Google's Monopoly: Defaults Have A Powerful Effect On Users’ Search Behavior

United States of America v. Google LLC., Court Filing, retrieved on April 30, 2024, is part of HackerNoon’s Legal PDF Series. You can jump to any part of this filing here. This part is 22 of 37. D. Defaults Have A Powerful Effect On Users’ Search Behavior, Particularly On Mobile Devices 867. Defaults are powerful in virtually all contexts. Tr. 526:22–25 (Rangel (Pls. Expert)) (discussing consensus in the field of behavioral economics about default effects). Search is no different. In search, this is demonstrated by the massive sums Google pays each year to ensure it maintains its search defaults status. Infra ¶¶ 932–943 (§ VI.D.6). In fact, Google spends more money securing exclusive defaults than on all other search-related expenses combined. Tr. 7576:9–7577:6 (Raghavan (Google)) (discussing UPX7002.A at 1). Many general-search queries are controlled by the power of defaults, and that is especially true on mobile devices. 1. The “Power Of Defaults” 868. A default is an option pre-selected for a consumer by a third-party, such as a smartphone manufacturer, that requires an affirmative action by a consumer to change. Tr. 520:7–25 (Rangel (Pls. Expert)) (defining defaults from the perspective of behavioral economics); Tr. 154:3–7 (Varian (Google)) (describing a default as a “capability that’s available before the user takes any action”). A default search engine is the search engine that comes preselected on a search access point when a user first begins using the access point. Des. Tr. 68:19– 25 (Ramalingam (Yahoo) Dep.) (the default is the one set “[w]ithout any user input”); Tr. 322:19–323:4 (Barton (Google)) (the default search is “the one that comes out of the box” on a browser, desktop, or mobile phone). 869. Across a range of domains, defaults have a powerful impact on consumer decisions. Tr. 520: 7–25, 526:22–529:16 (Rangel (Pls. Expert)) (discussing impact of defaults on consumer decision-making and various real-world default biases including coffee-shop tipping, 401(k) plans, and organ donation). 870. In its own businesses, Google has long understood and embraced the “power of defaults,” a term first evangelized by Dr. Varian in 2007. UPX1001 at -465 (referencing Dr. Varian’s “‘power of defaults’ idea”). As one Google employee explained, “the default options presented (in anything from finance to gaming) are very powerful, and will probably end up being what most people choose (out of lack of knowledge about customization, or convenience).” UPX0081 at -438. 871. In 2018, Google’s internal behavioral economics[15] team used the power of defaults to manipulate ad spending by creating a $10 default for advertisers’ maximum daily budget. Tr. 532:22–535:12 (Rangel (Pls. Expert)) (discussing UPX0101 at -285–86). Despite being visible to advertisers and easy to change, the new default budget had a powerful impact on Google’s advertisers as 64% of “low-budget” spenders (advertisers who previously set their daily maximum to less than $10 per day) increased their daily budget. Id. 532:22–535:12 (discussing UPX0101 at -285–86). However, Google also found the default “too powerful” because it dragged down spending by advertisers who had been setting their daily maximums above $10. Id. 872. Google also recognizes the power of the few defaults that it does not own but instead calls these the “tyranny of default.” UPX0111 at -825 (June 20, 2017 memo on “Google Branding on Search Desktop”); UPX0099 at -390 (Dec. 12, 2016 Email from Ben Friedenson (Google) analyzing Microsoft Edge default settings and observing either a) ‘tyranny of the default’ or b) a surprisingly high preference for/loyalty to Bing” among Edge users). In 2005, Google complained to Microsoft about its decision to preset a default search engine in the Internet Explorer browser. Tr. 7677:5–7681:6 (Pichai (Google)) (contextualizing UPX0172). Google wrote, “As you know, most end users do not change defaults,” and only “a tiny fraction of end users . . . try to change the default.” UPX0172 at -731; supra ¶ 742. Google accused Microsoft of “put[ting] its own interests above those of end users.” UPX0172 at -731. Because users do not change defaults, Google was “deeply concerned” that Microsoft’s decision to preset its own search engine as the default in Internet Explorer could harm the competitive process. Tr. 7693:5–8 (Pichai (Google)) (confirming concerns written in UPX0172). 873. Google proposed that Microsoft implement, instead of a preset default, a choice screen that would allow the end user to select their default engine the first time he or she used the Internet Explorer search box. UPX0172 at -731. A choice screen would place “control in the hands of the end user, where it belongs.” Id. Giving the user choice of search engine and a straightforward means of changing it was “indispensable both to preserve competition on the merits in search and to avoid distortion of competition in related markets.” Id. at -732; Tr. 7685:9–15, 7686:1–6 (Pichai (Google)) (Google proposed a choice screen to benefit users by aligning with user preference). 874. Other industry participants recognize the power of defaults on consumer behavior as well. Considering a request from Google that Apple share more user-location data, Apple executives said having “data sharing” set to “off” by default “wouldn’t have any real effect since so few people change defaults.” UPX0094 at -029. Apple recently demonstrated the power of defaults in a public way; the company changed from a default to a choice screen for cross-app tracking on iPhones, which resulted in 80% of users forbidding cross-app tracking even though they had previously acceded to tracking by default. Tr. 635:15–637:8 (Rangel (Pls. Expert)) (discussing UPXD101 at 55). 2. Search Engine Defaults Are The Most Efficient Method Of Distribution 875. Defaults can generate a bias in favor of the default option even when the default is easy to change, when the stakes of the decision are high, such as with end-of-life care, and when the person making the decision is an expert, such as in medication prescriptions. Tr. 529:17– 531:21, 532:18–535:12 (Rangel (Pls. Expert)). 876. By comparison, using a GSE is a low-stakes and low-expertise activity, so default bias is especially strong. Id. 530:10–531:21, 532:18–535:12. Default effects in search are “sizeable and robust.” Id. 523:14–24, 541:2–22; Tr. 5749:17–5750:22 (Whinston (Pls. Expert)) (“[T]he power of defaults is very significant.”). Even Google’s expert, Prof. Murphy, agreed that empirical evidence suggests that having a default generates “additional search volume.” Tr. 9941:9–11 (Murphy (Def. Expert)). 877. Dr. Ramaswamy testified that “being the default on the browser is the most efficient way to get [a search engine] into the hands of your users” because of the “convenience of easy accessibility and” tapping into . . . engrained default behaviors are deciding factors when it comes to whether a search engine gets lots of usage.” Tr. 3689:15–24 (Ramaswamy (Neeva)); Tr. 1961:20–1962:5 (Weinberg (DuckDuckGo)) (being a search access point’s default search engine is the most successful method of distributing a search engine). 878. In a 2007 study, Google examined the effect of browser home page defaults and query share. UPX0123; UPX0093 at -904. In the study, Google found that, when considering the factors that might influence a user’s choice of search engines—such as results quality, search features, user experience, and brand strength—“one factor surprisingly trumps them all: the default home page setting.” UPX0093 at -904 (Abstract for 2007 Google study “On the Strategic Value of the Default Home Page to Google”); UPX0124 at -036 (Aug. 10, 2007 presentation for Mr. Pichai (Google) and Ms. Braddi (Google) showing the “strong relationship between home page and search share”). Google further found that defaults were “directly correlated” with query share. UPX0093 at -904. Google showed that “users who have home page set to Google do 50% more searches on Google compared to those who don’t.” UPX0093 at -904. At the time, Dr. Varian concluded that the study was “great stuff” and “a very convincing case.” UPX1001 at -465. 879. Dr. Varian also agreed with (and liked) the study’s conclusion that a default home page—a form of default distribution popular at the time—“can be a powerful strategic weapon in the Search battle.” Tr. 170:15–24 (Varian (Google)) (reviewing UPX0123 at -485); UPX0123 at -487; UPX2049 (Executive Management Group (EMG) representatives “agreed that [it] is important” to “focus[] on homepage market share” because it is “one of the most effective things we can do to make gains in search market share”). In response to the study’s findings, Google began “taking these defaults very, very seriously.” Tr. 3710:15–3712:20, 3714:12–20 (Ramaswamy (Neeva)) (discussing UPX0093); UPX0093 at -904 (email from Dr. Ramaswamy, while working at Google: “This study is very cool! We should definitely put some marketing push behind it”). 880. In contrast to defaults, app downloads and marketing are ineffective methods of distribution; the availability of search apps in mobile application stores—such as the Play Store or Apple’s App Store—do not erode the effectiveness of default distribution. Tr. 617:23–618:17 (Rangel (Pls. Expert)). A Google survey found that more than a third of iPhone users did not know that they could download the Google Search App and that the majority of those who were aware of the ability to install the app still did not see the need to depart from Safari. UPX0139 at -149. Mozilla has found it “exceedingly difficult” to convince mobile users to download its browser application because “people have to find and install and you have to struggle if you can make them your default.” Des. Tr. 135:3–5, 135:8–22 (Baker (Mozilla) Dep.). “[M]erely having an app in the app store is a very difficult way to compete with the preloaded defaults... [b]ecause each person who gets that phone has to make a conscious decision to go through a lot of work to get to your product." Id. 139:20–23, 140:1–24. Apple also views searching on an application, such as the Google Seach App, as inferior to searching on a browser and that a search app therefore "wasn't the best thing for customers." Tr. 2493:18-2495:23, 2496:6-10 (Cue (Apple)). Google's analysis about search on Apple devices has similarly found that more than half of iPhone users did not have the Google Search App installed and even about half of the users who did still preferred using Safari. UPX0139 at -150. 881. Paid marketing for GSEs, browsers, and search apps is also an ineffective distribution method. UPX0122 at -960 (illustrating various distribution methods for Google services and calling pre-installation the "highest value way to acquire users"). In 2016, Emily Moxley, a Google vice president, “guarantee[d]” that marketing Google Search in Windows 10 would "make no difference" in generating usage. PSX00216 at -126 (“Didn't work for safari, won't work here!"). A January 2015 Google "Post-Mortem" on Firefox's 2014 default switch to Yahoo found that marketing efforts were immaterial. UPX0145 at -441, -450 (“Early results indicate that Chrome/DSE/DHP promos have little/no effect."); UPX0169 at -180 (email to Sundar Pichai (Google) stating "user promos [on Firefox] haven't materially returned users to Google"). DuckDuckGo has found marketing to be unhelpful in generating traffic because teaching people to change the default search engines is difficult. Tr. 1957:15–1958:5 (Weinberg (DuckDuckGo)). 882. Ultimately, a default search engine garners “enormous usage simply by the power of the default.” Tr. 3710:7-3712:20 (Ramaswamy (Neeva)) (“[W]hoever controls that search box gets a lot of usage independent of the merits of the search engine. And so you get enormous usage simply by the power of the default.”). According to Microsoft CEO Satya Nadella: “The entire notion that users have choice and they go from one website to one website or one search into one search and it’s complete bogus. There’s defaults. The only thing that matter in terms of changing search behavior.” Tr. 3497:13–3498:4 (Nadella (Microsoft)). “Whoever owns [the] default will gain share no matter what the quality.” Tr. 2722:2–10 (Parakhin (Microsoft)) (“It is, I believe well-documented fact that people very rarely switch defaults.”). 883. Indeed, most search traffic arrives through defaults rather than through nondefault querying. UPX0083 at -967 (61% of Google’s queries in 2017 arrived at Google via a paid distribution deal, which was a steady increase from 48% in 2014, 52% in 2015, and 57% in 2016.); UPX1050 at -894 (65% of Google’s Apple Traffic arrives through “Safari Default Search” compared to just 13% through “Safari Organic Homepage”); Tr. 23:25–25:2 (Sept. 19, 2023 sealed PM session (Yoo (Google)) (Android “organic” queries have declined over time.); id. 25:16–27:3 (less than 10% of Android revenue comes from “organic” queries); Tr. 3102:11– 3104:25 (Tinter (Microsoft)) (discussing greater use of search defaults than non-default querying); UPX0080 at -509 (Email from Ms. Braddi (Google) stating that Google generated almost three times as much revenue on iOS from default traffic as “organic” revenue). 3. Behavioral Economics Explains The Power Of Defaults 884. The behavioral economics concepts of habit, choice friction, and user confusion and awareness explain the preeminence of default distribution. Tr. 541:2–22, 542:25–543:18, 547:16–549:3 (Rangel (Pls. Expert)). a) Habit 885. Habit is when consumers make decisions automatically based on prior outcomes rather than expending cognitive capacity to make an explicit choice. Tr. 541:23–542:24 (Rangel (Pls. Expert)) (consumers use habit to preserve limited cognitive capacity by, for example, requesting the same order at a coffee shop rather than considering every possible option on each visit). Because search is an activity carried out repetitively, on a familiar interface, and with immediate feedback in the form of a search result, search “[h]abits develop very strongly.” Id. 542:25–543:9. 886. Google’s users have identified “habit”—along with convenience and ease of access—as the reason why they continue to use the “Google Safari Default.” UPX0090 at -940– 41 (2010 Google iPhone user study respondent: “I use Google out of habit. There is no difference between the way the results are presented from other engines.”). 887. Google itself recognizes the impact of habit on consumer behavior in search. To address the threat of competition from Microsoft’s search product in 2003, Dr. Varian recommended that Google “[g]et users addicted to our interface and tools.” UPX0151 at -158– 59; Tr. 140:22–142:9, 144:13–19 (Varian (Google)) (discussing context of UPX0151); Tr. 545:12–546:8 (Rangel (Pls. Expert)) (discussing UPX0151 at -158–59 as reflecting what behavioral economists would consider “an extreme habit”). The 2007 Google study examining the effect of browser home page defaults and query share, supra ¶ 878, attributed increased searching when Google was the default homepage to the fact that “[u]sers do not always make a deliberate choice of search engine.” UPX0123 at -485 (“Most users stay with pre-configured home page settings”); Tr. 551:10–552:13 (Rangel (Pls. Expert)) (habit is the opposite of a consumers’ deliberate choice). And Google pursued default exclusivity in its search distribution contracts precisely because it afforded Google “an opportunity to be discovered and used on a repeating basis.” Tr. 338:13–17 (Barton (Google)). 888. Mr. Nadella explained that the Google search habit has become ingrained like many other common habits; he testified, “[Y]ou get up in the morning, you brush your teeth and you search on Google. And so therefore, with that such level of habit forming, the only way to change is by changing defaults.” Tr. 3497:13–3498:4 (Nadella (Microsoft)). Dr. Ramaswamy similarly asserted that “the convenience of easy accessibility and tapping into . . . engrained behaviors are the deciding factors when it comes to whether a search engine gets a lot of usage. . . . [D]efaults become habits.” Tr. 3689:15–3690:4 (Ramaswamy (Neeva)). 889. Habit is so profound in search that generative AI is unlikely to meaningfully impact Google’s search usage share. Tr. 3512:8–3513:9 (Nadella (Microsoft)) (Generative AI “doesn’t address all the other challenges . . . which is the user habit [that] comes from search defaults, settings, so the distribution advantage Google has today doesn’t go away.”). Dr. Raghavan agreed that ChatGPT “hasn’t put a dent in Google’s market share.” Tr. 7533:6– 7534:19 (Raghavan (Google)). 890. As a result of controlling the defaults on search access points, habit creates bias in Google’s favor. Tr. 541:23–542:24 (Rangel (Pls. Expert)). b) Choice Friction 891. On the infrequent occasions when a consumer makes an explicit choice of search engine, choice friction causes Google’s default to remain sticky. Tr. 541:2–22 (Rangel (Pls. Expert)) (changing only the choice architecture for implementing a consumer’s decision, such as the number of steps to execute the decision or the difficulty of executing those steps, can powerfully affect consumer decision-making); id. 551:10–552:13 (in search, users act in “explicit or deliberative” decision-making mode only “infrequently”). Increasing choice friction increases the stickiness of a particular default. Id. 554:4–21 (“[T]he more choice friction it takes to change the defaults, the stickier the defaults are.”). 892. Google’s internal behavioral economics team has taught the company that even “[s]eemingly small friction points in user experiences can have a dramatically disproportionate effect” on consumers’ decisions. UPX0103 at -214; UPX0848 at -612 (“each barrier in terms of time attention, and demand for thinking is 10X what you think it is”). A June 2019 presentation about private searching on Google, prepared for Google’s product counsel, observed that Google was failing to “fully address user needs” because “multiple toggles and settings to get into a more ‘private state’ outside of Incognito . . . create user friction and are difficult to understand.” UPX0811 at -413. 893. In search, Dr. Rangel found that bias towards Google’s defaults arises from at least four sources of choice friction faced by users changing the pre-set default search engine. First, a user must become aware that there is a default search engine and that it can be changed. Tr. 552:15–553:25 (Rangel (Pls. Expert)); Tr. 2636:3–9 (Cue (Apple)) (before changing the default search engine on an iPhone, a user must know that a search engine is pre-selected as the default). Second, the user must discover what alternative options to the default search engine exist. Tr. 552:15–553:25 (Rangel (Pls. Expert)). Third, the consumer must learn the steps necessary to change the default. Id.; Tr. 2636:10–2637:8 (Cue (Apple)) (an iPhone user must understand how to change the default search engine on iPhone, which may require going to Apple, Google, or Bing’s support websites to learn the steps). Fourth, the consumer must actually take the steps necessary to implement the change. Tr. 552:15–553:25 (Rangel (Pls. Expert)). 894. On Apple’s iPhone, the steps to change the default search engine require overcoming “substantial” choice friction. Tr. 562:21–563:8 (Rangel (Pls. Expert)); Tr. 2169:23– 2170:20 (Giannandrea (Apple)) (reviewing the four steps to change the default search engine on iPhone in UPXD004); Tr. 2630:7-16 (Cue (Apple)) (describing four steps to change the default search engine on iPhone); Des. Tr. 55:20–56:14 (Ribas (Microsoft) Dep.) (“even though, in theory, it's possible to [change the Safari default on iPhones], it's not" and even "some [Microsoft] corporate vice presidents” and “smart engineers" have asked how to make a change). 895. An Android user who has already overcome the choice frictions of understanding that there is a default, discovering the alternatives, and learning the steps to change the Google default search widget must still execute ten steps to implement the change, which constitutes "considerable choice friction" for the user. Tr. 559:14–562:20 (Rangel (Pls. Expert)) (reviewing UPXD101 at 21–38); Tr. 979:19–980:2 (Kolotouros (Google)) (explaining that DX0738 at 2-3 shows that replacing a search widget requires one swipe, two touch-and-holds, three taps, and a selection); Tr. 2722:11-13 (Parakhin (Microsoft)) (requiring multiple clicks to change a search engine default makes doing so proportionally more difficult). 896. It also takes several clicks to change the default search engine on Mozilla's Firefox browser. Des. Tr. 187:18-20, 187:23 (Baker (Mozilla) Dep.). Changing the default search engine in other browsers is often even more difficult. Id. 49:16–23 (competing browsers' preset default search engines are “often not nearly as easy to change the default as in Firefox”); Tr. 1958:15–1961:8 (Weinberg (DuckDuckGo)) (Getting people to switch defaults is difficult because "it's all just way harder than it needs to be."). 897. Consequently, changing the default search engine requires “mental effort” and is not a "zero cost" action. Tr. 563:9-25 (Rangel (Pls. Expert)); Tr. 3491:23-3493:8 (Nadella (Microsoft)) ("changing of defaults is hard"); Tr. 3796:5–3798:22 (Ramaswamy (Neeva)) (dismissing the “pious prose around 'competition being a click away""). In contrast, controlling defaults boosts Google by making it “very, very seamless and easy for users” to access Google. Tr. 7661:1–11 (Pichai (Google)) (“[W]e know [making Google the default on Safari] would lead to increased usage of our products and services, particularly Google Search in this case. So there’s clear value in that, and that’s what we’re looking to do.”); Tr. 1537:9–12 (Roszak (Google)) (a benefit of Google’s default distribution deals is reducing choice friction). 898. Google has long recognized and even harnessed the effect of choice friction to protect its defaults. Tr. 554:17–558:19 (Rangel (Pls. Expert)) (discussing representative examples of Google recognizing and harnessing choice friction). For as long as the company has required it, Google’s default exclusivity has ensured that a user cannot change the default search engine to a rival without going into the device’s settings. Tr. 324:14–325:11 (Barton (Google)). This makes changing default search engines “difficult.” Id. 338:18–24 (explaining that users would “have a difficult time finding or changing [the default search engine] to Google” were it not pre-set as the default search engine). 899. Google’s MADAs with OEMs ingrain the importance of choice friction. Tr. 556:12–557:9 (Rangel (Pls. Expert)). In January 2020, Google amended the MADA to restrict OEMs from making it easier for a user to change a default by prohibiting OEMs from including “processes, instructions, promotions, or other means that directs, instructs or encourages the End User to change” their default search engine away from Google. Id.; UPX5358 at -004 (§ 2.5) (LG MADA (2020 amend.); JX0099 at -998 (§ 2.9) (Motorola MADA (2020 amend.)); UPX5511 at -987–88 (§ 2.9) (Samsung MADA (2020 amend.)); Tr. 799:14– 801:21 (Kolotouros (Google)) (agreeing that MADA’s amendments prohibited OEMs from instructing or encouraging a user to change default settings or placements). 900. Google’s new prohibition against instructing or even encouraging a user to make changes introduced choice friction by impairing users’ awareness of the ability to change a default. Tr. 556:12-557:9 (Rangel (Pls. Expert)) (analyzing new prohibitions in the context of a Samsung MADA amendment). 901. google has insisted on more default search friction, even when the partner preferred less. In 2018, Samsung changed the interface in its S browser to add a drop-down menu in the address bar that would allow a consumer to easily and permanently change their default search engine. Supra¶¶ 296, 808; Tr. 856:2-857:17 (Kolotouros (google)) (Samsung’s S Browser included a dropdown in the address bar that this changed the S Browser default search engine). google informed Samsung that this change breached the RSA and considered terminating Samsung’s RSA. Id. 857:14-858:1; Tr. 557:18 (Rangel (Pls. Expert)); UPX1011 at -289 (Google considered terminating Samsung’s RSA due to a potential fourth strike). Samsung acceded to Google’s demands and threats, and users who might wish to change the search default must take additional steps to do so. Tr. 997:1-6, 887:24-888:13,890:3-9 (Kolotouros Google)). c) User Confusion and Unawareness 902. Default bias in favor of Google is bolstered further by user confusion and unawareness about what a default search engine is or even that it can be changed. Tr. 547:16-549:3 (Rangel ({ls. Expert)) (“[M]any [users] are confused. They don’t know that there is a default. They don’t know what the default search engine is. They don’t know they can be changed.”); id. 552:15-553:25 (“[M]any people are not aware [that there is a default search engine], and people are confused, even about what’s the default search engine that they have.“). 903. Many internet users do not pay attention to their search engine and instead simply use the default search engine in their browser. Tr. 3685:15--3686:19 (Ramaswamy (Neeva)) (testifying about UPX0940 at -492, which found that almost half of surveyed users search using the “Browser’s default“ GSE). For example, one 2020 Google study found that more than 80% of Google’s daily active users on Apple’s iOS access Google through Apple’s pre-set default browser, Safari, and were “often unaware they’re using Google.” UPX2051 at -519–20 (user unawareness of Google use in Safari “a risk for [Google]”). Google’s expert Prof. Murphy agreed that some users are probably unaware that they even have alternatives for their default search engine. Tr. 9942:7–10 (Murphy (Def. Expert)). 904. Often, even tech-savvy consumers do not know the difference between a browser and a search engine. Tr. 3685:15–3686:19 (Ramaswamy (Neeva)) (user feedback revealed that “many perfectly tech-savvy people could not tell the difference between a browser and a search engine”); Tr. 702:5–23 (Rangel (Pls. Expert)) (users “don’t understand the difference between searching and browser”). Google interviewed Microsoft users in 2017 and found they were confused about the difference between browsers, search engines, and even email. UPX0066 at -073 (recording feedback from one Bing user who said, “Most of the time I use Google, everybody knows that”); PSX00216 at -126 (“one fundamental issue [with marketing Google in Windows 10] is that users on Edge don’t even realize they aren’t using Google, and they don’t feel they are missing out”). 4. Market Evidence Confirms Default Effects In Search 905. Because defaults “have a strong impact” with a “75% take rate,” in 2007, Google leadership rejected Apple’s request to amend the ISA to allow Apple to implement a search engine choice screen in Apple’s forthcoming version of the Safari browser for Windows. UPX0137 at -688–89; Tr. 7691:19–7692:13 (Pichai (Google)); Tr. 4980:13–4983:8 (Braddi (Google)) (describing meeting). In 2010, Google found that it “will lose more than 50% of Google users (and more non-Google users) if the [Safari mobile] default search engine changes to something else.” UPX0070 at -657. 906. In 2012, Apple replaced Google Maps with Apple Maps as the default maps application on iOS, which created an immediate, sizeable, and lasting change in usage. Tr. 564:11–566:1 (Rangel (Pls. Expert)) (discussing reproduction of chart in UPX0097 at -.001, showing a large and sudden drop in Google Maps usage on iOS when Apple changed the default while, during the same period, usage of Google Maps on Android increased); Tr. 1558:3–5, 1560:4–20, 1563:25–1566:4 (Roszak (Google)) (Google Maps lost 60% of its usage vis-à-vis its peak usage—when it was the default maps application on iOS). 907. In 2012, Apple publicly apologized for Apple Maps’ poor performance. Tr. 565:11–566:1 (Rangel (Pls. Expert)); Tr. 1569:15–18 (Roszak (Google)). As Prof. Rangel explained, the Apple Maps event illustrates that defaults are so powerful that they can generate bias among users “even when the default is assigned to an inferior product.” Tr. 565:11–566:1 (Rangel (Pls. Expert)). According to Mr. Nadella, the 2012 Apple Maps event demonstrated “the power of defaults” and gave Microsoft confidence that winning the iOS default would be a “big game-changer” in search competition. Tr. 3500:9–3502:2 (Nadella (Microsoft)) (“[Apple Maps is] exhibit number one on anybody who wants to sort of know anything about sort of the power of defaults, that case study is the best.”). 908. Mozilla’s 2014 decision to switch the default search engine on its Firefox browser from Google to Yahoo sharply and immediately shifted approximately 20% of queries from Google to Yahoo—a phenomenon that was reversed in 2017 when Mozilla reverted the Firefox default to Google. Tr. 628:6–20, 630:8–631:14, 660:12–661:18 (Rangel (Pls. Expert)) (describing Firefox default change and referring to redacted query share shifts in UPXD101 at 55); Tr. 5729:25–5731:4 (Whinston (Pls. Expert)) (discussing redacted query share shifts in UPXD104 at 23 and explaining, based on “jump[s]” in query shares for Yahoo and Google in 2014 and 2017, respectively, that “[i]f defaults didn’t matter, all of this would have been a nice, smooth—maybe not line, but curve. You wouldn’t have seen these jumps”); Des. Tr. 62:9–18 (Baker (Mozilla) Dep.) (describing Firefox default switch in 2014). Mozilla’s default switch cost Google an estimated half-a-billion dollars in 2015 along with a 30% decline in traffic from Firefox and a 45% decline in revenues from Firefox. Tr. 1610:2–1611:14 (Roszak (Google)) (discussing UPX0066 at -071). 909. As significant as these Firefox default effects were, they likely understate the power of the defaults controlled by the challenged agreements. First, because Firefox is not a pre-set default browser on any device in the United States, Firefox users are generally more tech savvy and therefore more likely than the average user to understand defaults and make a change. Tr. 628:6–20 (Rangel (Pls. Expert)). Second, Firefox is used almost exclusively on desktop, where defaults are less sticky than on mobile. Tr. 5731:6–21 (Whinston (Pls. Expert)) (lower switching is the result of Firefox being used primarily on desktop rather than mobile, which is more difficult to switch); Tr. 628:22–629:5 (Rangel (Pls. Expert)) (vast majority of Firefox queries come from desktops); infra ¶¶ 918–923. 910. Choice screens implemented by Google on Android devices in Russia and Europe illustrate the default effect by showing the consequence of removing a default. Tr. 610:6–611:18 (Rangel (Pls. Expert)). In 2017, Russian competition authorities required Google to implement a choice screen on Android devices, which resulted in a consistent and continuing loss in Google’s Android query share that was picked up by the Russian search engine Yandex. Tr. 2673:12– 2674:25 (Parakhin (Microsoft)) (describing Google’s settlement with Russian regulators and explaining that Yandex was not able to maintain market share on mobile before 2017 because “very few users change defaults”); Tr. 608:14–609:9, 609:15–610:9, 610:19–612:1 (Rangel (Pls. Expert)) (discussing query share shifts with reference to redacted figures in UPXD101 at 44); Tr. 10001:23–10002:10 (Murphy (Def. Expert)) (agreeing that the choice screen contributed to Google’s share decline in Russia); UPX0170 at -978 (depicting an approximate 19% search share loss for Google between 2017 and 2019 that resulted in Yandex overtaking Google as the leading search engine in Russia). Moreover, on Russian iPhones and desktops, where no choice screen was implemented and Google’s default persisted, query shares held constant, demonstrating that the choice screen was in fact responsible for the Android share changes. Tr. 612:4–613:5 (Rangel (Pls. Expert)) (referring to UPXD101 at 45); Tr. 5744:7–20 (Whinston (Pls. Expert)) (referring to UPXD104 at 27). Google agreed that the choice screen was a “key” cause for Google’s loss in Russian query share. UPX0170 at -979; UPX0760 at -596 (“[The Russian remedy] implementation (2016/17) predictably led to a decline Google’s queries share on mobile.”). 911. Similarly, the European choice screen implemented in 2021 following Google’s Android settlement with the European Commission shifted query shares and generated competition that had not existed when Google was the default across Android devices. UPX0764 at -077 (Quality-enhancing efforts undertaken by Google as part of the “Go Big in Europe” initiative in response to the choice screen included improving unique features, local content, and trust). Tr. 620:22–624:4 (Rangel (Pls. Expert)) (referring to UPXD101 at 49); Tr. 5735:1– 5737:19 (Whinston (Pls. Expert)) (referring to UPXD104 at 25) (Google’s selection on the choice screen was lower than its market share before adoption of the choice screen). 912. A regression analysis using data from the European choice screen predicts that Google’s U.S. rivals would acquire about 10% market share were a choice screen implemented in the United States. Tr. 5737:21–5738:10 (Whinston (Pls. Expert)). And although Google retained a high share in Europe following the choice screen, the “default echo” deriving from Google’s incumbency and brand recognition makes this outcome unsurprising and was, in fact, anticipated by Google. Tr. 622:25–625:14 (Rangel (Pls. Expert)); Tr. 5732:15–5734:18, 5735:1– 5738:22 (Whinston (Pls. Expert)) (referring to UPXD104 at 25 and explaining that the European choice screen showed consumer preferences at the time the choice screen was instituted and that there would be greater share moving away from Google if rivals were stronger); UPX1103 at -775 (Google email discussing the importance of brand recognition in differing outcomes with European and Russian choice screen experiments); supra ¶ 544 (discussing effect of brand recognition in Europe). 913. Finally, Google’s query share differentials between browsers where Google is the default—such as Chrome, Firefox, and Safari—compared to Microsoft’s Internet Explorer and Edge browsers, where Google is not the default, show that defaults drive search engine query shares. Tr. 5744:21–5745:20 (Whinston (Pls. Expert)) (referring to UPXD104 at 28–29); UPX0955 at -774 (2020 Email from Brave executive reporting that setting DuckDuckGo as the default search engine in Brave’s browser in Germany, Ireland, Australia, and New Zealand showed the “data is very clear: default is sticky and very powerful.”). In the context of desktop, Bing’s query share on Windows computers, where it is the default search engine on the only preinstalled browser, compared to Bing’s query share on Apple’s Mac computers, where Google is the default search engine on the only preinstalled browser, shows a near seven-fold default effect. Tr. 631:23–632:17 (Rangel (Pls. Expert)) (referring to UPXD101 at 56). 5. Ordinary-Course Modelling For Multi-Billion Dollar Decisions Confirms That Search Defaults Have High User Retention Rates, Especially On Mobile 914. In 2016, Apple and Google renegotiated the ISA, a multi-billion dollar agreement. Tr. 1630:10–14 (Roszak (Google)) (agreeing that the Safari default’s worth to Google was “an important question” that had “billion of dollars on the line“); infra¶935 (overviewing size of Google payments to Apple). 915. For the negotiations, Google's finance team modeled the impact of losing Apple's Safari default, which Google considered a "Code Red" scenario. Tr. 1616:20-1618:15 (Roszak (Google)); Tr. 536:10–537:21 (Rangel (Pls. Expert)) (discussing UPX0171 at -169, -177); UPX0171 at -169, -177 (acknowledging that Google was “vulnerable” if Apple switched the default search engine); UPX0085 at -228, -230, -240 (discussing "Code Red" in the context of Project Soy, which was code for clawing back users on Apple's iOS). 916. To model the outcomes of losing the search default on Apple devices, Google applied the revenue shifts from the default losses on Firefox, supra ¶ 908, and Apple Maps, supra ¶906. UPX1050 at -828, -886 (January 2016 modelling relying on Firefox for desktop recovery and on Apple Maps for iPhone recovery); Tr. 1620:13–1631:18 (Roszak (Google)) (reviewing UPX1050 at-828, -886); UPX6024 at -440, -443 (written 30(b)(6) response: Google used Firefox and Apple Maps events "to model certain clawback assumptions" during 2016 ISA negotiations with Apple.). 917. For desktop, Google based its modeling on the outcome of Mozilla's changing the Firefox default from Google to Yahoo in 2014 and concluded that Google would lose [redacted]% of its revenue ([redacted]% recovery) if it lost the Safari default on desktop. UPX1050 at -828, -886; Tr. 1620:13-1621:25, 1626:13–1627:10 (Roszak (Google)) (reviewing UPX1050 at -828, -886). 918. For mobile, Google based its modeling on the outcome of Apple's switching the iOS maps application default from Google Maps to Apple Maps in 2012 and concluded that it would lose [redacted]% of its revenue ([redacted]% recovery) if it lost the Safari default on mobile. UPX1050 at -828, -886; Tr. 1622:1–1624:16, 1626:13–1627:10 (Roszak (Google)). 919. In other words, Google predicted that its revenue loss on mobile (redacted%) would be more than double its revenue loss on desktop (redacted%). UPX1050 at -886; Tr. 1629:14–18 (Roszak (Google)). 920. The disparity between the default effect on mobile and desktop arises from the fact that “[d]efaults have more prominence in mobile due to screen size and UI [user interface].” UPX1050 at -886; Tr. 1628:8–1629:3 (Roszak (Google)). As Google explained in a later analysis, users “are much less likely to change default search engine on mobile.” UPX0139 at -119 (2018 Google presentation discussing the “[s]ignificant risk if [iPhone] default [search] engine is changed”). Mobile and desktop user experiences have “salient difference[s],” including the “very limited real state” on mobile devices. Tr. 6310:23–6311:16 (Nayak (Google)). 921. These form factor differences between mobile and desktop, such as smaller screen size, are an example of choice friction that enhances the power of defaults on mobile devices. Tr. 523:14–24, 625:17–626:6 (Rangel (Pls. Expert)); UPX0084 at -728 (Google 2016 2Q Earnings Diligence – TAC Appendix stating, “User behavior is more heavily influenced by default settings on mobile and tablet”); Tr. 3498:5–19 (Nadella (Microsoft)) (changing defaults is easier on desktop than on mobile because “there are many, many sort of friction points on mobile operating systems”). 922. Mobile users do, in fact, query mostly through defaults. Tr. 1632:3–14 (Roszak (Google)) (65% of Google’s queries from Apple iOS came through the Safari default); Tr. 3102:11–3104:25 (Tinter (Microsoft)) (fewer users reach search engines directly on mobile devices and instead rely on search access points with pre-set defaults); Tr. 3499:21–3500:8 (Nadella (Microsoft)) (although mobile devices have “multiple search access points, the one access point that matters is the search default on the browser”); UPX0083 at -967 (25% of mobile queries are non-default compared to 32% on tablet and 74% on desktop); UPX0115 at -142 (Microsoft estimates that defaults are responsible for 68% of mobile search volume). In 2020, Google found that 81% of all of Google’s daily active users accessed Google through Safari, the browser preloaded on iPhones where Google is the default. Tr. 7512:23–7514:8 (Raghavan (Google)) (discussing UPX2051 at -519). 923. The enhanced power of defaults on mobile is significant because the vast majority of querying today occurs on mobile devices. Tr. 5798:17–5799:5 (Whinston (Pls. Expert)) (“it’s important to remember mobile is where the market is growing” and that desktop queries are “flat and have been for a long time”); Tr. 2279:20–2280:9 (Giannandrea (Apple)) (in the last few years more queries are being done on mobile than on desktop); Tr. 3098:6–3099:3 (Tinter (Microsoft)) (the general search services market has changed from being “predominantly PC to predominantly mobile”). 924. Consequently, weighting between mobile and desktop revenue mix on Apple devices, Google projected a total revenue recovery that skewed heavily toward its mobile prediction— redacted% total recovery (redacted% total loss). UPX1050 at -886; Tr. 1627:11–24 (Roszak (Google)). 925. The predictions derived from the Firefox and Apple Maps events were Google’s “best guess expectation” that were applied in multiple scenarios. Tr. 1622:21–1624:11, 1743:1– 1744:10 (Google (Roszak)). These findings were shared with Google CFO Ruth Porat. Id. 1619:23–1620:12. Mr. Roszak, the leader of the Google finance team assessing default deals, testified that he was unaware of any better data upon which to base revenue recovery assumptions than the Firefox and Apple Maps events. Id. 1638:13–1639:20. Google also used the same Firefox and Apple Maps events to make the same recovery assumptions five years later, in 2021, when considering the possibility of Apple entering the general search services market. UPX0148 at -826; Tr. 1633:16–21, 1635:21–1638:1 (Roszak (Google)). 926. The Apple Maps event has also been an input to model recovery rates if Google lost the default on Samsung’s S Browser and if Chrome and the Google Search App were not pre-installed on Android devices as well. Tr. 1640:19–1642:14 (Roszak (Google)) (discussing UPX0323 at –540, which projected [redacted]% to [redacted]% recovery loss based in part on “iOS Maps default change”); UPX0146 at -412 (“S Browser default changes and we clawback redacted%” based partly on “iOS Maps Recovery”); UPX6024 at -442–43 (written 30(b)(6) response: “[I]n analyzing provisions of the RSA in 2020, Google personnel used clawback estimates of between [redacted]% (based on Google’s experience with Apple Maps) to redacted% (based on Google’s experience on Microsoft’s Edge browser) to assess revenue related to default search on third-party browsers.”). 927. Nor was Google alone in its analysis; both Microsoft and Apple independently reached conclusions regarding query recovery that were similar to Google’s. Tr. 10497:2– 10500:5 (Whinston (Pls. Expert)). For its own analysis during the same negotiations in 2016, Apple assumed that % of queries would continue to run through the default were Apple to change the Safari default to Bing. UPX0095 at -331. This is nearly identical to Google’s weighted % revenue recovery (i.e., [redacted]% revenue loss) modelling. Supra ¶ 924. Two years later when considering acquiring or co-owning Bing, Apple—in estimates provided to CEO Tim Cook—predicted that a new, non-Google default search engine could retain between [redacted]% and [redacted]% of queries with an intermediate model of [redacted]% query retention. UPX0460 at -176; Des. Tr. 211:4–24, 212:23–213:12 (Perica (Apple) Dep.) (query retention rates of [redacted]%, [redacted]%, and [redacted]% in UPX0460 at -176 came from the relevant experts at Apple). 928. In 2016, Microsoft assumed that [redacted]% of mobile search volume going through the Safari default would shift to Bing if it were to become the default. Tr. 10497:2–10500:4 (Whinston (Pls. Expert)) (discussing UPXD106 at 19); UPX0069 at -236. This too resembles Google's modelled [redacted]% mobile revenue loss attributable to a default switch. Supra ¶918. Also like Google, Microsoft assumed "higher churn” on desktop “as it is easier for users to switch defaults on desktop browser.” UPX0115 at -142 (estimating that Microsoft could capture significantly greater query share on mobile iPhones (54%) than desktop Macs (16%)); Tr. 3264:23-3265:14 (Tinter (Microsoft)) ("Defaults matter less" on desktop). In fact, Microsoft even distinguishes "PC search" from "mobile search" because of differences in keyboards and screen sizes. Id. 3094:12–24. “On a mobile platform more than anywhere else, even more than on the PC, default is the only thing that matters." Des. Tr. 143:7-23 (van der Kooi (Microsoft) Dep.). 929. In a 2018 presentation to Apple, Microsoft projected that obtaining the Safari default could boost Microsoft's query share on iPhones by 55 percentage points (or 29 fold), from approximately 2% to 57%. UPX0116 at .028; Tr. 3278:10–3279:15 (Tinter (Microsoft)) (“This was the highest fidelity estimates that we had based on the best of our knowledge and our experience and our ability of what we thought was likely to happen”). And like Google, Microsoft used the Apple Maps event "to sanity check” Microsoft's assumptions for modelling a potential deal for the Safari default. Id. 3266:13-3268:3. 930. In sum, these similar models across three sophisticated companies are significant—and more than “thought experiments”—because they were used in very detailed ways to make business decisions about enormous sums of money. Tr. 5717:1-7, 5719:20-5721:25, 5751:1– 5752:5 (Whinston (Pls. Expert); supra ¶¶ 925-927 (Google's modelling were based on best available data and shared with CFO); supra ¶ 927 (Apple's estimates were based on analysis from experts and shared with CEO); supra ¶ 929 (Microsoft’s estimates were “highest fidelity estimates”). As Mr. Tinter explained, the accuracy of these models was important because they were being used as the basis for decisions about “very large financial investments” and as the basis for negotiations with partners. Tr. 3278:23–3279:15 (Tinter (Microsoft)). 931. Drawing from market participants’ estimates and real-world data, Prof. Whinston predicted that at least 33% of all U.S. queries—including 49% of all U.S. mobile queries—would shift from Google to its rivals if Google lost all the defaults controlled by the challenged contracts; this would quadruple rivals’ total U.S. query share. Tr. 5749:17–5752:5 (Whinston (Pls. Expert)) (explaining analysis with reference to UPXD104 at 31); id. 10497:2–10499:15 (explaining analysis with reference to UPXD106 at 19). 6. Google’s Enormous Payments And Microsoft’s Enormous Bids For Defaults Corroborate The Power Of Search Defaults 932. Because of the effect of defaults on query share, control of defaults are “very valuable.” Tr. 7684:18–20 (Pichai (Google)); id. 7666:19–24 (agreeing “Google pays for these exclusive defaults because they have value in distributing Google Search”); id. 7661:1–11 (“[T]here’s clear value” in making Google Search the default search engine in Safari.); Tr. 830:22–23 (Kolotorous (Google)) (agreeing Google finds defaults “valuable”); Tr. 4935:11– 18 (Braddi (Google)) (agreeing Apple’s Safari default “is valuable to Google because it drives more queries to Google Search”); Tr. 154:20–21 (Varian (Google)) (Being the default has “some advantages.”); Tr. 329:18–20 (Barton (Google)) (agreeing that he “view[ed] search defaults as “valuable”). 933. Because “the default is valuable,” Google pays “tens of billions of dollars every year” to be the default search engine across search access points in the United States. Tr. 7669:4–14 (Pichai (Google)). Asked by the Court, based on his experience at Google, for the business justification for Google’s payments to search distributors, Dr. Ramaswamy explained that defaults are “enormously powerful because . . . pious prose around ‘competition being a click away,’ notwithstanding in practice, they [users] don’t change.” Tr. 3796:5–3798:22 (Ramaswamy (Neeva)). Mr. Roszak, Google vice president of finance with long-time responsibilities related to the default distribution deals, is not aware of any default distribution deal that Google has ever entered where the traffic acquisition cost exceeded the expected incremental return. Tr. 1534:24–1536:21, 1540:19–22 (Roszak (Google)). 934. The total TAC payments incurred by Google’ s search business (i.e., including payments to partners for default distribution of Google Search) more than tripled between 2014 and 2021—from $7.1 billion to $26.3 billion. Tr. 7521:2–16, 7576:16–21 (Raghavan (Google)) (testifying about UPX0453 and UPX7002.A); UPX7002.A at .001. 935. Google’s overall worldwide revenue share payments to Apple were approximately $20 billion in 2022. Tr. 2492:22–2493:6 (Cue (Apple)). In fiscal year 2020, Google’s payments to Apple—the largest company in the world by market capitalization— constituted 17.5% of Apple’s operating income. Tr. 5727:5–5728:11 (Whinston (Pls. Expert)) (referring to UPXD104 at 19); UPX1110 at -41046 (For each quarter in fiscal year 2020, Google’s payments to Apple as a percentage of operating income were 21.5%, 19.2%, 21.1%, and 11.6%.); UPX0635 at -352 (Apple recognized in 2018 that “G search deal is [redacted]% of corporate EBIT.”); compare UPX8105 at -203 ($66.288 billion in Apple worldwide operating income for fiscal year 2020) with Tr. 2491:1–2492:8 (Cue (Apple)) (reviewing UPX1109 at -038, Apple’s fiscal year 2020 revenue share payments from Google were $11.58 billion.). Further, Google’s revenue-share payments to Apple have grown consistently. Id. 2489:23–2490:25 (discussing UPX1109 at -036, agreeing that Google’s monthly revenue-share payments to Apple “have been on a pretty steady upward trajectory” and grew from $418 million in January 2017 to $1.5 billion in August 2021); UPX0460 at -149 (Global ISA payments grew from $2.2 billion in 2014 to $8.6 billion in 2018.); UPX1109 at -026 (“Google reported $1,532M (+46% Y/Y) revenue share for the month of August ’21, recording the highest revenue share in history for the fourth straight month.”); Tr. 2488:2–2489:6 (Cue (Apple)) (The April 2021 record streak in Google revenue-share payments to Apple continued.). In fiscal year 2020, Google’s payments to Apple for Safari address-bar queries alone exceeded $9.6 billion and surpassed a billion dollars every month by May 2021. UPX1109 at -036, -038 (total of payments for “Safari Rev Share” from Oct. 2019 to Sept. 2020); Tr. 2488:9–2492:21 (Cue (Apple)). 936. Google has paid billions of dollars in revenue share to Android RSA partners over the years. Tr. 7667:12–15 (Pichai (Google)); Tr. 5727:5–5728:11 (Whinston (Pls. Expert)) (explaining UPXD104 at 19, Google has paid its RSA partners, including Android partners, “billions and billions and billions of dollars.”). In 2020 alone, Google’s payments to carriers and Android OEMs for U.S. searches totaled more than $1.5 billion. Id. 5727:5–5728:11 (referring to UPXD104 at 19). Google also gives away the “must-have” Play Store on Android at no monetary cost to Android OEMs. Tr. 9426:17–18 (Rosenberg (Google)) (There is no license fee for the MADA.); Tr. 5727:5–5728:11 (Whinston (Pls. Expert)). 937. Google’s revenue-share payments to Mozilla make up approximately 80% of Mozilla’s revenue. Des. Tr. 41:18–24 (Baker (Mozilla) Dep.). In 2021, Google’s revenue-share payments to Mozilla exceeded $[redacted] million worldwide. Tr. 538:1–15 (Rangel (Pls. Expert)) (referring to UPXD101 at 10); Des. Tr. 190:12–17, 190:21–191:2, 191:5–10 (Baker (Mozilla) Dep.) (Google’s revenue share to Mozilla “ends up in the hundreds of millions of dollars.”). 938. Google’s payments to third-party browser partners—excluding Apple—totaled more than $150 million in 2020 for U.S. default search traffic. Tr. 5727:5–5728:11 (Whinston (Pls. Expert)) (referring to redacted figures in UPXD104 at 19). 939. The massive size of Google’s default revenue share payments confirms the value of these defaults; Google is a sophisticated, successful, for-profit corporation that would ensure these payments are worthwhile. Tr. 5727:5–5729:5 (Whinston (Pls. Expert)) (“[W]hen you see Google paying billions and billions and billions of dollars, there’s got to be a reason. There has to be a reason it’s worth doing it.”); Tr. 538:1–15 (Rangel (Pls. Expert)). Google’s willingness to invest billions of dollars to secure defaults shows that the default bias in search is “of sufficient magnitude” for “participants in the market [to] care about” and invest billions of dollars to influence them. Tr. 524:15–25, 538:1–15 (Rangel (Pls. Expert)). 940. Consistent with his fiduciary obligation to shareholders, Mr. Pichai would not authorize contracts to pay billions of dollars every year for search defaults if they were not valuable. Tr. 7669:15–25, 7773:13–18 (Pichai (Google)); Des. Tr. 29:19–30:1 (Porat (Google) Dep.) (agreeing that “the value of the acquired traffic that Google gets from the TAC payments exceeds . . . the cost of the payments that Google makes” and that “[o]therwise, [Google] wouldn’t make the payments”); Tr. 3796:5–3798:22 (Ramaswamy (Neeva)) (“[S]earch is one of the most profitable businesses ever. And so defaults in search, therefore, have a very meaningful impact on . . . Google’s top and bottom line.”). 941. At a [redacted]% revenue recovery rate, supra ¶ 924, Google stood to lose more than $7 billion in 2020 and more than $28 billion between 2016 and 2020 if Apple changed the Safari default. UPX1050 at -887; Tr. 1629:19–1631:18 (Roszak (Google)) (discussing UPX1050 at -887). 942. The 36% gross revenue share that Google agreed to pay Apple in the 2016 amendment to the ISA, supra ¶ 226, offers an economic perspective of the magnitude of the default effect in search. Tr. 5728:12–5729:5 (Whinston (Pls. Expert)). For Google’s revenueshare payments to make economic sense, the revenue Google expects it would lose if a rival won the default must exceed the traffic acquisition costs of obtaining the default. Id. As a result, Google must expect to lose at least 36% of Safari default traffic revenue if Apple were to switch the Safari default. Id. (discussing UPXD104 at 20); Des. Tr. 29:19–30:1 (Porat (Google) Dep.) (agreeing that “the value of the acquired traffic that Google gets from the TAC payments exceeds . . . the cost of the payments that Google makes” and that “[o]therwise, [Google] wouldn’t make the payments”); id. 32:15–32:22 (agreeing that “the purpose of [Google’s] payment [to Apple] is to get the additional or marginal traffic that [Google doesn’t] get today”); Tr. 1540:5–9 (Roszak (Google)) (defining incremental revenue). 943. Ultimately, because defaults are the most potent method of distribution, Mr. Nadella was prepared for Microsoft to lose “billions of dollars” to secure the Apple’s search defaults. Tr. 3502:21–3505:23 (Nadella (Microsoft)) (“[Microsoft was] always grounded in the fact that if [Apple] needed to switch, that [Microsoft] would have to pay, and [Microsoft] would have to pay and even subsidize the transfer . . . we were going to be negative for multiple years perhaps, and I was willing to do essentially a very long term and think perpetually.”); Des. Tr. 111:20–112:20 (van der Kooi (Microsoft) Dep.) (Microsoft’s former CEO Steve Ballmer “was deeply committing to investing in the search business beyond . . . what was, let’s say, purely economically justifiable. . . . There is not really another company in the world that you can imagine that would have had the wherewithal and the staying power to incur that degree of losses over that period of time other than Microsoft.”). Continue Reading Here. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on April 30, 2024, storage.courtlistener is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. [15] Behavioral economics is a branch of economics that applies concepts from psychology and neuroscience to improve understanding of human decision-making—including decisions made by consumers. Tr. 516:3–8 (Rangel (Pls. Expert)); UPX0101 at -280 (“Behavioral Science: A method of predictive analysis that applies economic modeling and psychological insight to the study of how people make decisions and develop attitudes and beliefs.”). For almost a decade, Google has relied on collaboration with Irrational Labs, a behavioral economics consulting firm, and its own internal behavioral science team led by Maya Shankar, the former head of behavioral economics in President Barack Obama’s administration. Tr. 531:22–532:17 (Rangel (Pls. Expert)); e.g., UPX1102 (2015 Irrational Lab presentation to Google); UPX0103 (2021 presentation from Google’s internal Behavioral Economics team). United States of America v. Google LLC., Court Filing, retrieved on April 30, 2024, is part of HackerNoon’s Legal PDF Series. You can jump to any part of this filing here. This part is 22 of 37. United States of America v. Google LLC., Court Filing, retrieved on April 30, 2024, is part of HackerNoon’s Legal PDF Series . You can jump to any part of this filing here . This part is 22 of 37. HackerNoon’s Legal PDF Series here D. Defaults Have A Powerful Effect On Users’ Search Behavior, Particularly On Mobile Devices 867. Defaults are powerful in virtually all contexts. Tr. 526:22–25 (Rangel (Pls. Expert)) (discussing consensus in the field of behavioral economics about default effects). Search is no different. In search, this is demonstrated by the massive sums Google pays each year to ensure it maintains its search defaults status. Infra ¶¶ 932–943 (§ VI.D.6). In fact, Google spends more money securing exclusive defaults than on all other search-related expenses combined. Tr. 7576:9–7577:6 (Raghavan (Google)) (discussing UPX7002.A at 1). Many general-search queries are controlled by the power of defaults, and that is especially true on mobile devices. 1. The “Power Of Defaults” 868. A default is an option pre-selected for a consumer by a third-party, such as a smartphone manufacturer, that requires an affirmative action by a consumer to change. Tr. 520:7–25 (Rangel (Pls. Expert)) (defining defaults from the perspective of behavioral economics); Tr. 154:3–7 (Varian (Google)) (describing a default as a “capability that’s available before the user takes any action”). A default search engine is the search engine that comes preselected on a search access point when a user first begins using the access point. Des. Tr. 68:19– 25 (Ramalingam (Yahoo) Dep.) (the default is the one set “[w]ithout any user input”); Tr. 322:19–323:4 (Barton (Google)) (the default search is “the one that comes out of the box” on a browser, desktop, or mobile phone). 869. Across a range of domains, defaults have a powerful impact on consumer decisions. Tr. 520: 7–25, 526:22–529:16 (Rangel (Pls. Expert)) (discussing impact of defaults on consumer decision-making and various real-world default biases including coffee-shop tipping, 401(k) plans, and organ donation). 870. In its own businesses, Google has long understood and embraced the “power of defaults,” a term first evangelized by Dr. Varian in 2007. UPX1001 at -465 (referencing Dr. Varian’s “‘power of defaults’ idea”). As one Google employee explained, “the default options presented (in anything from finance to gaming) are very powerful, and will probably end up being what most people choose (out of lack of knowledge about customization, or convenience).” UPX0081 at -438. 871. In 2018, Google’s internal behavioral economics[15] team used the power of defaults to manipulate ad spending by creating a $10 default for advertisers’ maximum daily budget. Tr. 532:22–535:12 (Rangel (Pls. Expert)) (discussing UPX0101 at -285–86). Despite being visible to advertisers and easy to change, the new default budget had a powerful impact on Google’s advertisers as 64% of “low-budget” spenders (advertisers who previously set their daily maximum to less than $10 per day) increased their daily budget. Id. 532:22–535:12 (discussing UPX0101 at -285–86). However, Google also found the default “too powerful” because it dragged down spending by advertisers who had been setting their daily maximums above $10. Id. 872. Google also recognizes the power of the few defaults that it does not own but instead calls these the “tyranny of default.” UPX0111 at -825 (June 20, 2017 memo on “Google Branding on Search Desktop”); UPX0099 at -390 (Dec. 12, 2016 Email from Ben Friedenson (Google) analyzing Microsoft Edge default settings and observing either a) ‘tyranny of the default’ or b) a surprisingly high preference for/loyalty to Bing” among Edge users). In 2005, Google complained to Microsoft about its decision to preset a default search engine in the Internet Explorer browser. Tr. 7677:5–7681:6 (Pichai (Google)) (contextualizing UPX0172). Google wrote, “As you know, most end users do not change defaults,” and only “a tiny fraction of end users . . . try to change the default.” UPX0172 at -731; supra ¶ 742. Google accused Microsoft of “put[ting] its own interests above those of end users.” UPX0172 at -731. Because users do not change defaults, Google was “deeply concerned” that Microsoft’s decision to preset its own search engine as the default in Internet Explorer could harm the competitive process. Tr. 7693:5–8 (Pichai (Google)) (confirming concerns written in UPX0172). 873. Google proposed that Microsoft implement, instead of a preset default, a choice screen that would allow the end user to select their default engine the first time he or she used the Internet Explorer search box. UPX0172 at -731. A choice screen would place “control in the hands of the end user, where it belongs.” Id. Giving the user choice of search engine and a straightforward means of changing it was “indispensable both to preserve competition on the merits in search and to avoid distortion of competition in related markets.” Id. at -732; Tr. 7685:9–15, 7686:1–6 (Pichai (Google)) (Google proposed a choice screen to benefit users by aligning with user preference). 874. Other industry participants recognize the power of defaults on consumer behavior as well. Considering a request from Google that Apple share more user-location data, Apple executives said having “data sharing” set to “off” by default “wouldn’t have any real effect since so few people change defaults.” UPX0094 at -029. Apple recently demonstrated the power of defaults in a public way; the company changed from a default to a choice screen for cross-app tracking on iPhones, which resulted in 80% of users forbidding cross-app tracking even though they had previously acceded to tracking by default. Tr. 635:15–637:8 (Rangel (Pls. Expert)) (discussing UPXD101 at 55). 2. Search Engine Defaults Are The Most Efficient Method Of Distribution 875. Defaults can generate a bias in favor of the default option even when the default is easy to change, when the stakes of the decision are high, such as with end-of-life care, and when the person making the decision is an expert, such as in medication prescriptions. Tr. 529:17– 531:21, 532:18–535:12 (Rangel (Pls. Expert)). 876. By comparison, using a GSE is a low-stakes and low-expertise activity, so default bias is especially strong. Id. 530:10–531:21, 532:18–535:12. Default effects in search are “sizeable and robust.” Id. 523:14–24, 541:2–22; Tr. 5749:17–5750:22 (Whinston (Pls. Expert)) (“[T]he power of defaults is very significant.”). Even Google’s expert, Prof. Murphy, agreed that empirical evidence suggests that having a default generates “additional search volume.” Tr. 9941:9–11 (Murphy (Def. Expert)). 877. Dr. Ramaswamy testified that “being the default on the browser is the most efficient way to get [a search engine] into the hands of your users” because of the “convenience of easy accessibility and” tapping into . . . engrained default behaviors are deciding factors when it comes to whether a search engine gets lots of usage.” Tr. 3689:15–24 (Ramaswamy (Neeva)); Tr. 1961:20–1962:5 (Weinberg (DuckDuckGo)) (being a search access point’s default search engine is the most successful method of distributing a search engine). 878. In a 2007 study, Google examined the effect of browser home page defaults and query share. UPX0123; UPX0093 at -904. In the study, Google found that, when considering the factors that might influence a user’s choice of search engines—such as results quality, search features, user experience, and brand strength—“one factor surprisingly trumps them all: the default home page setting.” UPX0093 at -904 (Abstract for 2007 Google study “On the Strategic Value of the Default Home Page to Google”); UPX0124 at -036 (Aug. 10, 2007 presentation for Mr. Pichai (Google) and Ms. Braddi (Google) showing the “strong relationship between home page and search share”). Google further found that defaults were “directly correlated” with query share. UPX0093 at -904. Google showed that “users who have home page set to Google do 50% more searches on Google compared to those who don’t.” UPX0093 at -904. At the time, Dr. Varian concluded that the study was “great stuff” and “a very convincing case.” UPX1001 at -465. 879. Dr. Varian also agreed with (and liked) the study’s conclusion that a default home page—a form of default distribution popular at the time—“can be a powerful strategic weapon in the Search battle.” Tr. 170:15–24 (Varian (Google)) (reviewing UPX0123 at -485); UPX0123 at -487; UPX2049 (Executive Management Group (EMG) representatives “agreed that [it] is important” to “focus[] on homepage market share” because it is “one of the most effective things we can do to make gains in search market share”). In response to the study’s findings, Google began “taking these defaults very, very seriously.” Tr. 3710:15–3712:20, 3714:12–20 (Ramaswamy (Neeva)) (discussing UPX0093); UPX0093 at -904 (email from Dr. Ramaswamy, while working at Google: “This study is very cool! We should definitely put some marketing push behind it”). 880. In contrast to defaults, app downloads and marketing are ineffective methods of distribution; the availability of search apps in mobile application stores—such as the Play Store or Apple’s App Store—do not erode the effectiveness of default distribution. Tr. 617:23–618:17 (Rangel (Pls. Expert)). A Google survey found that more than a third of iPhone users did not know that they could download the Google Search App and that the majority of those who were aware of the ability to install the app still did not see the need to depart from Safari. UPX0139 at -149. Mozilla has found it “exceedingly difficult” to convince mobile users to download its browser application because “people have to find and install and you have to struggle if you can make them your default.” Des. Tr. 135:3–5, 135:8–22 (Baker (Mozilla) Dep.). “[M]erely having an app in the app store is a very difficult way to compete with the preloaded defaults... [b]ecause each person who gets that phone has to make a conscious decision to go through a lot of work to get to your product." Id. 139:20–23, 140:1–24. Apple also views searching on an application, such as the Google Seach App, as inferior to searching on a browser and that a search app therefore "wasn't the best thing for customers." Tr. 2493:18-2495:23, 2496:6-10 (Cue (Apple)). Google's analysis about search on Apple devices has similarly found that more than half of iPhone users did not have the Google Search App installed and even about half of the users who did still preferred using Safari. UPX0139 at -150. 881. Paid marketing for GSEs, browsers, and search apps is also an ineffective distribution method. UPX0122 at -960 (illustrating various distribution methods for Google services and calling pre-installation the "highest value way to acquire users"). In 2016, Emily Moxley, a Google vice president, “guarantee[d]” that marketing Google Search in Windows 10 would "make no difference" in generating usage. PSX00216 at -126 (“Didn't work for safari, won't work here!"). A January 2015 Google "Post-Mortem" on Firefox's 2014 default switch to Yahoo found that marketing efforts were immaterial. UPX0145 at -441, -450 (“Early results indicate that Chrome/DSE/DHP promos have little/no effect."); UPX0169 at -180 (email to Sundar Pichai (Google) stating "user promos [on Firefox] haven't materially returned users to Google"). DuckDuckGo has found marketing to be unhelpful in generating traffic because teaching people to change the default search engines is difficult. Tr. 1957:15–1958:5 (Weinberg (DuckDuckGo)). 882. Ultimately, a default search engine garners “enormous usage simply by the power of the default.” Tr. 3710:7-3712:20 (Ramaswamy (Neeva)) (“[W]hoever controls that search box gets a lot of usage independent of the merits of the search engine. And so you get enormous usage simply by the power of the default.”). According to Microsoft CEO Satya Nadella: “The entire notion that users have choice and they go from one website to one website or one search into one search and it’s complete bogus. There’s defaults. The only thing that matter in terms of changing search behavior.” Tr. 3497:13–3498:4 (Nadella (Microsoft)). “Whoever owns [the] default will gain share no matter what the quality.” Tr. 2722:2–10 (Parakhin (Microsoft)) (“It is, I believe well-documented fact that people very rarely switch defaults.”). 883. Indeed, most search traffic arrives through defaults rather than through nondefault querying. UPX0083 at -967 (61% of Google’s queries in 2017 arrived at Google via a paid distribution deal, which was a steady increase from 48% in 2014, 52% in 2015, and 57% in 2016.); UPX1050 at -894 (65% of Google’s Apple Traffic arrives through “Safari Default Search” compared to just 13% through “Safari Organic Homepage”); Tr. 23:25–25:2 (Sept. 19, 2023 sealed PM session (Yoo (Google)) (Android “organic” queries have declined over time.); id. 25:16–27:3 (less than 10% of Android revenue comes from “organic” queries); Tr. 3102:11– 3104:25 (Tinter (Microsoft)) (discussing greater use of search defaults than non-default querying); UPX0080 at -509 (Email from Ms. Braddi (Google) stating that Google generated almost three times as much revenue on iOS from default traffic as “organic” revenue). 3. Behavioral Economics Explains The Power Of Defaults 884. The behavioral economics concepts of habit, choice friction, and user confusion and awareness explain the preeminence of default distribution. Tr. 541:2–22, 542:25–543:18, 547:16–549:3 (Rangel (Pls. Expert)). a) Habit a) Habit 885. Habit is when consumers make decisions automatically based on prior outcomes rather than expending cognitive capacity to make an explicit choice. Tr. 541:23–542:24 (Rangel (Pls. Expert)) (consumers use habit to preserve limited cognitive capacity by, for example, requesting the same order at a coffee shop rather than considering every possible option on each visit). Because search is an activity carried out repetitively, on a familiar interface, and with immediate feedback in the form of a search result, search “[h]abits develop very strongly.” Id. 542:25–543:9. 886. Google’s users have identified “habit”—along with convenience and ease of access—as the reason why they continue to use the “Google Safari Default.” UPX0090 at -940– 41 (2010 Google iPhone user study respondent: “I use Google out of habit. There is no difference between the way the results are presented from other engines.”). 887. Google itself recognizes the impact of habit on consumer behavior in search. To address the threat of competition from Microsoft’s search product in 2003, Dr. Varian recommended that Google “[g]et users addicted to our interface and tools.” UPX0151 at -158– 59; Tr. 140:22–142:9, 144:13–19 (Varian (Google)) (discussing context of UPX0151); Tr. 545:12–546:8 (Rangel (Pls. Expert)) (discussing UPX0151 at -158–59 as reflecting what behavioral economists would consider “an extreme habit”). The 2007 Google study examining the effect of browser home page defaults and query share, supra ¶ 878, attributed increased searching when Google was the default homepage to the fact that “[u]sers do not always make a deliberate choice of search engine.” UPX0123 at -485 (“Most users stay with pre-configured home page settings”); Tr. 551:10–552:13 (Rangel (Pls. Expert)) (habit is the opposite of a consumers’ deliberate choice). And Google pursued default exclusivity in its search distribution contracts precisely because it afforded Google “an opportunity to be discovered and used on a repeating basis.” Tr. 338:13–17 (Barton (Google)). 888. Mr. Nadella explained that the Google search habit has become ingrained like many other common habits; he testified, “[Y]ou get up in the morning, you brush your teeth and you search on Google. And so therefore, with that such level of habit forming, the only way to change is by changing defaults.” Tr. 3497:13–3498:4 (Nadella (Microsoft)). Dr. Ramaswamy similarly asserted that “the convenience of easy accessibility and tapping into . . . engrained behaviors are the deciding factors when it comes to whether a search engine gets a lot of usage. . . . [D]efaults become habits.” Tr. 3689:15–3690:4 (Ramaswamy (Neeva)). 889. Habit is so profound in search that generative AI is unlikely to meaningfully impact Google’s search usage share. Tr. 3512:8–3513:9 (Nadella (Microsoft)) (Generative AI “doesn’t address all the other challenges . . . which is the user habit [that] comes from search defaults, settings, so the distribution advantage Google has today doesn’t go away.”). Dr. Raghavan agreed that ChatGPT “hasn’t put a dent in Google’s market share.” Tr. 7533:6– 7534:19 (Raghavan (Google)). 890. As a result of controlling the defaults on search access points, habit creates bias in Google’s favor. Tr. 541:23–542:24 (Rangel (Pls. Expert)). b) Choice Friction b) Choice Friction 891. On the infrequent occasions when a consumer makes an explicit choice of search engine, choice friction causes Google’s default to remain sticky. Tr. 541:2–22 (Rangel (Pls. Expert)) (changing only the choice architecture for implementing a consumer’s decision, such as the number of steps to execute the decision or the difficulty of executing those steps, can powerfully affect consumer decision-making); id. 551:10–552:13 (in search, users act in “explicit or deliberative” decision-making mode only “infrequently”). Increasing choice friction increases the stickiness of a particular default. Id. 554:4–21 (“[T]he more choice friction it takes to change the defaults, the stickier the defaults are.”). 892. Google’s internal behavioral economics team has taught the company that even “[s]eemingly small friction points in user experiences can have a dramatically disproportionate effect” on consumers’ decisions. UPX0103 at -214; UPX0848 at -612 (“each barrier in terms of time attention, and demand for thinking is 10X what you think it is”). A June 2019 presentation about private searching on Google, prepared for Google’s product counsel, observed that Google was failing to “fully address user needs” because “multiple toggles and settings to get into a more ‘private state’ outside of Incognito . . . create user friction and are difficult to understand.” UPX0811 at -413. 893. In search, Dr. Rangel found that bias towards Google’s defaults arises from at least four sources of choice friction faced by users changing the pre-set default search engine. First, a user must become aware that there is a default search engine and that it can be changed. Tr. 552:15–553:25 (Rangel (Pls. Expert)); Tr. 2636:3–9 (Cue (Apple)) (before changing the default search engine on an iPhone, a user must know that a search engine is pre-selected as the default). Second, the user must discover what alternative options to the default search engine exist. Tr. 552:15–553:25 (Rangel (Pls. Expert)). Third, the consumer must learn the steps necessary to change the default. Id.; Tr. 2636:10–2637:8 (Cue (Apple)) (an iPhone user must understand how to change the default search engine on iPhone, which may require going to Apple, Google, or Bing’s support websites to learn the steps). Fourth, the consumer must actually take the steps necessary to implement the change. Tr. 552:15–553:25 (Rangel (Pls. Expert)). 894. On Apple’s iPhone, the steps to change the default search engine require overcoming “substantial” choice friction. Tr. 562:21–563:8 (Rangel (Pls. Expert)); Tr. 2169:23– 2170:20 (Giannandrea (Apple)) (reviewing the four steps to change the default search engine on iPhone in UPXD004); Tr. 2630:7-16 (Cue (Apple)) (describing four steps to change the default search engine on iPhone); Des. Tr. 55:20–56:14 (Ribas (Microsoft) Dep.) (“even though, in theory, it's possible to [change the Safari default on iPhones], it's not" and even "some [Microsoft] corporate vice presidents” and “smart engineers" have asked how to make a change). 895. An Android user who has already overcome the choice frictions of understanding that there is a default, discovering the alternatives, and learning the steps to change the Google default search widget must still execute ten steps to implement the change, which constitutes "considerable choice friction" for the user. Tr. 559:14–562:20 (Rangel (Pls. Expert)) (reviewing UPXD101 at 21–38); Tr. 979:19–980:2 (Kolotouros (Google)) (explaining that DX0738 at 2-3 shows that replacing a search widget requires one swipe, two touch-and-holds, three taps, and a selection); Tr. 2722:11-13 (Parakhin (Microsoft)) (requiring multiple clicks to change a search engine default makes doing so proportionally more difficult). 896. It also takes several clicks to change the default search engine on Mozilla's Firefox browser. Des. Tr. 187:18-20, 187:23 (Baker (Mozilla) Dep.). Changing the default search engine in other browsers is often even more difficult. Id. 49:16–23 (competing browsers' preset default search engines are “often not nearly as easy to change the default as in Firefox”); Tr. 1958:15–1961:8 (Weinberg (DuckDuckGo)) (Getting people to switch defaults is difficult because "it's all just way harder than it needs to be."). 897. Consequently, changing the default search engine requires “mental effort” and is not a "zero cost" action. Tr. 563:9-25 (Rangel (Pls. Expert)); Tr. 3491:23-3493:8 (Nadella (Microsoft)) ("changing of defaults is hard"); Tr. 3796:5–3798:22 (Ramaswamy (Neeva)) (dismissing the “pious prose around 'competition being a click away""). In contrast, controlling defaults boosts Google by making it “very, very seamless and easy for users” to access Google. Tr. 7661:1–11 (Pichai (Google)) (“[W]e know [making Google the default on Safari] would lead to increased usage of our products and services, particularly Google Search in this case. So there’s clear value in that, and that’s what we’re looking to do.”); Tr. 1537:9–12 (Roszak (Google)) (a benefit of Google’s default distribution deals is reducing choice friction). 898. Google has long recognized and even harnessed the effect of choice friction to protect its defaults. Tr. 554:17–558:19 (Rangel (Pls. Expert)) (discussing representative examples of Google recognizing and harnessing choice friction). For as long as the company has required it, Google’s default exclusivity has ensured that a user cannot change the default search engine to a rival without going into the device’s settings. Tr. 324:14–325:11 (Barton (Google)). This makes changing default search engines “difficult.” Id. 338:18–24 (explaining that users would “have a difficult time finding or changing [the default search engine] to Google” were it not pre-set as the default search engine). 899. Google’s MADAs with OEMs ingrain the importance of choice friction. Tr. 556:12–557:9 (Rangel (Pls. Expert)). In January 2020, Google amended the MADA to restrict OEMs from making it easier for a user to change a default by prohibiting OEMs from including “processes, instructions, promotions, or other means that directs, instructs or encourages the End User to change” their default search engine away from Google. Id.; UPX5358 at -004 (§ 2.5) (LG MADA (2020 amend.); JX0099 at -998 (§ 2.9) (Motorola MADA (2020 amend.)); UPX5511 at -987–88 (§ 2.9) (Samsung MADA (2020 amend.)); Tr. 799:14– 801:21 (Kolotouros (Google)) (agreeing that MADA’s amendments prohibited OEMs from instructing or encouraging a user to change default settings or placements). 900. Google’s new prohibition against instructing or even encouraging a user to make changes introduced choice friction by impairing users’ awareness of the ability to change a default. Tr. 556:12-557:9 (Rangel (Pls. Expert)) (analyzing new prohibitions in the context of a Samsung MADA amendment). 901. google has insisted on more default search friction, even when the partner preferred less. In 2018, Samsung changed the interface in its S browser to add a drop-down menu in the address bar that would allow a consumer to easily and permanently change their default search engine. Supra¶¶ 296, 808; Tr. 856:2-857:17 (Kolotouros (google)) (Samsung’s S Browser included a dropdown in the address bar that this changed the S Browser default search engine). google informed Samsung that this change breached the RSA and considered terminating Samsung’s RSA. Id. 857:14-858:1; Tr. 557:18 (Rangel (Pls. Expert)); UPX1011 at -289 (Google considered terminating Samsung’s RSA due to a potential fourth strike). Samsung acceded to Google’s demands and threats, and users who might wish to change the search default must take additional steps to do so. Tr. 997:1-6, 887:24-888:13,890:3-9 (Kolotouros Google)). c) User Confusion and Unawareness c) User Confusion and Unawareness 902. Default bias in favor of Google is bolstered further by user confusion and unawareness about what a default search engine is or even that it can be changed. Tr. 547:16-549:3 (Rangel ({ls. Expert)) (“[M]any [users] are confused. They don’t know that there is a default. They don’t know what the default search engine is. They don’t know they can be changed.”); id. 552:15-553:25 (“[M]any people are not aware [that there is a default search engine], and people are confused, even about what’s the default search engine that they have.“). 903. Many internet users do not pay attention to their search engine and instead simply use the default search engine in their browser. Tr. 3685:15--3686:19 (Ramaswamy (Neeva)) (testifying about UPX0940 at -492, which found that almost half of surveyed users search using the “Browser’s default“ GSE). For example, one 2020 Google study found that more than 80% of Google’s daily active users on Apple’s iOS access Google through Apple’s pre-set default browser, Safari, and were “often unaware they’re using Google.” UPX2051 at -519–20 (user unawareness of Google use in Safari “a risk for [Google]”). Google’s expert Prof. Murphy agreed that some users are probably unaware that they even have alternatives for their default search engine. Tr. 9942:7–10 (Murphy (Def. Expert)). 904. Often, even tech-savvy consumers do not know the difference between a browser and a search engine. Tr. 3685:15–3686:19 (Ramaswamy (Neeva)) (user feedback revealed that “many perfectly tech-savvy people could not tell the difference between a browser and a search engine”); Tr. 702:5–23 (Rangel (Pls. Expert)) (users “don’t understand the difference between searching and browser”). Google interviewed Microsoft users in 2017 and found they were confused about the difference between browsers, search engines, and even email. UPX0066 at -073 (recording feedback from one Bing user who said, “Most of the time I use Google, everybody knows that”); PSX00216 at -126 (“one fundamental issue [with marketing Google in Windows 10] is that users on Edge don’t even realize they aren’t using Google, and they don’t feel they are missing out”). 4. Market Evidence Confirms Default Effects In Search 905. Because defaults “have a strong impact” with a “75% take rate,” in 2007, Google leadership rejected Apple’s request to amend the ISA to allow Apple to implement a search engine choice screen in Apple’s forthcoming version of the Safari browser for Windows. UPX0137 at -688–89; Tr. 7691:19–7692:13 (Pichai (Google)); Tr. 4980:13–4983:8 (Braddi (Google)) (describing meeting). In 2010, Google found that it “will lose more than 50% of Google users (and more non-Google users) if the [Safari mobile] default search engine changes to something else.” UPX0070 at -657. 906. In 2012, Apple replaced Google Maps with Apple Maps as the default maps application on iOS, which created an immediate, sizeable, and lasting change in usage. Tr. 564:11–566:1 (Rangel (Pls. Expert)) (discussing reproduction of chart in UPX0097 at -.001, showing a large and sudden drop in Google Maps usage on iOS when Apple changed the default while, during the same period, usage of Google Maps on Android increased); Tr. 1558:3–5, 1560:4–20, 1563:25–1566:4 (Roszak (Google)) (Google Maps lost 60% of its usage vis-à-vis its peak usage—when it was the default maps application on iOS). 907. In 2012, Apple publicly apologized for Apple Maps’ poor performance. Tr. 565:11–566:1 (Rangel (Pls. Expert)); Tr. 1569:15–18 (Roszak (Google)). As Prof. Rangel explained, the Apple Maps event illustrates that defaults are so powerful that they can generate bias among users “even when the default is assigned to an inferior product.” Tr. 565:11–566:1 (Rangel (Pls. Expert)). According to Mr. Nadella, the 2012 Apple Maps event demonstrated “the power of defaults” and gave Microsoft confidence that winning the iOS default would be a “big game-changer” in search competition. Tr. 3500:9–3502:2 (Nadella (Microsoft)) (“[Apple Maps is] exhibit number one on anybody who wants to sort of know anything about sort of the power of defaults, that case study is the best.”). 908. Mozilla’s 2014 decision to switch the default search engine on its Firefox browser from Google to Yahoo sharply and immediately shifted approximately 20% of queries from Google to Yahoo—a phenomenon that was reversed in 2017 when Mozilla reverted the Firefox default to Google. Tr. 628:6–20, 630:8–631:14, 660:12–661:18 (Rangel (Pls. Expert)) (describing Firefox default change and referring to redacted query share shifts in UPXD101 at 55); Tr. 5729:25–5731:4 (Whinston (Pls. Expert)) (discussing redacted query share shifts in UPXD104 at 23 and explaining, based on “jump[s]” in query shares for Yahoo and Google in 2014 and 2017, respectively, that “[i]f defaults didn’t matter, all of this would have been a nice, smooth—maybe not line, but curve. You wouldn’t have seen these jumps”); Des. Tr. 62:9–18 (Baker (Mozilla) Dep.) (describing Firefox default switch in 2014). Mozilla’s default switch cost Google an estimated half-a-billion dollars in 2015 along with a 30% decline in traffic from Firefox and a 45% decline in revenues from Firefox. Tr. 1610:2–1611:14 (Roszak (Google)) (discussing UPX0066 at -071). 909. As significant as these Firefox default effects were, they likely understate the power of the defaults controlled by the challenged agreements. First, because Firefox is not a pre-set default browser on any device in the United States, Firefox users are generally more tech savvy and therefore more likely than the average user to understand defaults and make a change. Tr. 628:6–20 (Rangel (Pls. Expert)). Second, Firefox is used almost exclusively on desktop, where defaults are less sticky than on mobile. Tr. 5731:6–21 (Whinston (Pls. Expert)) (lower switching is the result of Firefox being used primarily on desktop rather than mobile, which is more difficult to switch); Tr. 628:22–629:5 (Rangel (Pls. Expert)) (vast majority of Firefox queries come from desktops); infra ¶¶ 918–923. 910. Choice screens implemented by Google on Android devices in Russia and Europe illustrate the default effect by showing the consequence of removing a default. Tr. 610:6–611:18 (Rangel (Pls. Expert)). In 2017, Russian competition authorities required Google to implement a choice screen on Android devices, which resulted in a consistent and continuing loss in Google’s Android query share that was picked up by the Russian search engine Yandex. Tr. 2673:12– 2674:25 (Parakhin (Microsoft)) (describing Google’s settlement with Russian regulators and explaining that Yandex was not able to maintain market share on mobile before 2017 because “very few users change defaults”); Tr. 608:14–609:9, 609:15–610:9, 610:19–612:1 (Rangel (Pls. Expert)) (discussing query share shifts with reference to redacted figures in UPXD101 at 44); Tr. 10001:23–10002:10 (Murphy (Def. Expert)) (agreeing that the choice screen contributed to Google’s share decline in Russia); UPX0170 at -978 (depicting an approximate 19% search share loss for Google between 2017 and 2019 that resulted in Yandex overtaking Google as the leading search engine in Russia). Moreover, on Russian iPhones and desktops, where no choice screen was implemented and Google’s default persisted, query shares held constant, demonstrating that the choice screen was in fact responsible for the Android share changes. Tr. 612:4–613:5 (Rangel (Pls. Expert)) (referring to UPXD101 at 45); Tr. 5744:7–20 (Whinston (Pls. Expert)) (referring to UPXD104 at 27). Google agreed that the choice screen was a “key” cause for Google’s loss in Russian query share. UPX0170 at -979; UPX0760 at -596 (“[The Russian remedy] implementation (2016/17) predictably led to a decline Google’s queries share on mobile.”). 911. Similarly, the European choice screen implemented in 2021 following Google’s Android settlement with the European Commission shifted query shares and generated competition that had not existed when Google was the default across Android devices. UPX0764 at -077 (Quality-enhancing efforts undertaken by Google as part of the “Go Big in Europe” initiative in response to the choice screen included improving unique features, local content, and trust). Tr. 620:22–624:4 (Rangel (Pls. Expert)) (referring to UPXD101 at 49); Tr. 5735:1– 5737:19 (Whinston (Pls. Expert)) (referring to UPXD104 at 25) (Google’s selection on the choice screen was lower than its market share before adoption of the choice screen). 912. A regression analysis using data from the European choice screen predicts that Google’s U.S. rivals would acquire about 10% market share were a choice screen implemented in the United States. Tr. 5737:21–5738:10 (Whinston (Pls. Expert)). And although Google retained a high share in Europe following the choice screen, the “default echo” deriving from Google’s incumbency and brand recognition makes this outcome unsurprising and was, in fact, anticipated by Google. Tr. 622:25–625:14 (Rangel (Pls. Expert)); Tr. 5732:15–5734:18, 5735:1– 5738:22 (Whinston (Pls. Expert)) (referring to UPXD104 at 25 and explaining that the European choice screen showed consumer preferences at the time the choice screen was instituted and that there would be greater share moving away from Google if rivals were stronger); UPX1103 at -775 (Google email discussing the importance of brand recognition in differing outcomes with European and Russian choice screen experiments); supra ¶ 544 (discussing effect of brand recognition in Europe). 913. Finally, Google’s query share differentials between browsers where Google is the default—such as Chrome, Firefox, and Safari—compared to Microsoft’s Internet Explorer and Edge browsers, where Google is not the default, show that defaults drive search engine query shares. Tr. 5744:21–5745:20 (Whinston (Pls. Expert)) (referring to UPXD104 at 28–29); UPX0955 at -774 (2020 Email from Brave executive reporting that setting DuckDuckGo as the default search engine in Brave’s browser in Germany, Ireland, Australia, and New Zealand showed the “data is very clear: default is sticky and very powerful.”). In the context of desktop, Bing’s query share on Windows computers, where it is the default search engine on the only preinstalled browser, compared to Bing’s query share on Apple’s Mac computers, where Google is the default search engine on the only preinstalled browser, shows a near seven-fold default effect. Tr. 631:23–632:17 (Rangel (Pls. Expert)) (referring to UPXD101 at 56). 5. Ordinary-Course Modelling For Multi-Billion Dollar Decisions Confirms That Search Defaults Have High User Retention Rates, Especially On Mobile 914. In 2016, Apple and Google renegotiated the ISA, a multi-billion dollar agreement. Tr. 1630:10–14 (Roszak (Google)) (agreeing that the Safari default’s worth to Google was “an important question” that had “billion of dollars on the line“); infra¶935 (overviewing size of Google payments to Apple). 915. For the negotiations, Google's finance team modeled the impact of losing Apple's Safari default, which Google considered a "Code Red" scenario. Tr. 1616:20-1618:15 (Roszak (Google)); Tr. 536:10–537:21 (Rangel (Pls. Expert)) (discussing UPX0171 at -169, -177); UPX0171 at -169, -177 (acknowledging that Google was “vulnerable” if Apple switched the default search engine); UPX0085 at -228, -230, -240 (discussing "Code Red" in the context of Project Soy, which was code for clawing back users on Apple's iOS). 916. To model the outcomes of losing the search default on Apple devices, Google applied the revenue shifts from the default losses on Firefox, supra ¶ 908, and Apple Maps, supra ¶906. UPX1050 at -828, -886 (January 2016 modelling relying on Firefox for desktop recovery and on Apple Maps for iPhone recovery); Tr. 1620:13–1631:18 (Roszak (Google)) (reviewing UPX1050 at-828, -886); UPX6024 at -440, -443 (written 30(b)(6) response: Google used Firefox and Apple Maps events "to model certain clawback assumptions" during 2016 ISA negotiations with Apple.). 917. For desktop, Google based its modeling on the outcome of Mozilla's changing the Firefox default from Google to Yahoo in 2014 and concluded that Google would lose [redacted]% of its revenue ([redacted]% recovery) if it lost the Safari default on desktop. UPX1050 at -828, -886; Tr. 1620:13-1621:25, 1626:13–1627:10 (Roszak (Google)) (reviewing UPX1050 at -828, -886). 918. For mobile, Google based its modeling on the outcome of Apple's switching the iOS maps application default from Google Maps to Apple Maps in 2012 and concluded that it would lose [redacted]% of its revenue ([redacted]% recovery) if it lost the Safari default on mobile. UPX1050 at -828, -886; Tr. 1622:1–1624:16, 1626:13–1627:10 (Roszak (Google)). 919. In other words, Google predicted that its revenue loss on mobile (redacted%) would be more than double its revenue loss on desktop (redacted%). UPX1050 at -886; Tr. 1629:14–18 (Roszak (Google)). 920. The disparity between the default effect on mobile and desktop arises from the fact that “[d]efaults have more prominence in mobile due to screen size and UI [user interface].” UPX1050 at -886; Tr. 1628:8–1629:3 (Roszak (Google)). As Google explained in a later analysis, users “are much less likely to change default search engine on mobile.” UPX0139 at -119 (2018 Google presentation discussing the “[s]ignificant risk if [iPhone] default [search] engine is changed”). Mobile and desktop user experiences have “salient difference[s],” including the “very limited real state” on mobile devices. Tr. 6310:23–6311:16 (Nayak (Google)). 921. These form factor differences between mobile and desktop, such as smaller screen size, are an example of choice friction that enhances the power of defaults on mobile devices. Tr. 523:14–24, 625:17–626:6 (Rangel (Pls. Expert)); UPX0084 at -728 (Google 2016 2Q Earnings Diligence – TAC Appendix stating, “User behavior is more heavily influenced by default settings on mobile and tablet”); Tr. 3498:5–19 (Nadella (Microsoft)) (changing defaults is easier on desktop than on mobile because “there are many, many sort of friction points on mobile operating systems”). 922. Mobile users do, in fact, query mostly through defaults. Tr. 1632:3–14 (Roszak (Google)) (65% of Google’s queries from Apple iOS came through the Safari default); Tr. 3102:11–3104:25 (Tinter (Microsoft)) (fewer users reach search engines directly on mobile devices and instead rely on search access points with pre-set defaults); Tr. 3499:21–3500:8 (Nadella (Microsoft)) (although mobile devices have “multiple search access points, the one access point that matters is the search default on the browser”); UPX0083 at -967 (25% of mobile queries are non-default compared to 32% on tablet and 74% on desktop); UPX0115 at -142 (Microsoft estimates that defaults are responsible for 68% of mobile search volume). In 2020, Google found that 81% of all of Google’s daily active users accessed Google through Safari, the browser preloaded on iPhones where Google is the default. Tr. 7512:23–7514:8 (Raghavan (Google)) (discussing UPX2051 at -519). 923. The enhanced power of defaults on mobile is significant because the vast majority of querying today occurs on mobile devices. Tr. 5798:17–5799:5 (Whinston (Pls. Expert)) (“it’s important to remember mobile is where the market is growing” and that desktop queries are “flat and have been for a long time”); Tr. 2279:20–2280:9 (Giannandrea (Apple)) (in the last few years more queries are being done on mobile than on desktop); Tr. 3098:6–3099:3 (Tinter (Microsoft)) (the general search services market has changed from being “predominantly PC to predominantly mobile”). 924. Consequently, weighting between mobile and desktop revenue mix on Apple devices, Google projected a total revenue recovery that skewed heavily toward its mobile prediction— redacted% total recovery (redacted% total loss). UPX1050 at -886; Tr. 1627:11–24 (Roszak (Google)). 925. The predictions derived from the Firefox and Apple Maps events were Google’s “best guess expectation” that were applied in multiple scenarios. Tr. 1622:21–1624:11, 1743:1– 1744:10 (Google (Roszak)). These findings were shared with Google CFO Ruth Porat. Id. 1619:23–1620:12. Mr. Roszak, the leader of the Google finance team assessing default deals, testified that he was unaware of any better data upon which to base revenue recovery assumptions than the Firefox and Apple Maps events. Id. 1638:13–1639:20. Google also used the same Firefox and Apple Maps events to make the same recovery assumptions five years later, in 2021, when considering the possibility of Apple entering the general search services market. UPX0148 at -826; Tr. 1633:16–21, 1635:21–1638:1 (Roszak (Google)). 926. The Apple Maps event has also been an input to model recovery rates if Google lost the default on Samsung’s S Browser and if Chrome and the Google Search App were not pre-installed on Android devices as well. Tr. 1640:19–1642:14 (Roszak (Google)) (discussing UPX0323 at –540, which projected [redacted]% to [redacted]% recovery loss based in part on “iOS Maps default change”); UPX0146 at -412 (“S Browser default changes and we clawback redacted%” based partly on “iOS Maps Recovery”); UPX6024 at -442–43 (written 30(b)(6) response: “[I]n analyzing provisions of the RSA in 2020, Google personnel used clawback estimates of between [redacted]% (based on Google’s experience with Apple Maps) to redacted% (based on Google’s experience on Microsoft’s Edge browser) to assess revenue related to default search on third-party browsers.”). 927. Nor was Google alone in its analysis; both Microsoft and Apple independently reached conclusions regarding query recovery that were similar to Google’s. Tr. 10497:2– 10500:5 (Whinston (Pls. Expert)). For its own analysis during the same negotiations in 2016, Apple assumed that % of queries would continue to run through the default were Apple to change the Safari default to Bing. UPX0095 at -331. This is nearly identical to Google’s weighted % revenue recovery (i.e., [redacted]% revenue loss) modelling. Supra ¶ 924. Two years later when considering acquiring or co-owning Bing, Apple—in estimates provided to CEO Tim Cook—predicted that a new, non-Google default search engine could retain between [redacted]% and [redacted]% of queries with an intermediate model of [redacted]% query retention. UPX0460 at -176; Des. Tr. 211:4–24, 212:23–213:12 (Perica (Apple) Dep.) (query retention rates of [redacted]%, [redacted]%, and [redacted]% in UPX0460 at -176 came from the relevant experts at Apple). 928. In 2016, Microsoft assumed that [redacted]% of mobile search volume going through the Safari default would shift to Bing if it were to become the default. Tr. 10497:2–10500:4 (Whinston (Pls. Expert)) (discussing UPXD106 at 19); UPX0069 at -236. This too resembles Google's modelled [redacted]% mobile revenue loss attributable to a default switch. Supra ¶918. Also like Google, Microsoft assumed "higher churn” on desktop “as it is easier for users to switch defaults on desktop browser.” UPX0115 at -142 (estimating that Microsoft could capture significantly greater query share on mobile iPhones (54%) than desktop Macs (16%)); Tr. 3264:23-3265:14 (Tinter (Microsoft)) ("Defaults matter less" on desktop). In fact, Microsoft even distinguishes "PC search" from "mobile search" because of differences in keyboards and screen sizes. Id. 3094:12–24. “On a mobile platform more than anywhere else, even more than on the PC, default is the only thing that matters." Des. Tr. 143:7-23 (van der Kooi (Microsoft) Dep.). 929. In a 2018 presentation to Apple, Microsoft projected that obtaining the Safari default could boost Microsoft's query share on iPhones by 55 percentage points (or 29 fold), from approximately 2% to 57%. UPX0116 at .028; Tr. 3278:10–3279:15 (Tinter (Microsoft)) (“This was the highest fidelity estimates that we had based on the best of our knowledge and our experience and our ability of what we thought was likely to happen”). And like Google, Microsoft used the Apple Maps event "to sanity check” Microsoft's assumptions for modelling a potential deal for the Safari default. Id. 3266:13-3268:3. 930. In sum, these similar models across three sophisticated companies are significant—and more than “thought experiments”—because they were used in very detailed ways to make business decisions about enormous sums of money. Tr. 5717:1-7, 5719:20-5721:25, 5751:1– 5752:5 (Whinston (Pls. Expert); supra ¶¶ 925-927 (Google's modelling were based on best available data and shared with CFO); supra ¶ 927 (Apple's estimates were based on analysis from experts and shared with CEO); supra ¶ 929 (Microsoft’s estimates were “highest fidelity estimates”). As Mr. Tinter explained, the accuracy of these models was important because they were being used as the basis for decisions about “very large financial investments” and as the basis for negotiations with partners. Tr. 3278:23–3279:15 (Tinter (Microsoft)). 931. Drawing from market participants’ estimates and real-world data, Prof. Whinston predicted that at least 33% of all U.S. queries—including 49% of all U.S. mobile queries—would shift from Google to its rivals if Google lost all the defaults controlled by the challenged contracts; this would quadruple rivals’ total U.S. query share. Tr. 5749:17–5752:5 (Whinston (Pls. Expert)) (explaining analysis with reference to UPXD104 at 31); id. 10497:2–10499:15 (explaining analysis with reference to UPXD106 at 19). 6. Google’s Enormous Payments And Microsoft’s Enormous Bids For Defaults Corroborate The Power Of Search Defaults 932. Because of the effect of defaults on query share, control of defaults are “very valuable.” Tr. 7684:18–20 (Pichai (Google)); id. 7666:19–24 (agreeing “Google pays for these exclusive defaults because they have value in distributing Google Search”); id. 7661:1–11 (“[T]here’s clear value” in making Google Search the default search engine in Safari.); Tr. 830:22–23 (Kolotorous (Google)) (agreeing Google finds defaults “valuable”); Tr. 4935:11– 18 (Braddi (Google)) (agreeing Apple’s Safari default “is valuable to Google because it drives more queries to Google Search”); Tr. 154:20–21 (Varian (Google)) (Being the default has “some advantages.”); Tr. 329:18–20 (Barton (Google)) (agreeing that he “view[ed] search defaults as “valuable”). 933. Because “the default is valuable,” Google pays “tens of billions of dollars every year” to be the default search engine across search access points in the United States. Tr. 7669:4–14 (Pichai (Google)). Asked by the Court, based on his experience at Google, for the business justification for Google’s payments to search distributors, Dr. Ramaswamy explained that defaults are “enormously powerful because . . . pious prose around ‘competition being a click away,’ notwithstanding in practice, they [users] don’t change.” Tr. 3796:5–3798:22 (Ramaswamy (Neeva)). Mr. Roszak, Google vice president of finance with long-time responsibilities related to the default distribution deals, is not aware of any default distribution deal that Google has ever entered where the traffic acquisition cost exceeded the expected incremental return. Tr. 1534:24–1536:21, 1540:19–22 (Roszak (Google)). 934. The total TAC payments incurred by Google’ s search business (i.e., including payments to partners for default distribution of Google Search) more than tripled between 2014 and 2021—from $7.1 billion to $26.3 billion. Tr. 7521:2–16, 7576:16–21 (Raghavan (Google)) (testifying about UPX0453 and UPX7002.A); UPX7002.A at .001. 935. Google’s overall worldwide revenue share payments to Apple were approximately $20 billion in 2022. Tr. 2492:22–2493:6 (Cue (Apple)). In fiscal year 2020, Google’s payments to Apple—the largest company in the world by market capitalization— constituted 17.5% of Apple’s operating income. Tr. 5727:5–5728:11 (Whinston (Pls. Expert)) (referring to UPXD104 at 19); UPX1110 at -41046 (For each quarter in fiscal year 2020, Google’s payments to Apple as a percentage of operating income were 21.5%, 19.2%, 21.1%, and 11.6%.); UPX0635 at -352 (Apple recognized in 2018 that “G search deal is [redacted]% of corporate EBIT.”); compare UPX8105 at -203 ($66.288 billion in Apple worldwide operating income for fiscal year 2020) with Tr. 2491:1–2492:8 (Cue (Apple)) (reviewing UPX1109 at -038, Apple’s fiscal year 2020 revenue share payments from Google were $11.58 billion.). Further, Google’s revenue-share payments to Apple have grown consistently. Id. 2489:23–2490:25 (discussing UPX1109 at -036, agreeing that Google’s monthly revenue-share payments to Apple “have been on a pretty steady upward trajectory” and grew from $418 million in January 2017 to $1.5 billion in August 2021); UPX0460 at -149 (Global ISA payments grew from $2.2 billion in 2014 to $8.6 billion in 2018.); UPX1109 at -026 (“Google reported $1,532M (+46% Y/Y) revenue share for the month of August ’21, recording the highest revenue share in history for the fourth straight month.”); Tr. 2488:2–2489:6 (Cue (Apple)) (The April 2021 record streak in Google revenue-share payments to Apple continued.). In fiscal year 2020, Google’s payments to Apple for Safari address-bar queries alone exceeded $9.6 billion and surpassed a billion dollars every month by May 2021. UPX1109 at -036, -038 (total of payments for “Safari Rev Share” from Oct. 2019 to Sept. 2020); Tr. 2488:9–2492:21 (Cue (Apple)). 936. Google has paid billions of dollars in revenue share to Android RSA partners over the years. Tr. 7667:12–15 (Pichai (Google)); Tr. 5727:5–5728:11 (Whinston (Pls. Expert)) (explaining UPXD104 at 19, Google has paid its RSA partners, including Android partners, “billions and billions and billions of dollars.”). In 2020 alone, Google’s payments to carriers and Android OEMs for U.S. searches totaled more than $1.5 billion. Id. 5727:5–5728:11 (referring to UPXD104 at 19). Google also gives away the “must-have” Play Store on Android at no monetary cost to Android OEMs. Tr. 9426:17–18 (Rosenberg (Google)) (There is no license fee for the MADA.); Tr. 5727:5–5728:11 (Whinston (Pls. Expert)). 937. Google’s revenue-share payments to Mozilla make up approximately 80% of Mozilla’s revenue. Des. Tr. 41:18–24 (Baker (Mozilla) Dep.). In 2021, Google’s revenue-share payments to Mozilla exceeded $[redacted] million worldwide. Tr. 538:1–15 (Rangel (Pls. Expert)) (referring to UPXD101 at 10); Des. Tr. 190:12–17, 190:21–191:2, 191:5–10 (Baker (Mozilla) Dep.) (Google’s revenue share to Mozilla “ends up in the hundreds of millions of dollars.”). 938. Google’s payments to third-party browser partners—excluding Apple—totaled more than $150 million in 2020 for U.S. default search traffic. Tr. 5727:5–5728:11 (Whinston (Pls. Expert)) (referring to redacted figures in UPXD104 at 19). 939. The massive size of Google’s default revenue share payments confirms the value of these defaults; Google is a sophisticated, successful, for-profit corporation that would ensure these payments are worthwhile. Tr. 5727:5–5729:5 (Whinston (Pls. Expert)) (“[W]hen you see Google paying billions and billions and billions of dollars, there’s got to be a reason. There has to be a reason it’s worth doing it.”); Tr. 538:1–15 (Rangel (Pls. Expert)). Google’s willingness to invest billions of dollars to secure defaults shows that the default bias in search is “of sufficient magnitude” for “participants in the market [to] care about” and invest billions of dollars to influence them. Tr. 524:15–25, 538:1–15 (Rangel (Pls. Expert)). 940. Consistent with his fiduciary obligation to shareholders, Mr. Pichai would not authorize contracts to pay billions of dollars every year for search defaults if they were not valuable. Tr. 7669:15–25, 7773:13–18 (Pichai (Google)); Des. Tr. 29:19–30:1 (Porat (Google) Dep.) (agreeing that “the value of the acquired traffic that Google gets from the TAC payments exceeds . . . the cost of the payments that Google makes” and that “[o]therwise, [Google] wouldn’t make the payments”); Tr. 3796:5–3798:22 (Ramaswamy (Neeva)) (“[S]earch is one of the most profitable businesses ever. And so defaults in search, therefore, have a very meaningful impact on . . . Google’s top and bottom line.”). 941. At a [redacted]% revenue recovery rate, supra ¶ 924, Google stood to lose more than $7 billion in 2020 and more than $28 billion between 2016 and 2020 if Apple changed the Safari default. UPX1050 at -887; Tr. 1629:19–1631:18 (Roszak (Google)) (discussing UPX1050 at -887). 942. The 36% gross revenue share that Google agreed to pay Apple in the 2016 amendment to the ISA, supra ¶ 226, offers an economic perspective of the magnitude of the default effect in search. Tr. 5728:12–5729:5 (Whinston (Pls. Expert)). For Google’s revenueshare payments to make economic sense, the revenue Google expects it would lose if a rival won the default must exceed the traffic acquisition costs of obtaining the default. Id. As a result, Google must expect to lose at least 36% of Safari default traffic revenue if Apple were to switch the Safari default. Id. (discussing UPXD104 at 20); Des. Tr. 29:19–30:1 (Porat (Google) Dep.) (agreeing that “the value of the acquired traffic that Google gets from the TAC payments exceeds . . . the cost of the payments that Google makes” and that “[o]therwise, [Google] wouldn’t make the payments”); id. 32:15–32:22 (agreeing that “the purpose of [Google’s] payment [to Apple] is to get the additional or marginal traffic that [Google doesn’t] get today”); Tr. 1540:5–9 (Roszak (Google)) (defining incremental revenue). 943. Ultimately, because defaults are the most potent method of distribution, Mr. Nadella was prepared for Microsoft to lose “billions of dollars” to secure the Apple’s search defaults. Tr. 3502:21–3505:23 (Nadella (Microsoft)) (“[Microsoft was] always grounded in the fact that if [Apple] needed to switch, that [Microsoft] would have to pay, and [Microsoft] would have to pay and even subsidize the transfer . . . we were going to be negative for multiple years perhaps, and I was willing to do essentially a very long term and think perpetually.”); Des. Tr. 111:20–112:20 (van der Kooi (Microsoft) Dep.) (Microsoft’s former CEO Steve Ballmer “was deeply committing to investing in the search business beyond . . . what was, let’s say, purely economically justifiable. . . . There is not really another company in the world that you can imagine that would have had the wherewithal and the staying power to incur that degree of losses over that period of time other than Microsoft.”). Continue Reading Here. Continue Reading Here . Here About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on April 30, 2024, storage.courtlistener is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on April 30, 2024, storage.courtlistener is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. storage.courtlistener [15] Behavioral economics is a branch of economics that applies concepts from psychology and neuroscience to improve understanding of human decision-making—including decisions made by consumers. Tr. 516:3–8 (Rangel (Pls. Expert)); UPX0101 at -280 (“Behavioral Science: A method of predictive analysis that applies economic modeling and psychological insight to the study of how people make decisions and develop attitudes and beliefs.”). For almost a decade, Google has relied on collaboration with Irrational Labs, a behavioral economics consulting firm, and its own internal behavioral science team led by Maya Shankar, the former head of behavioral economics in President Barack Obama’s administration. Tr. 531:22–532:17 (Rangel (Pls. Expert)); e.g., UPX1102 (2015 Irrational Lab presentation to Google); UPX0103 (2021 presentation from Google’s internal Behavioral Economics team).

Google's Monopoly: Defaults Have A Powerful Effect On Users’ Search Behavior

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

Untitled Story

20% of All General Searches in the U.S. Go Through the Default on User-Downloaded Version of Chrome

20% of All General Searches in the U.S. Go Through the Default on User-Downloaded Version of Chrome

Meet the Plaintiffs in the United States v Google Antitrust Lawsuit

The U.S. Government: Google Has Monopoly Power In The U.S. General Search Services Market

Google Has Monopoly Power In The U.S. Search Ads Market

20% of All General Searches in the U.S. Go Through the Default on User-Downloaded Version of Chrome

20% of All General Searches in the U.S. Go Through the Default on User-Downloaded Version of Chrome

Meet the Plaintiffs in the United States v Google Antitrust Lawsuit

The U.S. Government: Google Has Monopoly Power In The U.S. General Search Services Market

Google Has Monopoly Power In The U.S. Search Ads Market

Light-Mode

Classic

Newspaper

Dark-Mode

Neon Noir

Minty

HN StartUps