect of Epic Games, Inc. v. Apple Inc. (“Apple 16 I”), 559 F. Supp. 3d 898 (N.D. Cal. 2021), and Epic Games Inc. v. Apple Inc. (“Apple II”), 67 17 F.4th 946 (9th Cir. 2023), cert. denied, 144 S. Ct. 682 (2024). Apple II affirmed in part and 18 reversed in part Apple I. Google says that the Apple I court found a “market for mobile game 19 transactions in which both Google and Apple competed,” which was a market definition the circuit 20 affirmed. Dkt. No. 925 at 2-3. Consequently, in Google’s view, Epic had “already litigated and 21 lost” the issue of “competition between Apple’s App Store and Google Play.” Id. Because Epic 22 did not propose at trial a market that included Apple, Google contends that Epic failed to prove a 23 valid relevant market at all, which necessarily doomed all of its antitrust claims. Id. at 4. 24 This is not the first time Google has tried to make this point. It is in effect a re-do of the 25 same argument that the Court rejected in prior proceedings because Google had failed to establish 26 the elements of preclusion. Dkt. No. 700 at 2. Nothing has happened since to change the Court’s 27 conclusion. 28 6 1 2 litigation, applies where four conditions are met: (1) the issue at stake was identical in both 3 proceedings; (2) the issue was actually litigated and decided in the prior proceedings; (3) there was 4 a full and fair opportunity to litigate the issue; and (4) the issue was necessary to decide the 5 merits.” Snoqualmie Indian Tribe v. Washington, 8 F.4th 853, 864 (9th Cir. 2021) (cleaned up). 6 United States District Court Northern District of California “Issue preclusion, which bars the relitigation of issues actually adjudicated in previous The market definition issues that were litigated in Apple I and Apple II were plainly not the 7 same as the issues litigated here. In the case against Apple, Epic “proposed two single-brand 8 markets: the aftermarkets for iOS app distribution and iOS in-app payment solutions, derived from 9 a foremarket for smartphone operating systems.” Apple II, 67 F.4th at 970 (emphasis omitted). 10 The district court rejected Epic’s proposed single-brand markets mainly because there was a 11 “failure of proof.” Id. Epic “presented no evidence regarding whether consumers unknowingly 12 lock themselves into Apple’s app-distribution and IAP restrictions when they buy iOS devices.” 13 Id. On appeal, the circuit court determined that the district court “did not clearly err in rejecting 14 Epic’s proposed relevant markets”; “[i]n particular, Epic failed to produce any evidence showing 15 -- as our precedent requires -- that consumers are generally unaware of Apple’s app-distribution 16 and IAP restrictions when they purchase iOS devices.” Id. at 973. 17 Epic took a very different approach to the markets in this case. It did not, for obvious 18 reasons in a case that did not include Apple, advocate for “aftermarkets for iOS app distribution 19 and iOS in-app payment solutions, derived from a foremarket for” iOS devices. Id. at 970 20 (emphasis omitted). Nor did it argue for aftermarkets for Android app distribution and Android 21 in-app payment solutions, derived from a foremarket for Android devices. It took a wholly 22 different approach for the antitrust claims against Google, and offered wholly different evidence 23 about relevant markets than that offered in the case against Apple. The holdings in Apple I and 24 Apple II about Epic’s proposed foremarket/aftermarkets for Apple products, and Epic’s deficient 25 evidentiary support for those markets, have no preclusive effect here. 26 Consequently, Epic was perfectly free at trial to argue for Android-only relevant markets, 27 just as Google was free to argue for a different result. Each side took maximum advantage of this 28 freedom to hotly dispute the definition of the product markets for Epic’s antitrust claims. The jury 7 United States District Court Northern District of California 1 was presented with evidence sponsored by each side, including witness testimony, documents, and 2 expert witness opinions, on the question of the relevant product markets. Google took every 3 opportunity to tell the jury that Google and Apple compete, and so should be considered to be in 4 the same relevant market. If there was one theme Google pressed relentlessly to the jury, it was 5 this one. Epic presented substantial evidence showing that the Android-only product markets 6 made factual and economic sense for this case. For example, Epic’s economics expert, Professor 7 Douglas Bernheim, testified that the fact that Apple and Android compete in the market for 8 smartphones does not mean that they are in the same market for app distribution. Trial Tr. at 9 2423:23-2424:1. The jury also heard from Dr. Bernheim that, based on a SSNIP test and other 10 widely accepted analytical tools, his conclusion was that the Apple App Store does not compete in 11 the same relevant market as the Google Play Store. Id. at 2424:17-2427:16, 2462:2-16. 12 In the end, the jury did precisely what it was called upon to do by resolving the hotly 13 disputed evidence to define the product markets as stated in the verdict. The possibility that the 14 jury might have come out differently is no basis for judgment as a matter of law in Google’s favor. 15 See White, 312 F.3d at 1010. Google also has not demonstrated that the product market verdicts 16 were clearly contrary to the weight of the evidence. 17 B. Aftermarket Theory 18 Despite the plain record of what happened at trial, Google says that Epic was actually 19 proposing a “‘single-brand aftermarket’ theory of market definition” that it failed to prove. Dkt. 20 No. 925 at 5. This is a rather odd argument because Epic never presented or even mentioned a 21 “single-brand aftermarket” in this case, and Google’s suggestion to the contrary is utterly bereft of 22 any evidence. 23 To start, there was no “single brand” in play here. As the parties stipulated in the final jury 24 instructions, Android is a mobile operating system; it is not a brand. See Dkt. No. 850 at ECF 25 p. 16 (Instruction No. 12 (Stipulations of Fact)) ¶ 15. The undisputed evidence showed at trial that 26 Android devices are manufactured by many companies, including Google, Samsung, Motorola, 27 OnePlus, Xiaomi, and other OEMs. This is in sharp contrast to iOS devices, which are 28 manufactured by Apple alone. See Apple II, 67 F.4th at 966. Epic expressly argued for a single8 1 brand aftermarket in the Apple case for iOS devices, and the circuit stated that, “in some instances 2 one brand of a product can constitute a separate market.” Id. at 976 (cleaned up). That 3 observation, and the discussion that followed it, are not relevant here because Epic never proposed 4 or argued for a market consisting of only “one brand of a product” with respect to Google. Id. Google says that it doesn’t matter that there are multiple brands of Android devices United States District Court Northern District of California 5 6 because “Google generates significant revenue from Android devices.” Dkt. No. 945 at 2. Even 7 taking that as true for discussion purposes, it hardly explains why many different and competing 8 OEM brands should be treated as a single brand. Google certainly did not present any evidence, 9 or case law or other authority, in support of that proposition. There simply is no evidentiary or 10 legal reason to treat Epic as though it had pursued a foremarket/aftermarket theory that it did not 11 propose, or to penalize it for not proving that theory at trial. 12 This case also differs from the Apple case in that the Apple App Store is the only app store 13 for iOS devices, which is not true for Android devices. Substantial evidence was presented at trial 14 that multiple Android app stores can be, and on occasion have been, available to consumers. 15 Google’s efforts to suppress rival app stores was another key theme at trial. To that end, an 16 internal Google document asked the “existential question”: “How do we continue to keep Play as 17 the preeminent distribution platform for Android?” Trial Tr. at 920:24-921:14. Other documents 18 referred to a “market” consisting of Android app developers only, see id. at 922:13-923:18, and 19 spoke of “store rivals” that were Android app stores. Id. at 952:3-13. Google’s CEO, Sundar 20 Pichai, testified that each Android OEM “had the potential to have an app store” which “would 21 compete with Google Play.” Id. at 1343:1-5. Overall, the evidence at trial demonstrated that the markets for Android app distribution 22 23 and in-app payment systems are different from the markets for Apple/iOS app distribution and in- 24 app payment systems that were at issue in Apple II. Epic did not pursue a “single-brand 25 aftermarket” here. 26 II. 27 28 JURY INSTRUCTIONS RE RULE OF REASON Google requests a new trial because of alleged legal errors in the jury instructions relating to the Rule of Reason. Dkt. No. 925 at 7-11. Its arguments are not well taken. 9 1 A. 2 On Step 1 of the Rule of Reason, Google says the jury was impermissibly allowed to United States District Court Northern District of California 3 Step One “conclude that individually lawful acts are unlawful in the aggregate.” Dkt. No. 925 at 10. 4 The record demonstrates otherwise. Before trial, the Court granted summary judgment for 5 Google on “‘plaintiffs’ claims that Google unlawfully prohibits the distribution of other app stores 6 on Google Play.’” Dkt. No. 700 at 1 (quoting Dkt. No. 483 at 6). Because governing case law 7 makes clear that Google had no duty to deal, the Court stated that plaintiffs could reference § 4.5 8 of the Developer Distribution Agreement by way of background and context only, but they could 9 not “argue or suggest that § 4.5 is unlawful either on its own or in combination with other alleged 10 practices.” Id. (emphasis added) (citing Verizon Communications, Inc. v. Trinko, 540 U.S. 398 11 (2004)). 12 The same guidance was stated in the jury instructions. The jury was instructed that “[i]t is 13 not unlawful for Google to prohibit the distribution of other app stores through the Google Play 14 Store, and you should not infer or conclude that doing so is unlawful in any way.” Dkt. No. 850 at 15 ECF p. 33 (Instruction No. 24). For the anticompetitive conduct required for Epic’s Section 2 16 claim, the jury was instructed that, “[i]n determining whether Google’s conduct was 17 anticompetitive or whether it was legitimate business conduct, you should determine whether the 18 conduct is consistent with competition on the merits, whether the conduct provides benefits to 19 consumers, and whether the conduct would make business sense apart from any effect it has on 20 excluding competition or harming competitors.” Id. at ECF p. 30 (Instruction No. 23) (emphases 21 added). Nothing in the instructions invited the jury to consider Google’s alleged conduct in the 22 aggregate, or gave them permission to consider whether independently legitimate conduct may 23 have combined to create an anticompetitive effect. 24 The verdict form underscored this by directing the jury to consider each type of conduct 25 separately. For Epic’s Section 1 claim, the jury was called upon to answer separately for three 26 types of agreements -- (1) DDA agreements; (2) agreements with Google’s alleged competitors or 27 potential competitors under Project Hug or Games Velocity Program; and (3) agreements with 28 10 1 OEMs that sell mobile devices (including MADA and RSA agreements) -- whether each type of 2 agreement was an “unreasonable restraint(s) of trade.” Dkt. No. 866 at 5 (Question No. 7). As the record established, and contrary to Google’s argument, the jury was in fact “guided 3 United States District Court Northern District of California 4 . . . to consider each category of conduct individually.” Dkt. No. 925 at 10. 5 B. Step Two 6 Google says that on Step 2 of the Rule of Reason analysis, “[t]he jury should have been 7 instructed to consider cross-market justifications,” i.e., procompetitive benefits not limited to the 8 relevant product markets at issue. Dkt. No. 925 at 7-8. But in Apple II, our circuit expressly took 9 up the issue of the “cognizability of cross-market rationales.” 67 F.4th at 989. There, Epic had 10 argued that, “even if Apple’s security and privacy restrictions are procompetitive, they increase 11 competition in a different market than the district court defined and in which Epic showed step- 12 one anticompetitive effects, and thus are not legally cognizable at step two.” Id. (emphasis in 13 original). Our circuit noted that the “Supreme Court’s precedent on this issue is not clear,” even 14 15 though on occasion it “has considered cross-market rationales in Rule of Reason and 16 monopolization cases.” Id. The circuit “decline[d] to decide this issue here.” Id. The circuit also 17 determined that Apple’s procompetitive justifications related to the relevant market. Id. at 990. Consequently, there was no legal mandate to expressly require the jury to consider cross- 18 19 market justifications, such as “Google’s competition with Apple in smartphones,” Dkt. No. 925 at 20 9, as Google urges. Contrary to Google’s argument, this is not at all a situation where the circuit 21 has not yet “stated ‘explicitly’ a legal point that ‘was implicit in [its] past decisions.’” Dkt. 22 No. 945 at 4 (citing Dang v. Cross, 422 F.3d 800, 807-08 (9th Cir. 2005)). It also bears repeating 23 that Google spared no opportunity at trial to tell the jury of its views about competition with 24 Apple. 25 C. 26 Google makes another argument contrary to circuit law for Step 3 of the Rule of Reason Step Three 27 analysis by stating that the Court improperly invited the jury to balance competitive effects. Dkt. 28 No. 925 at 11. But as Google acknowledges, Apple II expressly “held that Ninth Circuit precedent 11 1 requires balancing.” Id.; see Apple II, 67 F.4th at 994 (“where a plaintiff’s case comes up short at 2 step three, the district court must proceed to step four and balance the restriction’s anticompetitive 3 harms against its procompetitive benefits.”). There was no error in the Court’s balancing 4 instruction to the jury. 5 III. United States District Court Northern District of California 6 SUFFICIENCY OF EVIDENCE SUPPORTING JURY VERDICT Google’s primary claim for judgment as a matter of law or a new trial is that the verdict is 7 “unsupported by legally sufficient evidence.” Dkt. No. 925 at 11-27. True to form, Google 8 objects to just about everything adduced at trial that impugned Google’s conduct. The Court has 9 undertaken the laborious task of reviewing the record in light of Google’s many complaints, as the 10 ensuing discussion details. Some prefatory observations are in order. The true crux of Google’s 11 argument isn’t that the verdict was not based on substantial evidence, but rather that the jury didn’t 12 see the evidence in the way Google wanted. This is not a situation where the verdict was based on 13 speculation or where the evidence would allow only one conclusion that is contrary to what the 14 jury decided. 15 Another problem is that Google ignores in practice the standards for granting JMOL or a 16 new trial. For Rule 50, as discussed, all reasonable inferences are made in favor of the verdict and 17 Epic, as the nonmoving party. See Experience Hendrix, 762 F.3d at 842, 845. This is so 18 irrespective of whether the evidence might have supported a different result. See Pavao, 307 F.3d 19 at 918 (“A jury’s verdict must be upheld if it is supported by substantial evidence, which is 20 evidence adequate to support the jury’s conclusion, even if it is also possible to draw a contrary 21 conclusion.”). Google subverts this by asking in effect that all inferences be drawn for its benefit. 22 For Rule 59, where the review is free of inferences for either side, the jury verdict will stand 23 unless it is “against the clear weight of the evidence.” Silver Sage, 251 F.3d at 819. Google 24 slights this by insisting that a new trial is warranted simply because some evidence was disputed 25 and the jury might have decided differently. 26 27 28 A. Relevant Market 1. Limitation to Android In-App Payments The jury found a market consisting of “Android in-app billing services for digital goods 12 United States District Court Northern District of California 1 and services transactions.” Dkt. No. 866 at 3, 6 (Question Nos. 2, 8). Google says that “[t]he 2 evidence does not support the jury’s decision to exclude out-of-app payment systems from the 3 relevant product market,” highlighting websites in particular as a “reasonable substitute” that 4 should have been included in the relevant market. Dkt. No. 925 at 11-12. 5 Not so. Epic’s economics expert, Dr. Steven Tadelis, testified that the relevant product 6 market was properly limited to “any product that could do what Google Play Billing does,” i.e., 7 “any payment solution provider for digital content on Androids.” Trial Tr. at 2553:1-4. He 8 further testified that “web purchases are not a viable substitute for in-app purchases” because of 9 “friction” -- whereas in-app purchases can be completed in two to three steps by the user, web 10 store purchases required at least eight steps. Id. at 2554:4-2556:15. This increased friction was 11 likely to lead to increased dropoff along the process, where users do not complete the purchase. 12 Id. at 2556:19-2557:10. Google executive Purnima Kochikar was taken through the 18 steps a 13 user would have had to go through “to have the Amazon App Store show up on the Home Page of 14 their Android device.” Id. at 746:3-752:8. Kochikar called the sideloading experience “abysmal,” 15 id. at 753:22-755:5, and agreed that “the number of steps makes for a bad user experience,” and 16 “where there’s friction, people [often] fall out and don’t complete purchases.” Id. at 762:20-763:2. 17 Witness Eric Chu testified that YouTube engineers worked to “reduce the number of clicks,” 18 asking themselves, “[f]rom the moment the user wants to buy something[,] what can we do to 19 reduce [the] number of clicks and make it easier for them to purchase something[?]” Id. at 20 1441:2-1442:11; Dkt. No. 915-1 at ECF p. 85. The “[r]eason for that is obvious that the more 21 friction there is[,] the more likely we lose users along the buy flow.” Id. 22 The jury was instructed, without objection by Google, that “[i]n determining the product 23 market, the basic idea is that the products within it are interchangeable as a practical matter from 24 the buyer’s point of view.” Dkt. No. 850 at ECF p. 22 (Instruction No. 18). This means “they 25 must be, as a matter of practical fact and the actual behavior of consumers (meaning users and 26 developers), substantially or reasonably interchangeable to fill the same consumer needs or 27 purposes. Two products are within a single market if one item could suit buyers’ needs 28 substantially as well as the other.” Id. The jury reasonably relied on the testimony summarized 13 United States District Court Northern District of California 1 above, and similar evidence at trial, to conclude that from the developers’ and users’ points of 2 view, out-of-app payment systems were not reasonable substitutes for in-app payment systems. 3 Google’s citation to Tanaka v. University of Southern California, 252 F.3d 1059, 1063 (9th 4 Cir. 2001), is misdirected. Google cites it for the proposition that consumers’ personal preferences 5 are not relevant. But Tanaka involved a highly unique personal preference -- literally the 6 preference of just plaintiff Tanaka, a “star high school soccer player” who wanted to “remain in 7 the Los Angeles area” so she could “be close to her family.” 252 F.3d at 1061, 1063. Tanaka 8 challenged an intercollegiate athletic association rule that discouraged intra-conference transfers, 9 and although the association was national in scope, she alleged that the “relevant geographic 10 market is Los Angeles and the relevant product market is the ‘UCLA women’s soccer program,’” 11 based purely on her personal desires. Id. at 1063. The circuit concluded that Tanaka’s personal 12 preference to be near her family could not be a proper basis for defining the “‘area of effective 13 competition’ for student-athletes competing for positions in women’s intercollegiate soccer 14 programs.” Id. 15 The facts here could not be more different. This is not a case of one person trying to define 16 a purely personal market. Substantial evidence was presented at trial to the effect that an out-of- 17 app payment solution would not meet a developer’s or user’s needs as well as an in-app payment 18 solution. The evidence showed that this was not a matter of mere preference or taste, but a 19 product of design and function. Out-of-app payment solutions are a cumbersome mechanism for 20 sales, and so are not likely to be viewed by developers or users as reasonable substitutes for in-app 21 payment systems. The jury’s finding of an Android in-app payment solutions product market was 22 not against the great weight of the evidence. 23 24 2. Geographic Scope Substantial evidence supported the jury’s finding that the relevant geographic market was 25 “worldwide excluding China.” Dkt. No. 866 at 3, 6 (Question Nos. 2, 8). The jury was instructed 26 that the “relevant geographic market is the area in which Google faces competition from other 27 firms that compete in the relevant product markets and to which customers can reasonably turn for 28 purchases.” Dkt. No. 850 at ECF p. 23 (Instruction No. 19). The jury was further instructed that 14 1 “[w]hen analyzing the relevant geographic market, you should consider whether changes in prices 2 or product quality in one geographic area have substantial effects on price or sales in another 3 geographic area, which would tend to show that both areas are in the same relevant geographic 4 market.” Id. Contrary to Google’s argument, there is no absolute “legal test” of “‘consumer 5 substitution.’” Dkt. No. 945 at 6. United States District Court Northern District of California 6 Epic’s economics expert, Dr. Douglas Bernheim, testified that the appropriate geographic 7 boundary for the relevant market was “global, excluding China.” Trial Tr. at 2445:16-24. This 8 was because “competitive conditions” in different countries “are largely similar,” and Google’s 9 challenged conduct in this case was “global, excluding China.” Id. at 2446:1-11. It was 10 appropriate to carve out China because China was “very dissimilar” in that Google Android, 11 Google Play, and Google’s challenged conduct, were “not in China.” Id. at 2446:18-23. 12 Dr. Tadelis agreed with Dr. Bernheim’s analysis, and also concluded that the geographic market 13 was “global excluding China.” Id. at 2560:10-16. 14 Other witnesses at trial also testified to these facts. For example, Google witness James 15 Kolotouros testified that Google Play is not permitted in China and is not preinstalled on 16 smartphones distributed there. See id. at 1070:7-17. On the flip side, “every Android smartphone 17 outside of China comes preinstalled with Google Play.” Id. at 1070:18-21. Kolotouros also 18 testified that Google faced competition from companies such as Samsung, Xiaomi, Oplus, and 19 Vivo, which “had the potential to have app stores that competed with Google Play in markets 20 outside of China.” Id. at 1079:16-1080:24; see also id. at 1207:13-1210:11 (Google witness Jamie 21 Rosenberg’s testimony re internal Google document discussing Amazon App Store’s growing 22 popularity in Japan, and Google’s concern that Amazon might “scal[e] up and go[] global,” 23 becoming a more global threat to Google Play). Similarly, for in-app payment solutions, there 24 was trial testimony that Google Play Billing is “not offered in China,” and Google faces 25 competition from companies such as PayPal, Square, and Braintree outside of the United States. 26 Id. at 2586:21-2588:25. The jury’s geographic market findings were supported by adequate 27 evidence and were not against the great weight of the evidence. 28 15 United States District Court Northern District of California Anticompetitive Effect of Google’s Conduct 1 B. 2 For Epic’s monopolization claim, the jury was instructed that “[a]nticompetitive acts are 3 acts, other than competition on the merits, that have the effect of preventing or excluding 4 competition or frustrating the efforts of other companies to compete for customers within the 5 relevant market.” Dkt. No. 850 at ECF p. 30 (Instruction No. 23). For Epic’s restraint of trade 6 claim, the jury was instructed that “[a] harmful effect on competition, or competitive harm, refers 7 to a reduction in competition that results in the loss of some of the benefits of competition, such as 8 lower prices, increased output, or higher product quality.” Id. at ECF p. 37 (Instruction No. 28). 9 The jury was further instructed to consider the following factors: “(1) the effect of the challenged 10 restraint on prices, output, product quality, and service; (2) the purpose and nature of the 11 challenged restraint; (3) the nature and structure of the relevant market; (4) the number of 12 competitors in the relevant market and the level of competition among them, both before and after 13 the challenged restraint was imposed; and (5) whether Google possesses market power.” Id. 14 After considering the evidence in light of these instructions, the jury found that Google 15 “willfully acquired or maintained monopoly power by engaging in anticompetitive conduct.” Dkt. 16 No. 866 at 3 (Question No. 3). The jury also found each of these agreements to have been 17 unreasonable restraints of trade (and so impliedly to have had anticompetitive effects): (1) “DDA 18 agreements”; (2) “Agreements with Google’s alleged competitors or potential competitors under 19 Project Hug or Games Velocity Program”; and (3) “Agreements with OEMs that sell mobile 20 devices (including MADA and RSA agreements).” Id. at 5 (Question No. 7). 21 Each of these findings was supported by substantial evidence, and was not against the great 22 weight of the evidence. There was substantial evidence at trial that Google had engaged in 23 conduct, “other than competition on the merits, that ha[d] the effect of preventing or excluding 24 competition or frustrating the efforts of other companies to compete for customers within the 25 relevant market.” Dkt. No. 850 at ECF p. 30 (Instruction No. 23). It bears mention that Google 26 does not contest that Epic presented sufficient evidence on Google’s market power or the barriers 27 to entry that existed in both of the relevant markets found by the jury. See Dkt. No. 932 at 15, 28 n.11; compare with Dkt. No. 945. 16 The jury heard a great deal of testimony about Google’s agreements with existing or United States District Court Northern District of California 1 2 potential competitors in connection with Project Hug.3 Activision Blizzard King (ABK) was a 3 developer that signed a Project Hug deal. The developer of mobile games such as Candy Crush 4 and Call of Duty, ABK “had the highest estimated spend by users of all of the Project Hug 5 developers.” Trial Tr. at 445:5-13. ABK had been “quite vocal [in] complaining about Google 6 Play’s 30 percent fee,” and Google witness Koh testified that it had been communicated to Google 7 that ABK was considering the option of starting its own Android app store. Id. at 445:14-22, 8 463:5-8. Google estimated that it faced a risk of losing $243 million per year if ABK were to pull 9 its content from the Google Play Store. Id. at 463:24-464:9. Google internally discussed this risk, 10 as well as the possible “contagion risk” if ABK were to launch its own store and “attract[] more 11 content from other developers.” Id. at 463:5-464:24.4 Google then offered ABK a Project Hug 12 deal for $360 million. Id. at 465:3-466:8. Google witness Kochikar testified about Google’s 13 concern that ABK might launch its own Android app store, and Google’s hope that a Project Hug 14 deal would prevent that. See id. at 804:7-807:12. Riot Games was another developer that was 15 offered and took a Project Hug deal. An internal Google document stated, “A year ago, we pulled 16 all the stops, promised them $10 million co-marketing for before they signed GVP, for example, 17 to get Riot Games to stop their in-house app store effort.” Id. at 811:6-812:12. In exchange for significant payments from Google, developers who signed a Project Hug 18 19 agreement “could not launch [an app] either first or exclusively on any competing Android 20 distribution platform.” Trial Tr. at 442:23-443:2. Developers who agreed to Project Hug also 21 could not “launch a materially different version of the game that it had on Google Play on a 22 competing Android app distribution platform.” Id. at 444:10-15. Epic’s expert, Dr. Bernheim, 23 testified to the anticompetitive effects of these provisions. In his view, these provisions 24 25 26 27 28 3 The Games Velocity Program was a continuation of Project Hug. See Trial Tr. at 491:13-14; id. at 410:13-15 (Google witness Lawrence Koh affirming that “Project Hug was later renamed the Games Velocity Program”). Witness Koh testified that the “contagion risk” had to do with Google’s concern that “once top developers took their gaming content off of Google Play, that other developers would potentially follow suit.” Trial Tr. at 422:14-16. 17 4 1 “prevent[ed] any significant differentiation,” disincentivized developers from creating valuable 2 content, and would also have discouraged Project Hug developers from entering the app store 3 market themselves. Id. at 2403:7-2410:7. United States District Court Northern District of California 4 There was also substantial evidence of the anticompetitive effects of Google’s agreements 5 with OEMs, specifically the Mobile Application Distribution Agreement (MADA) and Revenue 6 Share Agreements (RSA). The MADA is an agreement that Google enters into with Android 7 OEMs. Pursuant to the MADA, OEMs must place Google Play on the default home screen of 8 their Android devices. Trial Tr. at 1351:14-21. Virtually all OEMs that manufacture Android 9 smartphones have entered into a MADA, and so Google Play is preinstalled on the default home 10 screen of nearly all Android smartphones. Id. at 1351:22-1352:8. Google’s CEO, Sundar Pichai, 11 acknowledged that “[t]ypically,” placement on the default home screen tends to lead to more 12 usage of an app. Id. at 1352:9-12. Preinstallation of Google Play on the default home screen is a 13 precondition for an OEM to have access to other key Google GMS apps and Android APIs 14 without which many Android applications cannot function. Id. at 1355:1-1356:4. Restrictions 15 like these made it difficult for competitors like Amazon to obtain “premium placement” for apps 16 such as its own app store, Dkt. No. 915-1 at ECF p. 107, and so it was difficult for alternative app 17 stores to get off the ground. 18 The terms of the Google Revenue Share Agreements with OEMs were even more 19 aggressive. The RSA 3.0 agreements are the third iteration of that contract, and they offer OEMs 20 the opportunity to enroll their devices in three different tiers. Trial Tr. at 1053:1-9. For the 21 “premier tier,” which offers the highest revenue share, an OEM “may not install any app store on 22 their device other than Google Play.” Id. at 1053:7-24. Epic’s expert, Dr. Bernheim, testified that 23 this kind of profit-sharing with a competitor disincentivizes competition, and so is anticompetitive. 24 Id. at 3189:22-3190:14. 25 There was additional trial testimony to the same effect. Google witness Kolotouros 26 testified that, with the exception of Samsung, most of Google’s major Android OEM partners 27 executed RSA 3.0 agreements. Id. at 1092:4-6. OnePlus was one such OEM, and outside of 28 China and India, OnePlus enrolled the vast majority of its devices in the premier tier. Id. at 18 United States District Court Northern District of California 1 1094:3-17. Although OnePlus wanted to enter into a partnership with Epic Games whereby the 2 Epic Games Store app would be preloaded onto OnePlus devices, Google declined to grant a 3 waiver to the premier tier restrictions. After that, OnePlus decided to “take Google’s revenue 4 share payments and keep nearly all of its devices outside of India in the premier tier instead of 5 preinstalling the Epic Game Store app.” Id. at 1094:18-25, 1098:24-1099:13. 6 The jury heard more evidence from which it could reasonably conclude that the RSA 3.0 7 agreements were anticompetitive. This included an internal Google document in which Google 8 employees discussed how “Google cannot stop OEMs from preloading the Amazon App Store due 9 to anticompetitive concerns on the MADA 2.0 only,” but “[w]e can do this through revenue share 10 deals.” Trial Tr. at 1074:8-17. The employees agreed that having “stricter placement restrictions 11 through revenue share” was something that would “help stem the tide of emerging app stores.” Id. 12 at 1074:22-1075:19. In the course of this discussion, another Google executive inquired about 13 why Google “doesn’t put everything in the MADA” and asked, “Is it anticompetitive concerns or 14 something more than that?” Witness Kolotouros responded, “This might be better discussed in 15 person as opposed to writing.” Id. at 1075:20-1076:12. This and much other evidence supported a 16 verdict against Google on the “purpose and nature of the challenged restraint,” namely the RSA 17 3.0 agreements. 18 There was additional evidence at trial that Google worked to suppress competition by 19 actively impeding users from “sideloading” competing app stores through increased “friction” and 20 “scare screens.” “Sideloading” referred to a direct installation process whereby a user “find[s] an 21 app via a mechanism that is not billing itself purely as an app store.” Trial Tr. at 2128:4-6. 22 “Friction” meant “the screens, the dialogues, the warnings that an operating system is going to put 23 up and show to users and sort of force the user to click through or interact with before the user can 24 actually accomplish the intended task.” Id. at 2113:21-25. Google’s CEO, Sundar Pichai, 25 acknowledged that “the more friction there is, the less likely the user completes that flow,” id. at 26 1361:11-13, and there was evidence that Google viewed friction as a means of impeding users 27 from sideloading third-party app stores. For example, a Google internal document titled, 28 “Amazon competitor deep dive,” noted that “Amazon [was] emerging as a major challenge to Play 19 1 in gaming globally.” Dkt. No. 886-50 (Trial Ex. 682) at ECF pp. 1-2. Another slide was titled, 2 “Amazon strongly promoting its 15%+ discount on IAPs available via Play, but for now switching 3 hurdle too high for most users.” Id. at ECF p. 11. Under the heading, “Significant hurdle to 4 switching to Amazon apk,” the Google document stated, “Process is quite complex, involves 14 5 steps (but motivated users will follow walkthroughs like this on YT).” Id. United States District Court Northern District of California 6 For the DDA agreements, Google says that “Epic failed to prove, and no reasonable jury 7 could have found, that the anti-steering restrictions in the DDA were anticompetitive because they 8 merely prevent developers who choose to use valuable Google services and intellectual property in 9 the Play store from depriving Google of compensation for that value.” Dkt. No. 925 at 22. But 10 this objection ignores the trial evidence about the anticompetitive nature of these anti-steering 11 restrictions and the DDA in general. For example, there was testimony that Paddle, a company 12 that offers to developers in-app payment solutions, was prevented from more effectively entering 13 the in-app payment services market because “Google Play’s terms of services for developers [i.e., 14 the DDA] expressly prohibit the usage of a third-party payment method.” Trial Tr. at 653:3-11. 15 This and similar trial evidence demonstrate that the jury’s findings on Google’s 16 anticompetitive conduct were well supported. There was sufficient evidence for the jury to agree 17 with Dr. Bernheim that Google “impairs competition without preventing it entirely,” Trial Tr. at 18 3181:1-3185:12, thereby satisfying the requirement that Google’s conduct “frustrat[ed] the efforts 19 of other companies to compete for customers within the relevant market.” Dkt. No. 850 at ECF 20 p. 30 (Instruction No. 23). Because the evidence discussed above is adequate to support the jury’s 21 verdict, the Court declines to address Google’s other arguments on the anticompetitive effect 22 element of Epic’s antitrust claims. 23 C. Tying 24 Substantial evidence supported the jury’s finding in favor of Epic on its tying claim, 25 namely that “Google unlawfully tied the use of the Google Play Store to the use of Google Play 26 Billing.” Dkt. No. 866 at 7 (Question No. 10). 27 The jury heard evidence that the Google Play Store and Google Play Billing are separate 28 products. It was not necessary for Epic to prove that there was separate demand for Google Play 20 United States District Court Northern District of California 1 Billing as a standalone product; rather, it was enough for Epic to prove that there was demand for 2 in-app billing services separate from the demand for app distribution services. See Jefferson 3 Parish Hospital District No. 2 v. Hyde, 466 U.S. 2, 21-22 (1984) (a “tying arrangement cannot 4 exist unless two separate product markets have been linked”; “in this case no tying arrangement 5 can exist unless there is a sufficient demand for the purchase of anesthesiological services separate 6 from hospital services”). Epic presented substantial evidence on this element. Epic witness 7 Steven Allison testified, for example, that in the case of the Epic Game Store, developers can use 8 Epic Direct Pay, which is Epic’s in-app payment solution, or “they can bring their own if they’ve 9 set their game up to do so.” Trial Tr. at 227:5-12. Down Dog’s CEO, Benjamin Simon, testified 10 that he would prefer Stripe or PayPal over Google Play Billing, and if he “had the ability at Down 11 Dog to use PayPal or Stripe on [Down Dog’s] Android app,” he would do so. Id. at 303:2-10. 12 There was substantial evidence that Google coerced its customer -- here, developers -- to 13 buy the tied product (Google Play Billing) in order to obtain the tying product (Google Play 14 Store). Numerous witnesses testified that developers whose apps are on the Google Play Store are 15 required through the DDA to use only Google Play Billing to sell any digital content that is to be 16 used inside of the app. See, e.g., Trial Tr. at 887:7-13, 889:9-21, 1185:18-21. 17 The jury had an ample evidentiary basis for rejecting Google’s business justification 18 defense as pretextual, and finding that Epic had successfully proven the existence of less 19 restrictive alternatives. The trial testimony established, for example, that developers who sold 20 digital content for use outside of the app were exempted from the requirement to use Google Play 21 Billing, which weakened Google’s business justification argument. See Trial Tr. at 1185:22-25. 22 Similarly, developers who sold physical goods were also exempted. See Dkt. No. 886-94 (Trial 23 Ex. 1436); see also Trial Tr. at 936:3-9. 24 Epic also adduced evidence that the Google Play Store was so profitable that Google did 25 not need to tax developers a 30% fee through Google Play Billing to be fully compensated for its 26 IP and other costs for the Google Play Store. For example, the jury saw an internal Google 27 document showing that some developers paid “more than a hundred million dollars per year more 28 than the value that they have obtained from Google”; and for the 100 most negative developers 21 United States District Court Northern District of California 1 (whose payments to Google exceeded the estimated value they received from Google), Google 2 internally estimated that on average, they “receiv[ed] a value equivalent to 19 percent,” but “still 3 were required to pay Google a 30 percent revenue share.” Trial Tr. at 608:6-611:22. Based on 4 Google Play’s revenue numbers, this worked out to $1.43 billion per year that the top 100 most 5 negative developers were overpaying to Google. See id. at 612:3-613:11. There was additional 6 evidence at trial that Google was concerned about public criticism calling out “Google’s 30 7 percent fee on in-app purchases made on apps distributed through Google Play” as “highway 8 robbery.” Id. at 417:17-418:24; see also, e.g., id. at 708:18-21 (internal Google document stating 9 that “the team estimates that if you compare the value of nonSearch-driven discovery versus 10 revenue share paid, Tinder is now deriving only 10 percent of the revenue share value versus the 11 30 percent share they pay.”). Overall, the jury had more than enough evidence at the balancing 12 step to conclude that any benefit from Google’s tie was outweighed by its competitive harms. 13 It was not improper for the jury to consider the DDA as an unreasonable restraint of trade 14 and to additionally consider Epic’s tying claim. The tying claim focused on Google’s coercive tie 15 of Google Play Billing to the Google Play Store, while Epic’s challenge to the DDA also 16 encompassed the DDA’s anti-steering provisions. Multiple legal claims may be based on the same 17 underlying conduct, and Google has not presented any authority to the contrary. 18 D. 19 Google challenges the jury’s implicit finding in favor of Epic on Step 3 of the Rule of Substantially Less Restrictive Alternatives 20 Reason, and says that no reasonable jury could find that Epic satisfied its burden to identify 21 substantially less restrictive alternatives that would be virtually as effective in serving Google’s 22 objectives without significantly increased cost. Dkt. No. 925 at 24-27. 23 Google overlooks the fact that the jury might simply have rejected Google’s proffered 24 justifications. The jury had ample evidence to do so. There was evidence at trial, for example, to 25 support Epic’s theory that the exclusionary provisions in the MADA and RSA agreements were 26 put into those agreements for anti-competitive reasons, rather than any legitimate business reasons. 27 See, e.g., Trial Tr. at 2920:19-2921:7 (Google witness Gennai testifying that RSA 3.0 premier tier 28 was developed “to respond to increasing app store competition from OEMs and large platforms 22 1 like Epic”); id. at 1078:3-1079:19 (changes to RSA proposed “to protect Google from key 2 strategic risks,” including risks of revenue loss due to “Chinese OEMs and Samsung . . . actively 3 investing in creating own app and services ecosystems”); id. at 1073:6-1076:12 (internal Google 4 document stating that “Google cannot stop OEMs from preloading the Amazon App Store due to 5 anticompetitive concerns on the MADA 2.0 only,” but could “prevent OEMs from preloading 6 competitive app stores” through “revenue share deals”). For the sideloading warnings, there was evidence at trial that the increased friction built in United States District Court Northern District of California 7 8 by Google were not related to Google’s assessment of the security risk posed by the material the 9 user was trying to sideload. Google CEO Pichai agreed that “[s]ome websites, such as those from 10 reputable developers, actually present very low security risks,” and yet “Google’s unknown 11 sources flow does not distinguish between those trusted developers and every other website.” 12 Trial Tr. at 1365:2-8; see also id. at 1693:11-1695:21 (sideloading / “unknown source” install 13 flows of 14 or 17 steps applied equally to apps from companies such as Microsoft or Adobe, 14 which are known to Google). Epic’s mobile security expert, Dr. James Mickens, testified that any 15 legitimate benefit of increased security protections for users could have been accomplished 16 through less restrictive means such as fewer screens, or a notarization process that differentiated 17 among the types of security risks presented by different apps or companies. See id. at 2149:2-7, 18 2151:6-8, 2152:4-6, 2157:2-24. His overall opinion, which the jury could reasonably have 19 credited and believed, was that the friction Google imposes is unwarranted and disproportionate, 20 and that Google could reduce the amount of friction while preserving the status quo on security. 21 For the parity provisions in Project Hug and the anti-steering provisions in the DDA, as 22 discussed above in Section III.B. supra, sufficient trial evidence supported a conclusion by the 23 jury that those were motivated by anticompetitive reasons, rather than legitimate business ones. 24 IV. 25 26 EVIDENTIARY RULINGS AND ADVERSE INFERENCE INSTRUCTION Google says that three evidentiary rulings entitle it to a new trial. Dkt. No. 925 at 27-29. The record demonstrates otherwise. Google Employees’ Use of Attorney-Client Privilege 27 A. 28 To start, Google says the Court “permitted Epic to question witnesses about markings 23 1 related to attorney-client privilege on produced documents.” Dkt. No. 925 at 27. Google also 2 claims, quite brazenly and wrongly in light of its willful conduct to hide material evidence, that it 3 had not “improperly withheld any document on the basis of privilege” and so “Epic’s questioning 4 of Google witnesses regarding privilege markings on documents gave the jury the incorrect 5 impression that Google had improperly asserted the attorney-client privilege.” Id. United States District Court Northern District of California 6 Google’s remarks are ill made. After the Court’s findings of fact against Google for 7 willfully failing to preserve Google Chat evidence, see Dkt. No. 469, more evidence emerged at 8 trial of a frankly astonishing abuse of the attorney-client privilege designation to suppress 9 discovery. CEO Pichai testified that there were occasions when he “marked e-mails privileged, 10 not because [he was] seeking legal advice but just to indicate that they were confidential,” as he 11 put it. Trial Tr. at 1321:17-24. He knew this was a misuse of the privilege. Id. at 1323:5-17. 12 Emily Garber, a Google in-house attorney, testified that there was a practice at Google of 13 “loop[ing] in” a lawyer based on a “misapprehension about the rules of privilege,” and that Google 14 employees “believed that including [an in-house lawyer] would make it more likely that the email 15 would be considered privileged.” Id. at 964:21-966:5. Garber called this “fake privilege,” a 16 practice that she appears to have found amusing rather than something a lawyer should have put an 17 immediate and full stop to. Id. at 964:21-23; Dkt. No. 887-86 (Trial Ex. 6487) at EXHIBIT 18 pp. 012-013. 19 On this record, there was no error in the Court’s evidentiary ruling that Epic could “present 20 fake privilege” and make arguments to the jury about it. Id. at 785:5-6. The Court was crystal 21 clear that Epic could not “do anything else with privilege,” and it commended the parties for not 22 saying “anything about” documents that had been “branded privileged” even though it should not 23 have been subject to an assertion of privilege. Id. at 785:8-12. 24 B. 25 Google repackages the prior preclusion argument as an ostensible evidentiary objection to Preclusion of Outcome of Epic v. Apple 26 say: “the Google Play store’s primary competitor is free to use the same basic service fee model 27 explains why it is important for Google to use that same model. The Court erred by preventing 28 Google from introducing evidence that Apple was unlikely to change its existing model in light of 24 1 the outcome of Epic v. Apple -- a market fact that supports Google’s procompetitive justifications 2 for that model and the alleged tie.” Dkt. No. 925 at 28. 3 4 Section I.A., supra. In addition, for the same reason that there was no error in the jury’s decision 5 to exclude Apple from the relevant product markets it found, these outside facts about Apple are 6 not nearly as relevant and important as Google urges. 7 C. 8 Google’s comments on the permissive inference instruction are even more poorly taken 9 United States District Court Northern District of California The point fares no better as an evidentiary objection than it did in the prior version. See Adverse Inference Instruction that those about the attorney-client privilege. The Court determined after an evidentiary hearing 10 held before trial that Google had willfully failed to preserve relevant Google Chat 11 communications, and allowed employees at all levels to hide material evidence. Dkt. No. 469. 12 The evidence presented at trial added more fuel to this fire. As discussed, Google in-house 13 attorney Garber testified about the company practice of asserting a fake privilege to shield 14 documents and communications from discovery. Other witnesses also amplified the seriousness 15 and pervasiveness of Google’s preservation abuses. For example, Google employee Margaret 16 Lam, who worked on RSA issues, said in a Chat message that she didn’t have a specific document 17 because “competition legal might not want us to have a doc like that at all :).” Trial Tr. at 991:16- 18 992:8 (smiley face emoji in original). She was a party to other Chats where, in a discussion about 19 MADA, she asked to turn history off because of “legal sensitivity”; she requested to turn history 20 off in a different conversation about RSAs, so there would be no “trail of us talking about waivers, 21 etc.” Dkt. No. 887-83 (Trial Ex. 6464), Dkt. No. 888-23 (Trial Ex. 8020). Witness Lam also 22 testified that the decision of which Chats to preserve had been left in her hands, but she had “no 23 idea” what was or was not relevant. Trial Tr. at 1012:6-1014:9. 24 Overall, there was an abundance of pretrial and trial evidence demonstrating “an ingrained 25 systemic culture of suppression of relevant evidence within Google.” Id. at 1044:15-17. The 26 Court had advised the parties before trial that an appropriate sanction might include a permissive 27 inference instruction to the jury. Dkt. No. 700 at 3-4. After the additional evidence of 28 malfeasance emerged during trial, the Court raised the question of whether a mandatory adverse 25 1 inference instruction would be more fitting. Trial Tr. at 1044:4-22. Even then, despite the 2 mountain of evidence against Google, the Court held an evidentiary hearing on the question 3 outside of the presence of the jury. The results of this hearing were disappointing. Google’s chief legal officer, Kent Walker, United States District Court Northern District of California 4 5 was the main witness. Despite the seriousness of these issues, and the likelihood that they could 6 affect other litigation matters where Google is a party, Walker showed little awareness of the 7 problems and had not investigated them in any way. Trial Tr. at 1834:18-1835:17. Much of his 8 testimony was in direct opposition to the facts established at the prior Google Chat hearing. See, 9 e.g., id. at 1829:16-1830:3. Overall, Walker did nothing to assuage the Court’s concerns. 10 In these circumstances, the salient question was not whether an adverse inference 11 instruction should be given at all, but whether the inference should be permissive or mandatory. 12 The Court would have been well within its discretion to order a mandatory inference, given the 13 volume of evidence of Google’s misconduct. Even so, the Court took the conservative approach 14 of permitting the jury to make an adverse inference rather than requiring it to. The parties had a 15 fair and balanced opportunity during trial to present evidence and arguments about Google Chat 16 preservation and Google’s conduct, and both sides took full advantage of that. The jury was free 17 to make or decline an inference as it saw fit. To further ensure fairness, the Court instructed Epic 18 that it could not make arguments about Google’s conduct predating August 2020. See Trial Tr. at 19 3237:4-8. If Epic opened that door, Google would have been permitted to respond, see id., but 20 Epic followed that instruction in its closing argument. See id. at 3352:3-3386:14, 3430:19-3435:7. In light of this record, Google’s complaints about the inference instruction are wholly 21 22 misdirected. It has not provided anything close to a good reason to conclude otherwise. 23 V. 24 ADVISORY JURY Google says, rather disingenuously, that the Court has not clarified “whether it was going 25 to treat the jury’s verdict as binding or advisory,” and requests that the Court treat the verdict as 26 advisory. Dkt. No. 925 at 29. It argues further that it “did not consent to a jury trial,” and even if 27 it did, it withdrew that consent. Id. at 29-30. 28 26 1 2 trial, and lost, for good reason. The Court expressly denied Google’s request to “abandon a jury 3 trial” on the eve of trial. Trial Tr. at 6:13-7:16. Google is in effect seeking reconsideration of that 4 ruling, for no good reason. 5 United States District Court Northern District of California Google again ignores that it already made these arguments to the Court prior to the start of To summarize the prior proceedings on this issue, under Federal Rule of Civil Procedure 6 39(c)(2), “[i]n an action not triable of right by a jury, the court, on motion or on its own: . . . 7 (2) may, with the parties’ consent, try any issue by a jury whose verdict has the same effect as if a 8 jury trial had been a matter of right, unless the action is against the United States and a federal 9 statute provides for a nonjury trial.” Google consented to a jury trial of Epic’s antitrust claims 10 against it. A party’s consent under Rule 39(c)(2) can be express or implied. See, e.g., Bereda v. 11 Pickering Creek Indus. Park, Inc., 865 F.2d 49, 52 (3d Cir. 1989); Thompson v. Parkes, 963 F.2d 12 885, 889 (6th Cir. 1992) (en banc); Alcatel USA, Inc. v. DGI Techs., Inc., 166 F.3d 772, 795-96 13 (9th Cir. 1999); Broadnax v. City of New Haven, 415 F.3d 265, 271-72 (2d Cir. 2005); Pals v. 14 Schepel Buick & GMC Truck, Inc., 220 F.3d 495, 501 (7th Cir. 2000). 15 Google gave unambiguous express and implied consent to a jury trial. The express consent 16 can be found in documents such as the parties’ Joint Submission re Trial Proposal, which stated, 17 for “Issues Triable to a Jury”: “The parties agree that all claims by all Plaintiffs are triable to a 18 jury, with the exception of the claims brought under California’s Unfair Competition Law, . . . , 19 and claims that the States have brought under the laws of 38 states other than California.” Dkt. 20 No. 505 at 3. The record also shows that the Court and the parties contemplated a jury trial for 21 Epic’s antitrust claims for years, without objection by Google and with its active participation in 22 the filing and discussion of jury instructions, proposed voir dire, and motions in limine. 23 Google’s filing on November 1, 2023, one day before jury selection and two court days 24 before the start of trial, was the first time it said it was “withdraw[ing] that consent.” See Dkt. No. 25 730 at 7. That was far too late. See Bereda, 865 F.2d at 55 (“Rule 39(c) does not permit the 26 district court to withdraw its prior consent to the litigants’ request for a nonadvisory jury.”); 27 Thompson, 963 F.2d at 889 (“Even if the court was correct that no jury trial right existed in this 28 case, F.R.Civ.Pro. 39(c) permits both sides to stipulate to a jury trial. To be sure, a district court 27 1 does not have to go along with the stipulation, but once that occurs, it does not have unbridled 2 discretion to change its mind.”); see also AMF Tuboscope, Inc. v. Cunningham, 352 F.2d 150, 155 3 (10th Cir. 1965) (where parties had stipulated to a jury trial that was set to begin on March 1, 4 abuse of discretion for district court to vacate the jury trial on February 28 and re-set the case for a 5 bench trial, based on court’s conclusion that the parties were not entitled to a jury trial as of right). Allowing Google to withdraw its consent two court days before trial would have caused 6 7 immense prejudice to Epic, which had been awaiting its day before a factfinder since filing its case 8 years prior, and which had spent many months preparing for a jury trial. A jury trial was proper, 9 and the jury’s verdict is properly treated as binding. CONCLUSION 10 Google’s motion did not present good grounds for judgment as a matter of law or for a new United States District Court Northern District of California 11 12 13 14 trial. IT IS SO ORDERED. Dated: July 3, 2024 15 16 JAMES DONATO United States District Judge 17 18 19 20 21 22 23 24 25 26 27 28 28

Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.

Why Is My Information Online?