Google’s MADAs And RSAs Are Not Necessary For Android’s Success And Do Not Benefit Search Consumers

C. Google’s MADAs And RSAs Are Not Necessary For Android’s Success And Do Not Benefit Search Consumers By Improving Device Quality

1310. Google has failed to show that MADA and RSA terms granting Google default exclusivity on Android devices—including the MADA’s preinstallation and placement requirements and the RSA’s default and search exclusivity clauses—are necessary for Google’s ongoing support of the Android ecosystem, infra ¶¶ 1310–1320 (§ X.C.1), or have benefited consumers by improving the Android user experience, infra ¶¶ 1321–1344 (§ X.C.2).

1. The “Android Model” Would Not Collapse Absent Google’s Restrictive Contract Terms

1311. Google’s business model for the Android ecosystem (the “Android Model”) would not collapse if MADAs and RSAs did not require search default exclusivity. Google has numerous reasons for supporting the Android ecosystem, including the substantial revenues Google earns from Play Store sales. Google’s assertion that Android cannot compete against iOS without the MADAs and RSAs is inconsistent with the evidence and undermined by the substantial sums that Google pays Apple every year to be the exclusive default on iPhones and other Apple devices.

a) Google Has Alternative Reasons To Support Android And License The Play Store For Free

1312. Google has a strong incentive to invest in Android, regardless of whether search rivals have their GSEs pre-installed on those devices. Google uses the Android platform to distribute flagship, non-search applications that generate substantial revenue for the company, including YouTube, Google Maps, Gmail, and Google Drive. Tr. 9572:1–16 (Rosenberg (Google)) (agreeing that Google selects apps for the MADA bundle based in part on whether they “generate revenue for Google”); Tr. 7716:12–18, 7717:2–12 (Pichai (Google)) (Google generates revenue through distribution of core applications required by the MADA, including Gmail and Google Drive.); UPX0296A at -500 (listing “generate revenue for Google” as the first guiding principle for determining which new apps to include in the MADA bundle); UPX0286 at -211 (“Secure more users for Google services through the Android mobile platform via preloads, default intents, and/or exclusivity for over 30 Google mobile internet services.”); UPX6059 at -034 (Google’s 2021 10-K showing that YouTube ads revenue increased from roughly $20 billion in 2020 to more than $28 billion in 2021).

1313. Google’s Play Store, which is only available on Android devices, is one of the core apps that generates revenue for Google. Tr. 9553:12–16, Tr. 9551:5–18 (Rosenberg (Google)) (Google generates revenue from the Play Store through transaction fees on payments for and within apps, in addition to revenue from displaying ads in the Play Store). In one 2020 estimate, Google projected it would earn upwards of $[redacted]billion from Play Store sales on Samsung phones over a four-year deal. UPX2111 at -120; UPX0317 at -155 (showing $[redacted] billion in Play Store revenue in 2019).

1314. The more consumers buy Android phones the more money Google makes from Play Store sales. Tr. 9553:17–19 (Rosenberg (Google)). If consumers switch from an Android phone to an iPhone, Google loses all potential Play Store revenue from that user, although the ISA allows Google to retain those users’ search queries on iOS. Id. 9553:20–23; UPX2111 at -099 (“If we lose share, we lose Play revenue, and our Search TAC goes up to whatever the Apple rate is. We still get Search revenue; it’s just more expensive.”).

As Mr. Rosenberg explained, “[Google] certainly ha[s] commercial interest in seeing Android succeed because of Play.” Tr. 9556:10–25 (Rosenberg (Google)).

1315. Sales of Google-manufactured Pixel phones, which run on Android, generate substantial revenue for the company. UPX0317 at -159 (showing Pixel’s “OEM shipment share” in United States increased from 17% in 2018 to 20% in 2019); UPX6059 at -007 (Google 10-K noting, under the “Hardware” category, that Google generates revenues from Pixel phones).

These products collectively make Android incredibly lucrative for Google, even exclusive of revenue Google earns from searches run on Android devices. UPX6059 at -031, -034 (Google’s 2021 10-K showing that the “Google other” category—which includes Play Store revenues, Pixel phone sales, and YouTube non-advertising revenues—generated $28 billion).

1316. Google has not shown that Android smartphone sales would suffer if Google could no longer exclude competing GSEs from those devices. The evidence indicates the opposite; after Google implemented a choice screen in Europe, Android’s market share has declined less there than it has in the United States. Tr. 10593:2–10594:15 (Whinston (Pls. Expert)). Similarly, Android’s market share in Russia has increased since Google implemented a choice screen there. Id.

1317. Google’s competition with Apple also provides Google an incentive to continue investing in Android. If Apple released its own GSE and set it as the default on all Apple devices, Google would make substantially less money from searches run by Apple users. Tr. 10538:5–10539:13 (Whinston (Pls. Expert)); supra ¶¶ 1093–1124 (§ VIII.B.3). That would drastically increase Google’s incentive to invest in Android, because it makes more money on Android phones than on Apple devices. Id.

b) Google’s Purported Interest In Promoting Android’s Competitiveness Is Undermined By Its ISA Payments To Apple

1318. Google’s revenue share agreements with Apple undermine its search-based incentive to invest in Android. Tr. 10538:5–10539:13 (Whinston (Pls. Expert)).

1319. Google’s revenue-share payments to Apple under the ISA provide billions in capital to Google’s chief competitor in the smartphone market, yet Google is blithely unconcerned about strengthening Apple as an Android rival. Tr. 4974:9–14 (Braddi (Google)) (Google has never discussed any concern over how sharing billions in revenue with Apple makes Apple a stronger competitor in phone market.); Tr. 9566:20–9567:2 (Rosenberg (Google)) (recalling no discussions about halting Google payments to Apple); UPX0643 at -305 (noting Samsung’s argument that Google is giving Apple a better deal than Samsung, making it more difficult for Samsung to compete with Apple).

1320. By controlling the search defaults on both Apple and Android devices, Google Search bears little risk. When Android users switch to the iPhone, Google still gets substantial revenue as the default search engine on iPhones. UPX2111 at -099 (If an Android user switches to the iPhone, Google “still get[s] search revenue; it’s just more expensive.”); UPX0076 at -215 (showing iOS search revenue as roughly equal to Android search revenue).

c) Absent Exclusivity, Google Would Continue To Pay Revenue Share

1321. That Google’s search rivals pay for nonexclusive search suggests Google would have an incentive to continue paying for search distribution absent default exclusivity on Android devices. Bing, DuckDuckGo, and Yahoo pay for non-exclusive distribution in the United States, including for searches on Apple devices and Mozilla’s Firefox browser. Supra ¶¶ 38, 1307. In Russia, Yandex pays a 20% revenue share when users select Yandex from the Android choice screen. UPX0170 at -975 (“Yandex . . . pays Google a 20% TAC for search traffic on new devices if a user chooses Yandex . . . as default . . . and clicks on ads served by them.”).

2. There Is No Compelling Evidence That MADAs And RSAs Benefit Consumers In The Search Market

1322. Google has not shown that MADA and RSA terms that grant Google default search exclusivity on Android devices benefit consumers, including by improving the user experience and thus the competitiveness of Android devices. Even if Google had, those purported procompetitive justification do not justify harms to consumers in the general search services market.

a) The MADA And RSA Do Not Benefit Consumers

i. The MADA’s And RSA’s Preinstallation, Placement, Default, And Exclusivity Terms Do Not Improve The Customer Experience On Android Devices

1323. Google has not shown that its preload, placement, and default provisions improve the customer experience on Android devices.

1324. To the contrary, Google’s MADA and RSA terms lead to a poorer user experience in several ways. Both Android OEMs and Android users have complained about the placement of the Google Search Widget across the home screen of Android smartphones. UPX0128 at - 547 (email from Hiroshi Lockheimer stating, “1) Users generally are tired (visually) of our widget. It’s in the middle of the screen, obscures your family picture, etc.” and “2) OEMS are VERY tired (visually) of our widget, because of 1) and also because they feel like they don’t get to differentiate.”); Tr. 10099:2–23 (Murphy (Def. Expert)) (OEMs may not want to preinstall Google’s widget but do so in exchange for the benefits gained under the MADA.); UPX0653 at -053 (“If removable, [around] 20% of users delete widget within 3 months of device activation.”).

1325. The MADAs and RSAs also frustrate distribution partners’ ability to differentiate their products. Android OEMs, such as Samsung and Motorola, compete against each other by differentiating their devices, which can lead to innovative and compelling user experiences. Tr. 9574:22–9575:4 (Rosenberg (Google)) (Android manufacturers compete by differentiating their devices, and “innovation is one of the ways they can differentiate.”); Tr. 10098:3–7 (Murphy (Def. Expert)) (“There certainly are dimension[s] in which [Android OEMs] benefit from differentiation.”). For example, the foldable phones now offered by Samsung, Google, and Motorola are an innovation that resulted from OEMs competing through differentiation. Tr. 9574:22–9575:4 (Rosenberg (Google)).

1326. Google’s preinstallation, placement, default, and exclusivity terms thwart differentiation by standardizing features between devices. UPX0997 at -059 (discussing Google and AT&T’s “philosophical differences on the UX on Android” and AT&T’s desire to “differentiate the experiences so not all Android devices look the same”).

Google’s distribution partners routinely complain that MADAs and RSAs leave them with little control over the user experience on their devices. UPX0482 at -727 (Google Search deals leave T-Mobile “absolutely no control over the customer experience or monetization. Google is the one that controls both.”); UPX1036 at -835 (T-Mobile executives expressing frustration with, among other things, Google “trying to mandate with exclusivity and control”).

1327. For instance, T-Mobile explored building a federated search experience in which it would provide search results, if responsive to a user query, from Bing, Amazon, and Branch. Des. Tr. 181:4–183:15 (Giard (T-Mobile) Dep.). T-Mobile believed it could have created a better search experience for consumers, as well as improved monetization for T-Mobile, but even experimenting with a federated search experience was blocked by T-Mobile’s 2017 RSA. Id.

Similarly, Samsung and AT&T both expressed interest in preloading expanded versions of Branch’s app-search tool that included additional functionality but were deterred by concerns over violating their RSAs’ search exclusivity clauses. Des. Tr. 247:6–249:25 (Ezell (AT&T) Dep.) (explaining decision to walk away from expanded Branch partnership); UPX1064 at -543 (Samsung email expressing concern over Branch features that Google could characterize as “web search”).

1328. Google’s distribution partners have sufficient incentive to ensure a consistent and secure user experience without the MADAs and RSAs. Tr. 10539:21–10540:10 (Whinston (Pls. Expert)); Des. Tr. 273:15–274:23 (Giard (T-Mobile) Dep.) (A consistent user experience is in T-Mobile’s interest.); Des. Tr. 291:24–292:22 (Ezell (AT&T) Dep.) (Consumer experience takes precedence in business decisions.).

1329. OEMs and carriers know that if users have a poor experience on devices they purchase, users may blame the responsible OEM and carrier and users will shift their business to competitors. Des. Tr. 62:24–63:10 (Giard (T-Mobile) Dep.). Moreover, users confused about how to use their devices may call carriers’ customer care lines, which increases costs to distributors. Id.

1330. Although carriers have an interest in earning revenue shares from Google, the carrier device sale and services efforts represent a small portion of the carriers’ overall business. Des. Tr. 289:4–291:11 (Ezell (AT&T) Dep.) (estimating that revenue share payments AT&T receives total less than [redacted]% of its wireless service revenue). As Mr. Ezell explained, AT&T and the other carriers’ priority is being competitive in their core business of offering wireless services. Id. Carriers would not take steps that would drive customers to competitors, including distributing devices with poor user interfaces. Des. Tr. 62:24–63:10 (Giard (T-Mobile) Dep.).

ii. The “MADA Barter” Does Not Enable The Sale Of Low-Cost Devices

1331. Google has not shown that the “MADA Barter” has enabled the sale of low-cost Android devices. Google presented no evidence of a causal link between the MADA and the existence of low-priced Android phones.

1332. Although Prof. Murphy testified that such a link exists, he acknowledged that he is not aware of data showing one. Tr. 10185:25–10186:13 (Murphy (Google’s Expert)). He also has seen no ordinary-course documents linking the MADA with the sale of low-cost Android phones. Id. 10187:3–8. Prof. Murphy has not quantified how many, if any, low-priced Android devices would leave the U.S. market if the MADA bundle was disallowed. Id. 10187:9–15.

1333. As compared to less restrictive alternatives, the “MADA barter” is also inefficient because it requires the same “price” for every Android device. Tr. 10540:11–10541:9 (Whinston (Pls. Expert)). Currently, Google bundles all its GMS services, requiring the same placement and preload terms—e.g., placement of the Google Search Widget on the homepage—for every device regardless of its features and capabilities.

Every Android smartphone, no matter its size, cost, features, or other characteristics, must accept these conditions as “payment” for the Play Store. This makes less sense for certain devices, including those that have lower prices, fewer features, and storage and processing constraints. By contrast, licensing the Play Store separate from other GMS apps would provide Google and Android OEMs greater flexibility to tailor any “payment” for the Play Store according to the circumstances of the particular device. Id.

1334. Google has unbundled the MADA in Europe, permitting OEMs to license the Play Store directly. Tr. 10158:7–10159:1 (Murphy (Google’s Expert)). Google licensed the Play Store for different amounts depending on the country in which the device was being sold and the features that were available on the phone. Tr. 10540:11–10541:9 (Whinston (Pls. Expert)).

iii. RSAs Do Not Prevent Opportunistic Behavior

1335. There is no persuasive evidence to support Prof. Murphy’s opinion that RSAs’ search exclusivity clauses prevent Android OEMs and carriers from acting opportunistically.

1336. On balance, Google employees do not appear concerned about distribution partners acting opportunistically. Tr. 10177:21–10178:11 (Murphy (Google’s Expert)). Materials shared with senior Google executives, including Ms. Porat, do not include preventing free riding or opportunistic conduct among the rationales in support of RSAs. Id.; UPX1128 at -097 (providing rationales for new Android carrier and OEM RSAs).

1337. Google’s search rivals also are not concerned about distribution partners acting opportunistically. Bing, DuckDuckGo, and Yahoo pay for distribution in the United States, including for searches on Apple devices, without the guarantee of search exclusivity. Supra ¶¶ 38, 1307.

1338. Prof. Murphy’s opinion is inconsistent with his opinion that Google’s Android agreements do not harm competition for search services. If Prof. Murphy is correct that distribution partners have no desire to work with Google’s search rivals, there is no risk that distribution partners would steer queries toward those services and therefore no risk of them acting opportunistically. Likewise, if Prof. Murphy is correct that distribution partners are not interested in splitting queries between multiple search providers, there is no risk of distribution partners acting opportunistically. Tr. 10193:7–17 (Murphy (Google’s Expert)) (not aware of any instance in which a browser split queries between multiple search engines).

b) Even If There Was Evidence Of MADA’s And RSA’s Procompetitive Justifications, They Do Not Justify Harms To Search Consumers

1339. Even if Google had demonstrated that the MADA and RSA preinstallation, placement, default, and exclusivity terms improve the customer experience on Android devices (which it has not), Google’s purported justifications would not justify harms Google causes to consumers in relevant markets because (1) benefits to consumers in the general search services market are speculative at best; and (2) there are less restrictive means of increasing consistency and improving the user experience on Android devices.

i. Benefits To Search Consumers Are Speculative

1340. Improvements to the user experience on Android devices caused by the MADA and RSA, if any, inure to consumers outside the general search and related advertising markets. Google has not shown that those alleged improvements increase search output and, if so, by how much.

1341. Google identifies no evidence showing that the search-related provisions in the MADA and RSA have on balance improved the user experience on Android devices.

ii. Less Restrictive Means

1342. Google would continue to earn substantial revenues from Android users if Google offered an unconditional revenue share, or payments conditioned on a choice screen, and both alternatives would allow Google to encourage device improvements without foreclosing competition. Tr. 5776:22–5778:3, 10529:1–10530:17 (Whinston (Pls. Expert)) (discussing less restrictive alternatives).

1343. Encouraging Android OEMs and carriers to take user-enhancing actions, like making regular security updates, does not require Google to exclude other GSEs from Android devices. Android OEMs and carriers have independent incentives to push out security updates and ensure users have a high-quality phone experience. Tr. 7717:13–7718:15 (Pichai (Google)); id. 7658:17–7659:12.

1344. There is no technical reason user-enhancing actions, including security updates, must be tied to search defaults or exclusivity. Tr. 9564:3–15 (Rosenberg (Google)). As Google acknowledged at trial, it could use other models—including unconditional revenue shares or other financial incentive—to encourage Android partners to make timely security updates. Tr. 7718:16–7719:1 (Pichai (Google)); Tr. 9564:16–25 (Rosenberg (Google)).