Google Search Ads in the U.S.: Why They Compose a Relevant Market

B. Search Ads In The United States Compose A Relevant Market

398. As Dr. Varian recognized, Search Ads in the United States compose a relevant market. UPX0452 at -001 (“[T]here is a market for search advertising of course.”); Tr. 419:9– 421:19 (Varian (Google)) (discussing UPX0452); Tr. 4600:2–14 (referring to UPXD102 at 3) (Whinston (Pls. Expert)). Search Ads compose a distinct advertising channel. Tr. 5376:22– 5377:10 (Jerath (Pls. Expert)) (“[S]earch ads more broadly are a distinct product category of advertising.”).

1. The Characteristics And Uses Of Search Ads Support Finding A Search Ads Market

a) Search Ads Target A Consumer’s Real-Time, Self-Expressed Declaration Of Intent

399. Search Ads are the only ad channel displayed on a SERP in response to a user query. Tr. 1347:8–10 (Dischler (Google)) (Google does not show display ads in response to a query); Tr. 8829:2–6 (Israel (Def. Expert)) (Ads shown in response to search queries are Search Ads, and display ads are not done that way.). As a result, Search Ads are uniquely able to respond to a consumer’s real-time, self-expressed declaration of intent, whereas other ad channels are not. Compare supra ¶¶ 120–127 (§ III.C.2.c) with supra ¶¶ 128–137 (§ III.C.2.d).

400. Search Ads’ ability to target consumers’ real-time intent is a significant difference between Search Ads and other types of ads. As Adam Juda, Google’s VP of Ads, explained in an internal Google email, “SearchAds are fundamentally different than DisplayAds because they are targeted to the user’s query (and thus are relevant to the users’ task in a different way than an awareness–generating ad).”

UPX0459 at -871; UPX0416 at -117 (Dr. Juda writes that Google Display Network is “nowhere close to Search-like from my perspective (very different CPCs, CVRs [conversion rates], CTRs [click-through rates], user intent, etc.).”); UPX0272 at -699 (Microsoft “Privacy and Search Advertising” memo stating, “Search advertising is different from Display advertising in many respects,” including that Search Ads are “intent-driven” and “relevance-driven”); UPX0439 at -112 (“PLAs shown on shopping queries match the user intent . . . whereas display ads are often unrelated to the user intent.”); UPX0413 at -734 (Dr. Varian explains, “[Q]uery terms are a far stronger signal of user intent than past behavior,” which is used to target display ads.); UPX0428 at -863.010–11 (“A signal of intent is the most valuable thing you could have, more valuable than what the user likes or where the user is located”; other ads “don’t really have a signal of ‘intent’ like Search ads do.”).

401. Google advises advertisers to choose Search Ads for “[h]ighly specific targeting,” in contrast to display ads, which “let you reach a relevant audience . . . as they browse millions of websites, apps, and Google-owned properties . . . .” UPX8088 at -393 (directing advertisers to choose Search Ads for “[h]ighly specific targeting,” versus display ads for “[a]wareness and consideration”); Tr. 4644:15–4645:3 (Whinston (Pls. Expert)) (discussing UPX8088 at -393, excerpted in UPXD102 at 31); Tr. 6587:18–6589:2 (Vallez (Skai)) (“pull ads” (e.g., Search Ads) are served when a consumer is actively searching for information and “push ads” (e.g., video, display, and social media ads) are a “lean-back” passive experience where users are served ads that may or may not be relevant).

402. Google’s own guidance to advertisers further emphasizes as a “key take-away[]” that display ads are unable to target consumers and serve relevant ads the way Search Ads can, explaining:

While Search ads show up to potential customers the moment that they start looking on Google for what you offer, Display ads show up while people are visiting sites across the Google Display Network. When running Display ads, you might not reach those who are actively searching for what you offer. That said, you’re still introducing your business to a specific target audience who is likely to be interested in your products or services. This may help you to reach a larger or completely new audience than simply through search.

UPX8056 at .002 (Google Ads page); Des. Tr. 112:4–22 (Jain (Google) Dep.) (“If you’re reading the New York Times online and there’s ads there, those are display ads, because you didn’t search for anything. You’re just browsing for content. Most of the ads on Facebook are display ads if you didn’t search for anything.”); UPX8089 at -398 (Google Ads Help, “While the Search Network can reach people while they search for specific goods or services, the Display Network can help you capture someone’s attention earlier in the buying cycle. You can put your ads in front of people before they start searching for what you offer . . . .”).

403. Because Search Ads can target user intent more specifically via queries, Search Ads do not need to rely on broader techniques, such as audience targeting, upon which display ads and traditional advertising predominately rely. Tr. 1174:15–1176:1 (Dischler (Google)) (because they are served in response to a query, Search Ads are a more focused advertising format than formats such as television that are aimed at large, diffuse audiences); Tr. 8829:13–25 (Israel (Def. Expert)) (display ads target based on page and person, Search Ads target based on query); Des. Tr. 55:18–56:7 (Miller (Google) Dep.) (primary targeting for Search Ads is intent of the users and audience signals are secondary); Des. Tr. 111:21–112:22 (Jain (Google) Dep.) (“[F]or Search, as someone types in a query, we have the bulk of the signal from that. . . . [T]he notion of audience targeting is predominantly applied to display ads, not Search ads.”); UPX0022 at -908 (“Personalized data is not very helpful for search advertising since the query is such a strong signal.”); Des. Tr. 175:20–23, 175:25–176:5 (Daniels (Thumbtack) Dep.) (main difference between Text Ads and television ads is specificity by which one can target customers, given the latter is “broader based and less targeted”).

404. For example, for non-search channels, Kohl’s targets “audiences” with “attributes that [they] will assign to really define that target audience”; Kohl’s does not engage in similar audience targeting for Search Ads other than changing its bid strategies for existing customers. Des. Tr. 30:15–31:12 (Raymond (Kohl’s) Dep.).

405. Search Ads’ ability to target consumer’s real-time declared intent remains significant when compared to intent signals that can be derived from social media sites like Facebook, where users may express general interest in a subject, such as by interacting with different groups. For example, an Expedia executive noted that “even the theory behind [Search Ads and social ads] is pretty different,” explaining:

[I]t’s really based on intent and where someone is in a purchase decision. . . . [W]e would rarely know in a social channel if someone is trying to book travel right now. We would know they have an affinity for it. Maybe they’re interacting with different groups or different ads. But we wouldn’t be as likely to know I’m trying to go to Miami October 20th through 22nd, which is kind of the most information you would know in Google and a pretty frequent use case in Google.

Tr. 6512:23–6513:24 (Hurst (Expedia)); Des. Tr. 56:11–57:14, 287:7–289:2 (Utter (Microsoft) Dep.) (“In a search engine, you are describing your intent with a very firm signal, a query. On Facebook as a social platform . . . [t]he intent in what a consumer is trying to do is actually very different.”); Des. Tr. 174:24–175:3, 175:6–19 (Daniels (Thumbtack) Dep.) (The “main difference” between Text Ads and Facebook ads is that the former let one “target . . . search queries” while the latter do not “have the ability to target based on specific text search queries” and use “mostly demographic-based targeting.”).

406. Google itself recognizes a significant difference in targetability between Search Ads and social ads on Facebook. Tr. 1484:19–1485:13 (Dischler (Google)) (a significant difference between Facebook’s ads and Google’s Search Ads is that Search Ads are shown in response to a user’s query, but Facebook ads are not); UPX0418 at .001 (“Google knows identity AND intent. Facebook doesn’t know what users want”); Des. Tr. 112:9–17 (Miller (Google) Dep.) (explaining UPX0418 at .001: “[T]his whole document was trying to give sales teams collateral and talking points to advertisers about why they should spend with Google over Facebook. This is referring to that many users come to Google and tell us what they’re looking for through the search query, and that is a very powerful signal.”); Des. Tr. 112:23–114:3, 115:15–116:1 (Jain (Google) Dep.) (presence of a query distinguishes Search Ads from ads on Facebook).

b) Search Ads Are More Suited And Effective At Harvesting Demand Than Are Other Ads

407. Search Ads are more suited and effective at achieving goals related to harvesting demand, whereas other types of ads are more suited and effective at achieving goals related to generating demand. Tr. 3819:10–17 (Lowcock (IPG)). Compare supra ¶¶ 120–127 (§ III.C.2.c) with supra ¶¶ 128–137 (§ III.C.2.d).

408. Google itself recognizes this as a key distinction. For example, Google acknowledges that “[o]ne way to think about the difference between search and display/brand advertising is to say that ‘search ads help satisfy demand while brand advertising helps to create demand.’” UPX0411 at -638; UPX0433 at -826 (“Search ‘answers demand’ whereas Display ‘creates demand.’”).

409. Indeed, Search Ads perform a different function than other ad types and are more effective at harvesting demand.

For example, when The Home Depot conducted separate “go dark” incrementality tests to measure the impacts of turning off Search Ads and display ads, the results showed that turning off Search Ads resulted in a meaningful drop in revenue, with a greater impact on sales compared to when display ads were turned off.

Tr. 5146:12–5148:3, 5148:17–5149:3, 5170:7–17 (Booth (The Home Depot)) (explaining tests and discussing PSX00676 at -240); PSX00676 at -240 (“THD topline revenue dropped [redacted]% when we turned off paid search [(PLAs & Text Ads).] For every $1 invested into paid search, THD gains an incremental $[redacted].”).

410. As another example, Arjan Dijk, Senior Vice President and Chief Marketing Officer of Booking.com, noted that although consumers do click on Facebook ads and make purchases on Booking.com, such conversions are “very minimal” relative to the performance of Search Ads. Tr. 5242:10–13 (Dijk (Booking.com)).

411. Similarly, Ms. Lim explained that Search Ads are not fungible with display ads because Search Ads “exist for a different reason” and have a “singular objective” of driving acquisitions. Tr. 4858:4–16 (Lim (JPMorgan)) (“[P]aid search is an acquisition driver. . . . Whereas, a digital ad may be trying to drive awareness or, you know, increase demand for something . . . . So various different objectives happening in the digital space, singular objective in paid search.”); Tr. 5170:18–5171:8 (Booth (The Home Depot)) (he would not consider stopping spend on Search Ads and putting that money to display or social ads).

412. By contrast, display ads are focused more on generating demand as compared to Search Ads. Tr. 3819:10–17 (Lowcock (IPG)) (“Display advertising is primarily intended to drive or create demand” and “[s]earch advertising is there to capture intent after you have driven awareness.”); Tr. 4003:18–21, 4004:1–4005:1 (Juda (Google)) (relative to a Search Ad, a display ad is more likely intended to drive awareness); Tr. 4841:2–12 (Lim (JPMorgan)) (“[P]aid social, that’s—those are channels that we use to do a variety of different things, mainly awareness . . . . [I]t has a distinctly different role at our firm than paid search.”); Tr. 5241:17–5243:17, 5279:21– 5280:9 (Dijk (Booking.com)) (Booking.com advertises on Facebook, Instagram, and TikTok to drive awareness and consideration).

c) Search Ads Are Distinct From Retargeted Display Ads

413. Search Ads are distinct from retargeted (sometimes referred to as remarketing) display ads. Tr. 455:15–456:17 (Varian (Google)); Tr. 5220:9–13 (Booth (The Home Depot)) (retargeting ads cannot replace Search Ads); UPX0414 at -697 (“With remarketing, . . . [t]he user is worth more than a random display ad clicker, but less than someone who just searched.”).

414. Search Ads and retargeted display ads differ along four primary characteristics. Tr. 5445:12–5448:9 (Jerath (Pls. Expert)) (discussing UPXD103 at 20 and giving examples demonstrating how the differences between the two ad channels can play out).

415. (1) What type of intent the targeting is based on. With Search Ads, targeting is based on real-time declared intent, whereas with retargeted display ads, targeting is based on inferred intent from signals, which can become stale. Tr. 5445:12–5448:9 (Jerath (Pls. Expert)); UPX0026 at -764–65 (“Most of the value of retargeted ads occurs in first hour or so after the user visits the advertiser’s web page.”); Tr. 456:18–457:17 (Varian (Google)) (acknowledging authorship of UPX0026 and adopting this observation as true); id. 456:6–17 (the value of retargeting fades over time); UPX9001.A at 9:3–7, 6:24–7:5 (transcript of 2020 video of Dr. Varian discussing Search Ads: “[R]ecency is everything. . . . [R]eminder ads are a form of intent.

[But] [i]t’s not as strong as the intent when you are searching something . . . .”); Tr. 5220:9–5221:1 (Booth (The Home Depot)) (Retargeting ads can be shown even if the person has already purchased the product.); Tr. 5449:2–17 (Jerath (Pls. Expert)) (discussing UPXD103 at 21).

416. (2) Whom the ad is able to target. Search Ads can target anybody who types a relevant query, whereas retargeted display ads can only be shown to people who have already visited the advertiser’s website. Tr. 5445:12–5448:9 (Jerath (Pls. Expert)); Tr. 455:25–456:5 (Varian (Google)).

417. (3) What the content of the ad reflects. Search Ads typically reflect what the consumer is actually searching for, whereas retargeted display ads often show different products than what the consumer was initially interested in. Tr. 5445:12–5448:9 (Jerath (Pls. Expert)).

418. (4) What the impact is of ongoing changes in consumer tracking. Search Ads are uniquely unaffected by changes in consumer tracking because the consumer is declaring her interest at the moment of the query, whereas retargeted display ads are expected to become substantially less effective due to impediments to tracking consumers after they have left the advertiser’s website. Tr. 5445:12–5448:9, 5654:5–5655:25 (Jerath (Pls. Expert)) (cookie deprecation has “very little” impact on search ads especially in contrast with other kinds of advertising).

d) Google Does Not Sell Search Ads Through The Same Auction Process As Other Ads

419. Display ads do not participate in an auction with either Text Ads or shopping ads. Tr. 4006:23–25 (Juda (Google)).

420. Google’s ad auctions for its Google Display Network and video advertisements in video campaigns are modified first price auctions, as opposed to second price auctions. UPX6032 at -655–56 (Google admission (Resp. 14) that video and display auctions “involve a pricing mechanism where the price the advertiser pays is determined in part by the highest bid”); -656 Response No. 15 (admitting Text Ads auction is “a type of generalized second price auction in which the price an advertiser pays is determined in part by the auction runner up”).

e) Because Search Ads Are Targeted Using The Query, They Are Unaffected By Privacy Initiatives And Limitations On Cookie Tracking

421. Display ads, including social ads, and other non-Search Ads target users using “cookies” and audience profiling. Tr. 5445:24–5447:13 (Jerath (Pls. Expert)) (discussing UPXD103 at 20); UPX0413 at -735 (cookies are “used extensively for targeting display ads”).

422. Because they are targeted using cookies and audience profiling, display ads and other non-Search Ads are affected by privacy initiatives aimed at restricting these tools. Tr. 5445:24–5447:13, 5448:10–5449:1 (Jerath (Pls. Expert)); UPX8048 at .002. For example, Apple’s recent privacy initiatives have impacted Facebook’s ability to sell retargeted ads. Tr. 8835:3–8836:8 (Israel (Def. Expert)) (Meta/Facebook was affected by Apple’s and other privacy initiatives); Des. Tr. 164:13–165:6, 172:15–24 (Levy (Meta) Dep.)

(As a result of Apple’s privacy initiatives, Meta’s “ability to help advertisers target their ads has been impacted negatively” and Apple’s iOS 14 privacy changes have made “it much harder or potentially not even possible” for Meta to effectively retarget ads.); UPX0923 at -696 (Facebook executive emailing Meta CEO, COO, and others, in 2019, projecting that Apple’s privacy initiatives would have “seismic ramifications to our ads business . . .[redacted]% of revenue potentially at risk . . . [t]argeting crippled, retargeting impossible . . .”); UPX2117 at -035 (CFO in Meta Platforms, Inc.’s Fourth Quarter 2021 Results Conference Call, on Feb. 2, 2022, characterizing “impact of iOS [privacy initiatives] overall as a headwind on [Meta’s] business in 2022 is on the order of 10 billion”); UPX1018 at -726 (Meta board of directors meeting document predicting “regulatory and technology changes will cause an approximate [redacted]% revenue change to our 2023 long range plan”).

423. Indeed, Apple’s privacy initiative dramatically decreased advertisers’ ability to retarget display ads on Facebook and elsewhere. Des. Tr. 123:22–125:16 (Raymond (Kohl’s) Dep.).

424. However, and as conceded by Dr. Israel, because Search Ads rely primarily on a query, they are uniquely unaffected by these initiatives. Tr. 8831:10–8832:8 (Israel (Def. Expert)); Tr. 5445:24–5447:13, 5654:5–5655:25 (Jerath (Pls. Expert)) (“History leads to the signal. The signal is used in display advertising to infer your interest and so forth. And search advertising, what am I interested in? I just tell the search engine using my query. So, the relative value of any historical data is much less in search, and it is much more in display.”); Des. Tr. 167:9–11, 167:12–21 (Levy (Meta) Dep.) (“the impact on Meta, at least proportional to our business or proportionally, it would be larger for us than it would for Google” excluding YouTube).

425. Apple’s privacy initiatives have also limited the ability of Meta and others to identify which users viewing display and social ads later converted. Des. Tr. 177:23–178:4 (Levy (Facebook) Dep.). However, because sellers of Search Ads can track clicks on their ads, advertisers are still able to tell if a user, coming to their site via a click on a Search Ad, converted. UPX8047 at .001–02.

f) Google Developed A Non-Search Offering Because Of The Different Uses Of Social And Search Ads

426. Google’s Demand Gen product, formerly known as Discovery Ads and introduced in 2019, includes visually rich ads displayed on Google’s feed surfaces (YouTube, Gmail, the Discover feed, and the Google app), sold as a bundle, and not served in response to a query. Tr. 7538:17–22 (Raghavan (Google)); Tr. 1196:15–24 (Dischler (Google).

Google launched Demand Gen/Discovery ads “because [Google] found advertisers were asking [it] why they couldn’t get the types of social ads they could get from Facebook and Instagram from Google.” Tr. 7539:6–10 (Raghavan (Google)); id. 7349:21–22 (similar for TikTok); Tr. 1197:3– 5 (Dischler (Google)); Tr. 4646:9–4647:7 (Whinston (Pls. Expert)). Google recognized that it had “no direct competitor to Facebook’s ad offering,” and others’ social ads, and desired to build a new product to capture incremental advertiser spend. UPX0029 at -541 (“This new offering will enable us to capture dollars that are currently flowing exclusively to Facebook.”)

427. Discovery/Demand Gen campaigns “are aimed at the latent intent that Facebook, Instagram, TikTok, and others are capturing,” and are targeted using recent search and browsing behavior instead of keywords. Tr. 7543:25–7544:3, 7544:6–20 (Raghavan (Google)).

In developing the Discovery product, Google recognized the distinguishing factor between its search products and social offerings is that search captures a user’s declared intent through the query, therefore capturing demand, while social offerings create demand for new products. UPX0458 at -701 (Google is good at “driving conversions by following user intent expressed as queries on Search/Shopping properties,” but advertisers also “need solutions that create intent in order to acquire new customers.”).

Google thus crafted Discovery Ads to provide a “familiar pitch” to social buyers, which Google distinguished from search buyers based explicitly on differing types of intent: UPX0033 at -145; Tr. 7541:8–7543:15 (Raghavan (Google)).

428. Google marketed and continues to market Demand Gen/Discovery Ads to social buyers. Midway through trial, Google made a post to X (formerly Twitter) asking “Are your social ads falling short of expectation?

It’s time to demand more from social advertising!” and promoting Demand Gen. UPX2041 at -303; Tr. 7540:16–7541:2, 7541:8–7543:15 (Raghavan (Google)) (with discovery ads, Google focused on creating ads that offered parity with Facebook ads); UPX0033 at -117 (“Discovery ads launched, allowing advertisers to extend the reach of their social ads to 2.9B+ Google Users . . . Discovery Ads pitch is crisp and simple . . . Simply reuse your social assets and campaigns to drive performance”); UPX0409 at -252 (internal Google training stating Discovery Campaigns will “enable us to capture dollars that are currently flowing exclusively to Facebook/Instagram as well as provide incremental reach to existing performance campaigns”).

429. Demand Gen/Discovery Ad campaigns did not result in a shift of spend away from search, but instead created incremental revenue for Google. Tr. 7541:8–7543:15 (Raghavan (Google)) (Google Discovery ads do not cannibalize Search Ad revenues); UPX0033 at -146 (Google’s discovery ads do not cannibalize other Google digital ads.).

2. Industry Recognition

a) Google Considers Search Ads To Be Distinct And Complementary Products To Other Digital Ads

430. Google views Search Ads as distinct from and complementary to other types of advertising. Tr. 5449:18–5450:17 (Jerath (Pls. Expert)) (discussing UPXD103 at 22). For example, Dr. Varian testified that he considers Facebook ads (i.e., displays ads) to be distinct from Search Ads. Tr. 458:4–5 (Varian (Google)). Dr. Varian also acknowledged forwarding an article to colleagues and calling out the article’s conclusion that display ad impressions increased searches by 30 to 45 percent. Id. 462:7–15 (discussing UPX1085 at -616); UPX1085 at -616 (email with subject “spillover from display to search ads”).

431. Google separates Search Ads from display ads in its internal financial documents. UPX0476 at -664.

432. Google has conducted internal studies concluding that Search Advertisers adopting display products “does not lead to cannibalized Search spend. In fact, we see a small but statistically significant increase in Search spend post Display adoption.” UPX1118 at -191.

433. Google further views consumer experiences on Facebook as “fundamentally better suited for demand gen[eration]” and views Facebook itself as “lean[ing] into its reputation for driving demand to differentiate from Google Search.” UPX0458 at -704; UPX0030 at -752 (“Overall, compare FB to display, not search due to different expectations for the role it plays – top of funnel, image ads, more prospecting, awareness. Intent isn’t there.”).

b) Advertisers And Other Industry Participants Consider Search Ads To Be Distinct And Complementary Products

434. Advertisers and other industry participants view Search Ads as distinct from and complementary to other types of advertising.

435. Advertisers view Search Ads as an “always on” or “mandatory” ad channel, in contrast to other ad channels. For example, Ms. Lim described “paid search” as “an always-on channel . . . like an evergreen media channel,” running “[e]very day of every week of every month year-round.” Tr. 4841:13–4842:3, 4849:4–25, 4858:18–4859:9 (Lim (JPMorgan)) (further confirming, in response to a question from the Court, that “it’s common that financial institutions use paid search to drive acquisition year-round”).

Other ad channels, however, “operate on a campaign timetable . . . [and have] a beginning, a middle, and an end.” Tr. 4841:13–4842:3, 4842:15–4843:2 (Lim (JPMorgan)); DX0412 at 665 (Kohl’s advertising “flighting approach” showing constant Search spend during ad campaign but significant variation among video, social, and display). Similarly, Mr. Lowcock described Search Ads as “mandatory” for any advertising campaign because Search Ads capture user intent and search behavior can be driven by other ad channels. Tr. 3819:10–17, 3824:20–3825:3, 3826:11–18, 3980:3–3981:6 (Lowcock (IPG)); Tr. 5449:18–5450:17 (Jerath (Pls. Expert)) (discussing UPXD103 at 24).

436. Also, advertisers do not view other forms of advertising as substitutes for Search Ads. For example, an Expedia executive explained that there “isn’t a great substitute for the volume of high-intent customers you can find on Google” paid search. Tr. 6506:18–6507:11, 6515:21–6516:1 (Hurst (Expedia)). Mr. Dijk also noted that Booking.com purchases Search Ads to reach high-intent customers and cannot reach them effectively through other ad channels. Tr. 5236:18–5237:23, 5237:9–12 (Dijk (Booking.com)). Mr. Lowcock also does not consider Search Ads and display ads to be “substitutable.” Tr. 3824:20–3825:3, 3980:3–3981:6, 3937:25– 3838:2 (Lowcock (IPG)).

437. Industry participants recognize other ad channels as complementary to Search Ads. Des. Tr. 78:18–79:9, 80:4–12 (Levy (Meta) Dep) (Search Ads and social ads complement each other); Des. Tr. 25:10–26:9 (van der Kooi (Microsoft) Dep) (Snapchat, Pinterest, and TikTok “were not taking share from Microsoft Advertising” because “search is sort of unique in its characteristics on how it performs for advertisers and what it is used for.”); UPX1014 at -067 (Kenshoo presentation to Facebook stating, “The digital marketing portfolio has many unique channels and formats . . . . [S]earch and social stick out as being the most complementary in a cross-channel relationship.”); id. at -079 (Search Ads and social ads are “fundamentally different”); UPX0920 at -206 (Skechers employee noting “social is a lean back experience and does drive a lot of sessions to the funnel in general and helps with our paid search effort.”); PSX00970 at -726 (ad agency presentation, noting that TikTok brand awareness campaign improved Search Ads cost per action (CPA), i.e., conversion, by driving consumers to search for the client, where their demand was then captured by Search Ads)).

438. Advertisers typically have separate marketing teams or groups that focus on Search Ads compared to other ad channels. Tr. 4837:1–10, 4838:4–13, 4839:9–4841:1, 4843:10– 16 (Lim (JPMorgan)) (JPMorgan has three departments within its paid media practice: search (covering paid search and SEO); paid social; and programmatic (banner ads).); Tr. 5112:25– 5113:1, 5113:12–21, 5117:20–5118:21, 5124:2–8 (Booth (The Home Depot)) (The Home Depot has separate teams for paid search, for “digital media” (covering paid social and display), and for traditional advertising (e.g., television and print)); Tr. 6590:23–6591:1 (Vallez (Skai)) (advertisers use different teams to support the different ad channels); Tr. 3805:20–24 (Lowcock (IPG)) (IPG has an agency that specializes in Search Ads); Des. Tr. 186:9–187:9, 190:2–13 (James (Amazon) Dep.) (identifying different individuals responsible for paid search, display, and video); PSX00970 at -668 (listing ad agency’s team to support the client with separate staff for “Paid Search” and “Paid Social”).

439. Similarly, advertisers typically have separate budgets for Search Ads and other ad channels. Tr. 4856:16–4858:16 (Lim (JPMorgan)) (“Paid search budgets are for paid search only. . . . [I]t is not transferable between a programmatic buy across webpages and paid search.

They are distinct and different and separate.”); Tr. 5142:14–5143:4 (Booth (The Home Depot)) (“[O]ur Google and Bing investments are pretty much interchangeable but distinct and separate from social or display.”); Des. Tr. 289:3–290:13 (Utter (Microsoft) Dep.) (advertisers typically have separate budgets for different ad channels, such as for search and for display, with “a lack of . . . fluidity or fungibility, kind of, across those types of accounts”); Des. Tr. 49:19–50:3 (Alberts (Dentsu) Dep.) (paid search budgets are separate from budgets for other advertising channels); Des. Tr. 28:24–30:11 (van der Kooi (Microsoft) Dep.) (Because advertisers have dedicated search budgets, advertising on platforms like Snapchat, Pinterest, and TikTok does not come at the expense of Search Ads).

3. Uniqueness Of Production Facilities

440. For Search Ads that appear on GSEs, the uniqueness of production facilities present in the general search services market apply in the Search Ads market. Supra ¶ 370. For Search Ads that appear on specialized search engines, production facilities require the capability either (1) to crawl the webpages of a given Internet domain or set of domains to search for answers to a narrow range of queries or (2) to license such web-crawl results. Infra ¶¶ 531–540 (§ V.A.2.a), 557–558 (§ V.A.3). Google and Bing run separate auctions for Search Ads versus other types of ads. Tr. 4018:21–4019:20 (Juda (Google)).

441. Providing Search Ads requires the infrastructure and capabilities to (1) match Search Ads to consumers’ real-time queries, (2) pull those ads into the relevant auction, (3) determine which ads in the auction will be shown, (4) determine where on the SERP the shown ads will be positioned, and (5) calculate the price for each ad shown, should it be clicked on. Tr. 5458:23–5459:6, 5462:11–5463:8 (Jerath (Pls. Expert)).

442. Advertisers buying Search Ads expect to manage, measure, and optimize their ad campaigns. Tr. 5458:23–5459:6, 5462:11–5463:8 (Jerath (Pls. Expert)); UPX8042 at -096 (Google Ads Help on Google Ads Tools). Thus, providing Search Ads further requires the infrastructure to offer advertisers the tools and data for setting up and managing advertising campaigns. Tr. 5458:23–5459:6, 5462:11–5463:8 (Jerath (Pls. Expert)); UPX8042 at -096 (Google Ads Help on Google Ads Tools). Importantly, these capabilities must allow advertisers, in real-time, to evaluate the performance of and make changes to their advertising. Tr. 5458:23– 5459:6, 5462:11–5463:8 (Jerath (Pls. Expert)); UPX8042 at -096 (Google Ads Help on Google Ads Tools).

4. Distinct Prices And Sensitivity To Price Changes

443. Search Ads are generally priced by click while display ads are primarily priced by impression. Supra ¶¶ 101, 113; Tr. 1194:20–24 (Dischler (Google)) (display ads are primarily sold based upon impressions); id. 1196:5–14 (banner ads are sold predominantly based upon impressions).

444. [Intentionally Left Blank].

5. A Monopolist Of Search Ads Could Sustain Prices Significantly Above, Or Quality Significantly Below, The Competitive Level

445. Advertisers do not have reasonable substitutes for Search Ads to discipline a price increase or quality degradation. Tr. 5237:9–23 (Dijk (Booking.com)) (Booking.com cannot turn to other digital ads to reach the high-intent users it can with Search Ads). Facebook and other non-search digital advertising platforms do not constrain Search Ads pricing. Id. 5241:17- 5243:17 (Booking.com advertises on Facebook, Instagram, and TikTok to drive awareness and consideration, but not to reach high-intent users); UPX0519 at .010-12 (summarizing an internal Google study that concluded that Search Ads CPCs are times higher than Facebook social ads).

446. Non-digital advertising is not a reasonable substitute for Search Ads and, thus, not in the relevant market. Tr. 5279:23–5280:9, (Dijk (Booking.com)) (television advertising is higher in the funnel as compared to Search Ads and less effective than Search Ads); Tr. 3826:11-3827:11 (Lowcock (IPG)) (TV ads build awareness and interest that can then be converted to a sale when the user enters a query and is served a Search Ad).

6. The Advertising Industry Uses The Consumer Purchase Funnel

447. The consumer purchase funnel represents a conceptual framework of how consumers make purchase decisions. Tr. 1414:9–22 (Dischler (Google)) (the funnel is used in the advertising world and within Google as a construct for how advertisers think); Tr. 5381:6-20 (Jerath (Pls. Expert)). The concept applies to both products and services. Id. 5383:10–11.

448. Google itself uses the consumer purchase funnel, as shown below. UPX0427 at -030. Further, Google has touted the value of full-funnel marketing. UPX8051 at .001 (“Think With Google" post observing, "For years, [marketers] have heard that a holistic, full-funnel approach is critical for meeting customer where they are. Now we have the data to prove it."). UPX0427 at -030.

449. Although different variations of the consumer purchase funnel appear in academia and in practice, they reflect the same fundamental ideas that consumers go through stages when making purchase decisions. Tr. 5383:12–22 (Jerath (Pls. Expert)); Tr. 3879:8–3881:14 (Lowcock (IPG)). For example, consumers need to be aware of a product before they can buy it. Tr. 3815:17–19 (Lowcock (IPG)); Tr. 5382:13–19 (Jerath (Pls. Expert)). The term “funnel” reflects “the successive narrowing” or “empirical generalization” that fewer people progress to each successive stage. Tr. 5382:20–5383:3 (Jerath (Pls. Expert)) (discussing UPXD103 at 6).

450. One example of the consumer purchase funnel presents four stages: awareness, interest, desire, and action, where (a) first the advertiser makes the consumer aware of its product, then (b) the advertiser gets the consumer interested in the product, then (c) the advertiser increases the consumer’s desire for the product, and then (d) the consumer takes the desired action. Tr. 5381:21–5382:12 (Jerath (Pls. Expert)) (discussing UPXD103 at 6); Tr. 3879:7–3881:7 (Lowcock (IPG)).

451. The consumer purchase funnel does not presuppose strict linear progression through the funnel stages, and consumers can skip stages or go up and down the funnel. Tr. 6648:19–23 (Vallez (Skai)) (each consumer journey is unique and often non-linear); Tr. 3881:23–3882:16 (Lowcock (IPG)) (consumers can exit and enter the funnel at different points); Tr. 5383:4–09 (Jerath (Pls. Expert)).

a) The Consumer Purchase Funnel Is Widely Recognized And Used

452. Google and other companies that offer advertising products recognize and use the consumer purchase funnel. UPX0426 at -058; Tr. 8861:25–8862:7 (Israel (Def. Expert)) (consumer purchase funnel is “just sort of the lingo everybody uses,” including Google when it speaks to advertisers); UPX0445 at -508 (marketing funnel used in 2016 by Facebook executives in internal communications). Similarly, advertisers and agencies recognize and use the consumer purchase funnel. Tr. 5238:3–5239:6 (Dijk (Booking.com)); Des. Tr. 45:4–46:16 (Alberts (Dentsu) Dep.) (noting “in the world of marketing, we often refer to the sales funnel”); Tr. 3814:25–3815:15 (Lowcock (IPG)); PSX00682 at -406–08 (using stages of the funnel in internal IPG training materials).

Although an old concept, the consumer purchase funnel remains relevant today. Tr. 5240:9–16 (Dijk (Booking.com)) (The marketing funnel is “incredibly relevant” to analyze advertising even with the data now available.); Tr. 5381:6–20, 5384:5– 5386:3 (Jerath (Pls. Expert)) (The funnel represents “timeless concepts that advertisers use.” It is taught today in business schools, and new academic research is being published.).

b) Different Advertising Channels Align With Different Stages Of The Consumer Purchase Funnel

453. The consumer purchase funnel helps explain how different advertising channels align with different stages of the consumer purchase journey. Tr. 5383:23-5384:4 (Jerath (Pls. Expert)) (“[A]dvertisers can think of the goals that they need to meet for their campaigns and also sort of align which channels of advertising they would want to use to meet those goals."); Tr. 4848:23-4849:3 (Lim (JPMorgan)) (“Each media channel typically has its designated purpose.").

The funnel also explains how advertisers vary their choice of channels depending on a particular campaign's goals and objectives. Tr. 5383:23–5384:4 (Jerath (Pls. Expert)); Tr. 3814:19-3815:13 (Lowcock (IPG)) (explaining “[o]rienting these campaigns with the customer journey is critical so that you can align all assets housed within the campaign to a common and consistent goal" in UPX0926 at -684); Tr. 6587:4-17 (Vallez (Skai)) (Skai aligns the different advertising channels/ad types (e.g., social, search), with the customer journey on the path to purchase); PSX00970 at -692–93, -696-67, -700–01 (ad agency client presentation, listing social ads for awareness and consideration (i.e., upper and middle funnel goals) and Search Ads for consideration and purchase (i.e., middle and lower funnels goals)).

454. For example, display ads and Search Ads work differently, with display ads helping to create demand, and Search Ads helping to satisfy demand. Tr. 453:23-455:3 (Varian (Google)) (discussing UPX0411 at -638); UPX0431 at -708 (same); Tr. 5122:1-20 (Booth (The Home Depot)) (Displays ads and Search Ads work together across the funnel.).As Google expert Dr. Israel concedes, advertisers see different types of advertisements forming different tools along the journey to try to accomplish the goal of making a sale. UPX1014 at -080 ("Though search and social both cater to massive, highly engaged audiences, they tend to serve very different purposes along the consumer’s path to conversion.”); Tr. 8863:8–8863:14 (Israel (Def. Expert)).

7. The United States Is The Relevant Geographic Market For Search Ads

455. The United States is the relevant geographic market for Search Ads. Tr. 4654:3– 21 (Whinston (Pls. Expert)); UPX0342 at -870 (Search Ads revenue update breaking Search Ads revenue out separately by country, including United States).