Cabrera v. Google, LLC

Full title: RENE CABRERA, et al., Plaintiffs, v. GOOGLE LLC, Defendant.

Court: United States District Court, Northern District of California

Date published: Aug 15, 2023

5:11-cv-01263-EJD (N.D. Cal. Aug. 15, 2023)

Opinion

5:11-cv-01263-EJD

08-15-2023

RENE CABRERA, et al., Plaintiffs, v. GOOGLE LLC, Defendant.

EDWARD J. DAVILA UNITED STATES DISTRICT JUDGE

ORDER DENYING DEFENDANT'S MOTION FOR SUMMARY JUDGMENT; DENYING DEFENDANT'S MOTION TO STRIKE EXPERTS; AND GRANTING IN PART AND DENYING IN PART PLAINTIFFS' MOTION FOR CLASS CERTIFICATION RE: ECF NOS. 426, 442, 615

EDWARD J. DAVILA UNITED STATES DISTRICT JUDGE

This action involves a long-running dispute between Defendant Google LLC and advertisers who used its AdWords program. In this dispute, Plaintiffs Rene Cabrera and RM Cabrera Company, Inc. (“RMC,” and collectively with Cabrera, “Plaintiffs”) allege that Google failed to properly apply Smart Pricing discounts and failed to limit advertisements to the geographic locations that advertisers were targeting. Now before the Court are several motions: Google's motion for summary judgment (ECF No. 615), Google's motion to strike Plaintiffs' expert reports (ECF No. 442), and Plaintiffs' motion for class certification (ECF No. 426). After careful consideration of the pleadings, the parties' briefs, oral argument at hearing, and the record in this matter, the Court DENIES Google's motion for summary judgment, DENIES Google's motion to strike, and GRANTS IN PART and DENIES IN PART Plaintiffs' motion for class certification.

I. BACKGROUND

The Court has issued multiple opinions in this matter extending back over a decade and has summarized the facts on several occasions. See, e.g., ECF Nos. 85, 122, 253, 480. Accordingly, the Court here will only provide a brief overview of the facts relevant to this Order.

A. Plaintiffs

Cabrera is a Florida citizen who co-owns RMC (formerly known as Training Options). Fifth Am. Class Action Compl. (“Compl.”) ¶ 18, ECF No. 567. In 2008, he opened an AdWords account to purchase advertising on behalf of Training Options, and it is these purchases that are now the subject of Cabrera's and RMC's individual claims against Google. Pls.' Responsive Separate Statement Undisputed Facts (“SUF”), Fact 22-23 (undisputed), ECF No. 623-1; Compl. ¶¶ 7, 44. However, in 2009, Cabrera sold the assets of Training Options and allowed the corporation to dissolve. SUF, Facts 26-27 (disputed). The Court found that this sale included Training Options' claims against Google and that Cabrera lacked standing to raise those claims. 2/26/19 Order, ECF No. 480. To remedy this issue, Cabrera requested that the purchaser of Training Options assign the relevant claims back to Cabrera, and the purchaser agreed. Mustokoff Decl., Ex. 2, ECF No. 495-3. Cabrera then reinstated Training Options (renamed as RM Cabrera Company) for the purpose of bringing the instant claims. SUF, Facts 29-30 (undisputed).

B. Factual Background

As a broad overview, this action pertains to Google's AdWords program, which allows advertisers to create and display ads on webpages within Google's advertising networks. SUF, Fact 1 (undisputed). There are two components to those advertising networks: (1) the “Search Network,” which consists of google.com and Google's partner search websites, and (2) the “Display Network,” which consists of other webpages whose owners partner with Google to display advertisements. SUF, Facts 4-7 (undisputed).

AdWords operates via an auction system. SUF, Fact 8 (undisputed). When a user visits a webpage or performs a search, advertisers will compete in a second-price auction to display their ads to that user. Id. Unlike a typical auction where the winning bidder pays the amount of its winning bid, in a second-price auction, the winning bidder only pays a de minimis amount more (here, a penny) than the second-highest bid. Id. Fact 11 (disputed). So, for example, if the winning bid is 60 cents and the second-highest bid is 50 cents, the winning advertiser will only pay 51 cents. Bids are typically made on a cost-per-click basis, which is the amount the advertiser is charged when someone clicks on its ads. Id. Fact 9 (undisputed).

1. Smart Pricing

One feature that Google offers as part of AdWords is known as “Smart Pricing.” SUF, Fact 12 (disputed); Compl. ¶ 25. This feature discounts the price that advertisers pay based on the likelihood that a click “converts” or leads to a “conversion,” i.e., a successful transaction as defined by the advertiser. SUF, Facts 13-14 (undisputed). Put differently, when a click on one web property is less likely to convert than a click on a benchmark property, Google will provide a discount. Id. Facts 15-16 (disputed); Compl. ¶ 26. In practice, Google provides the Smart Pricing discount by applying a multiplier, where lower multipliers represent a greater discount. See Mazzeo Decl., Ex. 25 (“7/10/15 Solomon Rep.”) at 31-33, ECF No. 424-33 (using multipliers to calculate damages).

According to Plaintiffs, Google is contractually obligated to Smart Price all clicks. Compl. ¶ 5. Before an advertiser can join the AdWords program, it must sign a standard contract that the parties refer to as the AdWords Agreement. SUF, Fact 2 (undisputed); see also Compl., Ex. A (“AdWords Agreement”). Although the AdWords Agreement does not explicitly address Smart Pricing, it states that “[c]harges are solely based on Google's measurements for the applicable Program, unless otherwise agreed to in writing.” AdWords Agreement ¶ 7. In turn, extrinsic evidence allegedly shows that Smart Pricing is a “measurement” within the meaning of that clause. Compl. ¶¶ 50-53.

Plaintiffs claim that Google has violated this measurements clause of the AdWords Agreement in two ways: (1) by failing to Smart Price some clicks altogether and (2) by failing to provide a sufficient Smart Pricing discount for others. Id. ¶ 15 & n.6.

2. Location Targeting

When advertisers set up their campaigns on AdWords, Google allows the advertisers to select certain regions that their ads will target, a feature the parties refer to as “Location Targeting.” SUF, Fact 35 (disputed); Compl. ¶ 60.

Specifically, advertisers were asked, “In what geographical locations do you want your ads to appear?” This question was accompanied by a link that opened a text box with the following:

Location targeting

You can target your ads to almost any set of locations, including countries, territories, regions, cities and custom areas. For example, you could target specific regions with the United States and a few large English-speaking cities in Europe. You can view or edit your targeting options from the Settings tab for your campaign.

Compl. ¶¶ 61-62 (citing Compl., Exs. N, O). At the time Cabrera enrolled in AdWords on behalf of RMC, he encountered a different language conveying a similar message:

Target customers by regions / cities

Highlight the cities and regions on the left where you'd like your ad to appear, then click “Add.” Select as many regions as you like. You may also type city names directly into the box below.

Compl. ¶ 63.¹ Cabrera selected Florida, North Carolina, Georgia, and Louisiana as the locations for RMC's advertising campaign to target. Id. ¶ 64.

1 The Court will collectively refer to the various iterations of statements presented on the Adwords Settings screen where advertisers set up Location Targeting as the “LT Settings Screen” statements.

Once an advertiser selects the locations it wishes to target, Google will use those selected locations to determine whether to present an advertisement to a particular Internet user based on certain factors. SUF, Fact 35 (disputed). One such factor is the user's estimated physical location, as derived from the user's IP address. Id. Another factor Google considers is the content of the user's search query that may indicate the user's interest in an area (“Area of Interest”) regardless of that user's actual physical location; this process is known as “query parsing.” Id. Through query parsing, Google may show ads to someone searching “New York plumbers,” even if that search is from an IP address located outside of New York. Id.

Accordingly, even though RMC only targeted Florida, North Carolina, Georgia, and Louisiana for its advertising campaign, Compl. ¶ 64, it paid for clicks from users physically located in New York, Virginia, California, Illinois, Texas, and other states. Kaplan Decl., Ex. 7 (“11/8/18 Gibson Decl.”) ¶ 8, ECF No. 622-12. RMC claims that this was a violation of the California Unfair Competition Law (“UCL”).² Compl. ¶¶ 69-85.

2 This claim is brought only by RMC, not Cabrera.

C. Procedural History

As the age of this suit may indicate, the road to this point has not been smooth. This action began in 2011, when former plaintiff Rick Woods filed suit. Class Action Compl., ECF No. 1. After extensive motion practice, including summary judgment, the parties finally arrived at the class certification stage in 2018. ECF Nos. 274, 276. The Court denied class certification at that time, finding that Woods' entanglement with class counsel raised a conflict that rendered him inadequate to represent the putative class. 8/23/2018 Order, ECF No. 366. However, the Court granted leave to amend to substitute in a new proposed class representative. Id.

Shortly thereafter, Rene Cabrera, one of the current plaintiffs, joined the suit. Third Am. Class Action Compl., ECF No. 368. In November 2018, Cabrera filed a renewed motion for class certification.³ Mot. for Class Certification (“CC Mot.”), ECF No. 426. Google filed an opposition, and Cabrera replied. Opp. to Mot. for Class Certification (“CC Opp.”), ECF No. 444; Reply in Further Support of Mot. for Class Certification (“CC Reply”), ECF No. 457. Google also filed multiple motions under Daubert to strike various experts that Plaintiffs had set forth in support of class certification, only one of which remains to be resolved in this Order. See Google's Mot. to Strike (“MTS”) 6-11, ECF No. 442.

3 The docket entry for this motion reflects that it was filed by both Woods and Cabrera, though the motion itself purports only to be filed by Cabrera.

While Cabrera's motion for class certification was pending, Google filed a motion to dismiss. The Court granted Google's motion and dismissed Cabrera for lack of standing. 2/26/19 Order. Subsequently, the parties settled Woods' individual claims, after which Cabrera appealed the dismissal of his claims. Notice of Appeal, ECF No. 526. In early 2021, the Ninth Circuit reversed, vacating the Court's 2/16/19 Order and remanding for further proceedings. 1/4/21 Ninth Cir. Mem., ECF No. 533. Later that year, Cabrera joined RMC as a named plaintiff, the third different plaintiff since this action began. See Compl.

Because several years had passed since the parties initially filed their class certification papers in 2018, and a new plaintiff had joined the suit, the Court called for a supplemental round of briefing. See Plfs.' Suppl. Mot. Class Certification (“Suppl. CC Mot.”), ECF No. 579; Google's Suppl. Opp. Mot. Class Certification (“Suppl. CC Opp.”), ECF No. 598; Plfs.' Reply in Further Support of Suppl. to Mot. for Class Certification (“Suppl. CC Reply”), ECF No. 608.

Google also filed a motion for summary judgment after the parties had completed supplemental briefing on class certification. See Google's Mot. for Summary Judgment (“MSJ”), ECF No. 615; Plfs.' Opp. to Mot. for Summary Judgment (“MSJ Opp.”), ECF No. 623; Google's Reply in Support of Mot. for Summary Judgment (“MSJ Reply”), ECF No. 629.

Plaintiffs' 2018 motion for class certification, Google's most recent motion for summary judgment, and Google's Daubert motion are now before the Court.

II. LEGAL STANDARDS

A. Summary Judgment

Under Federal Rule of Civil Procedure 56, a court may grant summary judgment only when the moving party shows that there is no genuine dispute of material fact. A genuine dispute exists if there is sufficient evidence that a reasonable fact finder could decide in favor of the nonmoving party. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986). And that dispute is material if it might affect the outcome of the suit. Id.

In determining if a genuine dispute of material fact exists, a court must “tak[e] the evidence and all reasonable inferences drawn therefrom in the light most favorable to the nonmoving party.” Torres v. City of Madera, 648 F.3d 1119, 1123 (9th Cir. 2011). The moving party bears the burden of persuading the Court that there is no genuine dispute of material fact, and it also bears the initial burden of producing evidence that demonstrates there is no dispute. Cunningham v. Medtronic, Inc., 2018 WL 4053446, at *2 (N.D. Cal. Aug. 24, 2018) (citing Celotex Corp. v. Catrett, 477 U.S. 317, 323 (1986)).

When the moving party does not bear the ultimate burden of persuasion at trial, it may satisfy its burden of production at summary judgment by either (a) “produc[ing] evidence negating an essential element of the nonmoving party's claim or defense” or (b) “show[ing] that the nonmoving party does not have enough evidence of an essential element to carry its ultimate burden of persuasion at trial.” Nissan Fire & Marine Ins. Co. v. Fritz Cos., Inc., 210 F.3d 1099, 1102 (9th Cir. 2000). If the moving party fails to satisfy its burden of production, the nonmoving party need not produce anything-summary judgment will already have been defeated. Id. at 1102-03. But if the moving party fulfills its burden of production, the nonmoving party must produce evidence supporting its claim or defense. Id. at 1103. In that scenario, the nonmoving party will defeat summary judgment by producing enough evidence to create a genuine issue of material fact; failing that, the moving party will prevail on its motion. Id.

However, when the moving party bears the ultimate burden of persuasion, its initial burden of production is to “establish ‘beyond controversy every essential element of'” its claim or defense. S. Cal. Gas Co. v. City of Santa Ana, 336 F.3d 885, 888 (9th Cir. 2003) (citation omitted). If the moving party satisfies this initial burden, the nonmoving party can nonetheless defeat summary judgment by showing “the evidence, taken as a whole, could lead a rational trier of fact to find in its favor.” Id.

B. Expert Testimony

Courts act as the gatekeeper of expert testimony to ensure that such testimony is reliable and relevant under Federal Rule of Evidence 702. Kumho Tire Co. v. Carmichael, 526 U.S. 137, 147 (1999); see also Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579 (1993). The proponent of expert testimony has the burden of proving admissibility. In re Korean Ramen Antitrust Litig., 281 F.Supp.3d 892, 931 (N.D. Cal. 2017) (citations omitted). Before an expert can offer her opinions, she must be qualified by “knowledge, skill, experience, training, or education.” Fed.R.Evid. 702. Once she is qualified, Rule 702 permits her to testify as long as “(a) the expert's scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and (d) the expert has reliably applied the principles and methods to the facts of the case.” Id. This multifactor inquiry is flexible, and “Rule 702 should be applied with a ‘liberal thrust' favoring admission.” Wendell v. GlaxoSmithKline LLC, 858 F.3d 1227, 1232 (9th Cir. 2017) (citations omitted).

Although courts must screen expert testimony for reliability, what they assess “is not the correctness of the expert's conclusions but the soundness of [her] methodology.” City of Pomona v. SQM N. Am. Corp., 750 F.3d 1036, 1044 (quoting Primiano v. Cook, 598 F.3d 558, 564 (9th Cir. 2010)). In other words, Daubert is not a “guarantee[] of correctness.” i4i Ltd. P'ship v. Microsoft Corp., 598 F.3d 831, 855 (Fed. Cir. 2010). “[T]he case law-particularly Ninth Circuit case law-emphasizes that a trial judge should not exclude an expert opinion merely because he thinks it's shaky, or because he thinks the jury will have cause to question the expert's credibility. So long as an opinion is premised on reliable scientific principles, it should not be excluded by the trial judge.” In re Roundup Prods. Liab. Litig., 390 F.Supp.3d 1102, 1109 (N.D. Cal. 2018).

C. Class Certification

Federal Rule of Civil Procedure 23 sets forth the two-step process for certifying class actions. First, a plaintiff must establish that “(1) the class is so numerous that joinder of all members is impracticable; (2) there are questions of law or fact common to the class; (3) the claims or defenses of the representative parties are typical of the claims or defenses of the class; and (4) the representative parties will fairly and adequately protect the interests of the class.” Fed.R.Civ.P. 23(a). Second, the plaintiff must separately show that the proposed class fits into one of the three categories of Rule 23(b). In the present case, Plaintiffs seek to invoke the third category, Rule 23(b)(3), which requires the Court to find that “the questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.” Fed.R.Civ.P. 23(b)(3).

To meet their obligations under Rule 23, plaintiffs “must actually prove-not simply plead-that their proposed class satisfies each requirement of Rule 23” by a preponderance of the evidence. Olean Wholesale Grocery Coop., Inc. v. Bumble Bee Foods LLC, 31 F.4th 651, 664-65 (9th Cir. 2022) (en banc) (citation omitted), cert. denied, 143 S.Ct. 424 (2022). Courts must conduct a “rigorous” analysis of the Rule 23 factors that will often “entail some overlap with the merits of the plaintiff's underlying claim.” Ellis v. Costco Wholesale Corp., 657 F.3d 970, 980 (9th Cir. 2011) (quoting Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 351 (2011)). However, “Rule 23 grants courts no license to engage in free-ranging merits inquiries at the certification stage. Merits questions may be considered to the extent-but only to the extent-that they are relevant to determining whether the Rule 23 prerequisites for class certification are satisfied.” Amgen Inc. v. Conn. Ret. Plans & Tr. Funds, 568 U.S. 455, 466 (2013).

III. GOOGLE'S MOTION FOR SUMMARY JUDGMENT ⁴

4 To the extent Plaintiffs argue that the Court's prior denials of summary judgment constitute law of the case and are entitled to controlling weight, 7/7/22 Hr'g Tr. 134:18-25, ECF No. 653, the Ninth Circuit has expressly held that “the denial of a summary judgment motion is never law of the case because factual development of the case is still ongoing.” Peralta v. Dillard, 744 F.3d 1076, 1088 (9th Cir. 2014).

A. Breach of Contract Claim (Smart Pricing)

Google seeks summary judgment on Plaintiffs' Smart Pricing claims on three grounds: (1) Plaintiffs have not offered any evidence showing that they were damaged because they have failed to account for the auction dynamics that determine AdWords pricing; (2) Cabrera has no claim because any damage was suffered by RMC, the third-party beneficiary of the AdWords Agreement; and (3) the AdWords Agreement does not specify a particular Smart Pricing formula, so there is no breach as to any clicks that were Smart Priced “to at least some degree.” MSJ 6-7.

1. Lack of Damage

Google opens by asserting Plaintiffs have not demonstrated appreciable and actual damage. This argument rests on two premises: the assumption that Smart Pricing applies to bids as opposed to the final cost-per-click paid, and the fact that AdWords conducts second-price auctions. MSJ 3, 7. From Google's perspective, the consequence of these two premises is that Plaintiffs would not have been damaged if, after Smart Pricing discounted their winning bid, the discounted bid was still greater than the second-place bid. Id. at 7-8 & n.4. That is because Plaintiffs would have paid the same amount-one cent more than the second-place bid- regardless of whether their bid was Smart Priced or not. Google contends that Plaintiffs have failed to adduce evidence showing this scenario did not occur, and therefore summary judgment against Plaintiffs is proper. Id. at 8. Plaintiffs respond that it is Google who has failed to affirmatively produce evidence showing that such scenario occurred, and they also argue that Smart Pricing applies to post-auction costs rather than bids. MSJ Opp. 9-11.

The Court begins by addressing the burden of production, as both parties fault the other for failing to produce evidence. The requirement of “appreciable and actual damage” is an element of a claim for breach of contract under California law. Aguilera v. Pirelli Armstrong Tire Corp., 223 F.3d 1010, 1015 (9th Cir. 2000).⁵ Plaintiffs, therefore, have the burden of persuasion at trial to show that they suffered actual damage. As a result, Google as the moving party can meet its burden of production at summary judgment “by pointing out . . . that there is an absence of evidence to support [Plaintiffs'] case.” Nissan Fire & Marine, 210 F.3d at 1105 (quoting Celotex, 477 U.S. at 325). Although Google could alternatively meet its burden by negating the element of actual damage, it is not obligated to do so. Id. at 1106. Consequently, Google was not required to produce evidence substantiating its damages argument.

5 California courts have suggested that nominal damages are available for breach of contract, so Aguilera was mistaken to require appreciable and actual damages. E.g., Elation Sys., Inc. v. Fenn Bridge LLC, 71 Cal.App. 5th 958, 967 (2021). However, this Court is bound to follow Aguilera as Ninth Circuit precedent. Ruiz v. Gap, Inc., 380 Fed.Appx. 689, 692 (9th Cir. 2010) (following Aguilera despite finding that California law supports the availability of nominal damages in contract actions).

Nonetheless, summary judgment is not warranted on this basis because there is a genuine dispute of material fact over the argument's underlying premise. Google's theory is based on the presumption that Smart Pricing is properly applied to bids rather than the post-auction cost of a click. However, if the AdWords Agreement requires Google to Smart Price post-auction costs instead, the dynamics of a particular auction would be immaterial. Under that interpretation of the AdWords Agreement, damage would depend on Google's failure to Smart Price the post-auction cost of a click, irrespective of the auction history or how bids were priced.

In arguing that there is no dispute, Google points to evidence that, in practice, it Smart Priced bids, not post-auction costs. See 4/16/15 Harrower Decl. ¶ 8, ECF No. 208-8. While Google may be correct about its practices, those practices do not affect the determination and interpretation of Google's obligations under the AdWords Agreement. This is not mere legal pedantry. Plaintiffs have marshalled evidence in support of their interpretation of the AdWords Agreement. See Kaplan Decl., Ex. 63, ECF No. 622-66 (summary of Google documents and public statements indicating that Smart Pricing applies to the post-auction cost of clicks rather than bids)⁶; 7/10/15 Solomon Rep. at 9-15 (summarizing both public and internal statements by Google that Smart Pricing applies to the cost of clicks). What is more, Google's own evidence supports Plaintiffs' position. Mandhania Decl., Ex. 2, ECF No. 615-4 at 2 (Google Learning Center page stating: “While you set one CPC bid, if our data shows that a click from a content page is less likely to turn into actionable business results . . . we reduce the price you pay for that click.”) (emphasis added). That is not to say Google's position lacks all evidence. See FAC, Ex. E at 2 (“Smart pricing . . . discounts advertiser bids accordingly.”) (emphasis added). But at this stage, there is enough of a dispute over the meaning of the AdWords Agreement to preclude summary judgment for lack of appreciable damage.

6 Many of the documents summarized in Exhibit 63 are identified only by a Bates number, making it impossible for the Court to determine the identities and weight of those documents. Still, there are enough documents labeled by URL-from which the Court can glean the identities of some documents-for the Court to understand the significance and meaning of the exhibits.

2. Presence of Third-Party Beneficiary

Next, Google advances two arguments related to RMC's status as a third-party beneficiary of the AdWords Agreement.⁷ Specifically, it asserts that Cabrera's breach of contract claim must fall in light of RMC's claim because RMC suffered the relevant damages instead of Cabrera. MSJ 9. Google also argues that permitting both RMC's and Cabrera's claims to proceed would result in an impermissible double recovery. Id. at 11. In rejoinder, Plaintiffs contend that the Ninth Circuit already addressed the viability of Cabrera's claim when it held that Cabrera suffered a cognizable injury, and they assert that any concerns about double recovery should be addressed at the damages phase, not at summary judgment. MSJ Opp. 13-14.

7 Google does not contest that RMC is an intended third-party beneficiary who can sue for breach of contract.

The Court starts with Google's argument that only RMC may sue for damages. Under California law, when a contract is intended to benefit a third party, the signatory who received the contractual promise (the “promisee”) may sue for damages only in limited situations. In re Marriage of Smith & Maescher, 21 Cal.App.4th 100, 107 (1993).⁸ If the third-party beneficiary is a “creditor”-meaning that it receives the benefits of the contract in satisfaction of a debt or obligation owed by the promisee-then the promisee can recover damages as long as there is no conflict with the rights of the beneficiary or the party who made the promise. Id. The rationale is that “the promisee may suffer substantial damages as a result of breach,” as a breach may cause the promisee to default on her obligations. Id. (quoting Restatement (Second) of Contracts § 305 cmt. a); see also Restatement (Second) of Contracts § 307 cmt. c (similar). On the other hand, if the beneficiary is a “donee”-meaning that the performance promised to the beneficiary is a gift- the promisee has no economic interest in the performance and cannot sue for damages. Smith & Maescher, 21 Cal.App.4th at 107 (citing Restatement (Second) of Contracts § 307 cmt. d).

8 The Smith & Maescher court technically interpreted Massachusetts contract law when discussing the ability of a promisee to recover damages. 21 Cal.App.4th at 106. However, it relied on common law principles from the Restatement (Second) of Contracts, and “the Restatement[,] . . . ‘in the absence of a contrary statute or decision in this state . . . is entitled to great consideration as an argumentative authority.'” Lake Almanor Assocs. L.P. v. Huffman-Broadway Grp., Inc., 178 Cal.App.4th 1194, 1200 (2009) (quoting Canfield v. Sec.-First Nat'l Bank, 13 Cal. 2d 1, 30-31 (1939)). The Court has found no decision or statute contrary to the Restatement sections cited by Smith & Maescher. Moreover, another court in this district has adopted the reasoning of Smith & Maescher when ruling on a California contract claim. See Ehret v. Uber Techs., Inc., 68 F.Supp.3d 1121, 1140 (N.D. Cal. 2014). The Court therefore follows the approach outlined in Smith & Maescher.

Although it is not clear from case law which party bears the burden at trial on the issue of creditor and donee beneficiaries, the Court need not answer that question now. Google has wholly ignored the distinction between creditor and donee beneficiaries, offering only the erroneous categorical statement that promisees like Cabrera can never sue for damages when there is an intended third-party beneficiary. MSJ 12. Because Google has not addressed the issue at all, it has failed to meet its burden of production whether or not it bears the ultimate burden at trial.⁹

9 The Court need not address Plaintiffs' argument that the Ninth Circuit already resolved Google's argument in the prior appeal in this matter, but it observes that the Ninth Circuit's decision dealt with Article III standing, not with issues of damages under California contract law. See 1/4/21 Ninth Cir. Mem.

For similar reasons, summary judgment on Google's double recovery argument is also unjustified. Even if Google is correct that the rule against double recovery requires dismissal of one of the Plaintiffs, it has offered neither evidence nor viable argument about which Plaintiff must be dismissed. The Court has already found as insufficient the argument that Cabrera cannot sue for damages because RMC is the intended third-party beneficiary. All that remains is Google's unsupported assertion that it is Cabrera who must go. MSJ 11-12. Such bald assertions are not enough to dismiss Cabrera on summary judgment. To the extent that Google is concerned about double recovery, it may address those concerns at the damages stage of litigation.

3. Smart Pricing Formula

For its final argument, Google contends that the AdWords Agreement does not contain a promise to apply a specific Smart Pricing formula. MSJ 13. According to Google, in the absence of a specific promise, the AdWords Agreement cannot be interpreted to require any particular discount, so summary judgment should be granted as to all clicks receiving at least some Smart Pricing discount. Id. at 15. Plaintiffs disagree, believing that the Court already considered and rejected this argument in an earlier order. MSJ Opp. 14-15.

Plaintiffs are correct that Google's argument is not a new one. Google placed the argument squarely before the Court in its prior motion for summary judgment. ECF No. 205 at 13 (motion); ECF No. 238 at 2-5 (reply). The Court did not explicitly address the argument when resolving that motion, 9/28/2017 Order, ECF No. 253, but it did not grant Google's motion on that basis, thereby implicitly rejecting the argument. Although the factual record now before the Court is different than the one from Google's earlier motion, the record remains the same as to the relevant issue-interpretation of the AdWords Agreement. In particular, the extrinsic evidence germane to interpreting the AdWords Agreement remains the same. Compare Second Am. Compl., Exs. B-G, ECF Nos. 87-2 to -7 (operative complaint for prior summary judgment motion), with Compl., Exs. B-G (operative complaint for current summary judgment motion). Thus, there is nothing in the new record that compels the Court to depart from its prior holding.

That said, to provide guidance to the parties as this action proceeds, the Court elaborates on its reasoning. The pertinent provision of the AdWords Agreement provides that “[c]harges are solely based on Google's measurements for the applicable Program.” AdWords Agreement ¶ 7. Google's view is that this language contains no promises regarding Smart Pricing, and the inquiry into contractual meaning can end there. It stresses that the AdWords Agreement is fully integrated, so extrinsic evidence cannot be used to add to its terms. MSJ 14 (citing Casa Herrera, Inc. v. Beydoun, 32 Cal.4th 336, 343 (2004)). But that is not quite accurate. Extrinsic evidence can still be used to “explain the meaning of a written contract . . . [if] the meaning urged is one to which the written contract terms are reasonably susceptible.” Casa Herrera, 32 Cal.4th at 343 (alteration in original) (citation omitted). In other words, “extrinsic evidence can be admitted to explain the meaning of the contractual language at issue, although it cannot be used to contradict it or offer an inconsistent meaning.” Hot Rods, LLC v. Northrop Grumman Sys. Corp., 242 Cal.App.4th 1166, 1175 (2015).

As the Court previously found, when read in the context of extrinsic evidence provided by Plaintiffs, the relevant contractual clause is reasonably susceptible to the interpretation that all clicks must be Smart Priced. 8/24/12 Order 8, ECF No. 85. The extrinsic evidence also addresses how Google should Smart Price clicks: In an FAQ about AdWords, Google explained, “Smart pricing uses conversion data to help determine if Search and Display Network partner sites are likely to convert at different rates and discounts advertiser bids accordingly.” Compl., Ex. E at 2 (emphasis added). A fact finder could conclude that this means Google must apply discounts “according” to its conversion data, scaling the size of the discount such that an advertiser receives a greater discount as conversion rates decrease. Under this understanding of the AdWords Agreement, Google is required to do more than apply some arbitrary discount to clicks. Plaintiffs can maintain claims for improper Smart Pricing if Google applies a discount that is not appropriately pegged to conversion rates.

As this illustrates, there is extrinsic evidence supporting Plaintiffs' theory of liability, and resolving the parties' contract interpretation dispute requires the Court to weigh that extrinsic evidence against contract language that does not expressly refer to Smart Pricing. Notwithstanding the need to weigh extrinsic evidence, Google urges the Court to resolve the contract interpretation issue now. MSJ 15. And it is true that, in some instances, the extrinsic evidence may be “so one-sided that no reasonable person could decide to the contrary.” Unicom Sys., Inc. v. Elec. Data Sys., Corp., 2005 WL 5801534, at *12 (C.D. Cal. Nov. 1, 2005) (quoting Bourque v. FDIC, 42 F.3d 704, 708 (1st Cir. 1994)). But that is not the case here. The Court finds that there is a genuine dispute of material fact over the proper interpretation of the AdWords Agreement, so summary judgment is not appropriate.

* * *

In conclusion, the Court finds that summary judgment in favor of Google on Plaintiffs' Smart Pricing breach of contract claim is not warranted. Despite Google's efforts to contest Plaintiffs' theory of damages and interpretation of the AdWords Agreement, Plaintiffs have advanced enough evidence to establish questions of fact unsuitable for summary judgment. And while the evidence may ultimately establish that Cabrera may not file suit for damages because RMC is the third-party beneficiary of the AdWords Agreement, Google has failed to produce evidence justifying summary judgment on that argument. As such, Google's motion for summary judgment on the Smart Pricing breach of contract claim is DENIED.

B. UCL Claim (Location Targeting)

With respect to RMC's Location Targeting UCL claim, Google moves for summary judgment on three bases: (1) RMC could not have reasonably relied on Google's “LT Settings Screen” statement; (2) the statement is not material to reasonable advertisers; and (3) RMC did not suffer a cognizable UCL injury. MSJ 16.

1. Reasonable Reliance

Google first contends that RMC could not have reasonably relied on the LT Settings Screen's statement because a Google Help Center disclosure had separately indicated that ads may be shown to users “regardless of the user's physical location.” MSJ 18. RMC responds that reasonable reliance is not an element under the UCL; that a separate disclosure cannot cure or shield a misrepresentation from liability; and, in any event, there is at least a genuine issue of fact as to whether reliance on the LT Settings Screen's statement was reasonable. MSJ Opp. 16-21.

As an initial matter, the Court agrees with RMC's characterization of the law: reasonable reliance is not required for it to prevail on the UCL claim. It is well-established that the elements for a UCL claim for fraudulent business practices are distinct from those of common law fraud. “A common law fraudulent deception must be actually false, known to be false by the perpetrator and reasonably relied upon by a victim who incurs damages. None of these elements are required to state a claim for injunctive relief under the UCL.” In re Tobacco II Cases, 46 Cal.4th 298, 312 (2009) (internal brackets and quotation marks omitted) (emphasis added); see also Berger v. Home Depot USA, Inc., 741 F.3d 1061, 1068 (9th Cir. 2014) (“[T]he victim's reliance on the false statements . . . [is] not required to show a violation of California's UCL.”), abrogated on other grounds by Microsoft Corp. v. Baker, 582 U.S. 23 (2017). Thus, a plaintiff asserting a UCL fraud claim can prevail “even without allegations of . . . reasonable reliance.” Collins v. eMachines, Inc., 202 Cal.App.4th 249, 258-59 (2011). That plaintiff need only show that “members of the public are likely to be deceived.” Tobacco II, 46 Cal.4th at 312. “This distinction reflects the UCL's focus on the defendant's conduct, rather than the plaintiff's damages.” Id. Accordingly, to the extent Google argues that they are entitled to summary judgment because RMC cannot show that its reliance was reasonable, Google's argument mistakes the standard for what is required for a UCL fraud claim.¹⁰

10 This conclusion only pertains to Google's argument to a “reasonable reliance” requirement and is not to be confused with the objective “reasonable person” standard that courts use to assess likelihood of deception under the UCL. See, e.g., Friedman v. AARP, Inc., 855 F.3d 1047, 1055 (9th Cir. 2017) (“We assess likelihood of deception under a ‘reasonable consumer standard.'”).

That said, even if the Court were to construe Google's argument to be challenging the evidence under the correct UCL standard-i.e., whether “members of the public are likely to be deceived”-there are nonetheless genuine issues of material fact as to whether a reasonable advertiser would have likely been deceived by the LT Settings Screen Statement. As Judge Freeman recently remarked in Roley v. Google, courts should “bear[] in mind that whether reliance on a misrepresentation was reasonable is a question of fact for the jury, and may be decided as a matter of law only if the facts permit reasonable minds to come to just one conclusion.” Roley v. Google LLC, 2021 WL 1091917, at *8 (N.D. Cal. Mar. 22, 2021) (internal quotation marks and brackets omitted) (emphasis added). In other words, for Google to prevail on its motion for summary judgment, it must demonstrate that the evidence could only support its conclusion that Plaintiffs' reliance on the LT Settings Screen statement was unreasonable. Such a conclusion, however, is not supported by the evidence submitted.

Google contends that Help Center disclosures rendered the LT Settings Screen Statement unlikely to deceive, but RMC has presented evidence to substantiate at least two relevant points that rebut Google's contention. First, in addition to the plain language of the LT Settings Screen Statement, Google had made other statements suggesting Location Targeting would appear only to users within the selected areas and regions. See MSJ Opp. 18; Kaplan Decl., Ex. 48, ECF No. 622-52) (“For each Adwords campaign, [advertisers] can select the countries or regions . . . for your ad. That campaign's ads will appear only to users located in those areas.”) (emphasis added); Kaplan Decl., Ex. 50, ECF No. 623-52 (“Geographic location is the physical area where you want your ads to show. When setting up a geographic location, remember to think about your target audience and where you can best find them. For instance, if you are a local wedding planner operating in San Francisco, California, it's probably not a good idea to target users all over the United States.”) (emphasis added). These additional statements counterbalance any curative effect from Help Center disclosures, undermining Google's ability to prevail on summary judgment.

Second, RMC has also presented evidence that suggests an appreciable number of reasonably prudent advertisers did, in fact, rely on and believe that Location Targeting would yield clicks only within the selected geographic area. See, e.g., Kaplan Decl. Ex. 51 at 10, ECF No. 622-54 (describing a help ticket from an advertiser complaining, “As recommended by Google, I have drawn very specific boundaries around the area in which I want my ads to appear (Hawaii only). In the last month, there have been 80 clicks outside this area and I'm being charged for those clicks against my will.”) (emphasis added); Kaplan Decl., Ex. 53, ECF No. 622-56 (remarking that the fact that “a user in Ohio searching on ‘widgets California' will see a California-targeted ad . . . is a total surprise to many advertisers”) (emphasis added); Kaplan Decl., Ex. 55, ECF No. 622-58 (commenting that a Google employee “constantly hear[s] from advertisers that they weren't aware of query parsing”); Kaplan Decl., Ex. 57, ECF No. 622-60 (highlighting “two articles that [Google Analytics specialists] said led them to believe that [Google] always use[s] IP addresses to derive the data for the geographic performance report,” as opposed to also deriving data from query parsing) (emphasis added).

Google is not entitled to summary judgment on their reasonable reliance argument because it is divorced from the legal standard for a UCL fraud claim. Even construing Google's argument to be directed towards the correct standard, RMC's proffered evidence indicates a genuine and robust dispute over the deceptiveness of the statement at issue here.

2. Materiality

In addition to its reasonable reliance argument, Google also contends that the Location Targeting claim fails because Plaintiffs cannot show the misrepresentations were material; that is, “RMC cannot show that it based its decision to sign up with AdWords, in whole or in part, on this alleged misrepresentation.” MSJ 20. Specifically, Google advances the notion that “[n]o reasonable advertiser would throw out more than 95% of its advertising campaign because of a purported issue with the remaining 5%.” Id. at 19. RMC rejects this argument as improperly contorting its claim, which, “rightfully framed, is that had it known that Google would charge for clicks from outside its four-state advertising zone, it would not have paid for those out-of-area clicks or at the very least would have paid far less for those clicks.” MSJ Opp. 22 (internal parenthesis omitted).

At the outset, the Court is not persuaded by Google's suggestion that a failure to show materiality would preclude Plaintiffs' UCL fraud claim. As noted above, all that is required for RMC to prevail on its UCL claim is that “members of the public are likely to be deceived” by the misrepresentation. Tobacco II, 46 Cal.4th at 312. The California Supreme Court has indicated that materiality is relevant only insofar as it may give rise to a presumption of actual reliance by the named plaintiffs.¹¹ Id. at 327; see also Engalla v. Permanente Med. Grp., Inc., 15 Cal.4th 951, 977 (1997) (holding that plaintiffs “need only make a showing that the misrepresentations were material, and that therefore a reasonable trier of fact could infer reliance from such misrepresentations, in order to survive this summary-judgment-like proceeding”). In other words, if a statement is sufficiently important (i.e., material), courts may assume or infer that a reasonable consumer would rely on that statement. However, it does not follow that materiality is an element of a UCL fraud claim-it is simply one method by which a plaintiff can satisfy actual reliance. Google's citation In re iPhone Application Litigation is inapposite. MSJ 19-20 (citing In re iPhone Application Litig., 6 F.Supp.3d 1004 (N.D. Cal. 2013)). Nowhere in In re iPhone Application does Judge Koh suggest that a misrepresentation must cross some materiality threshold to survive summary judgment; indeed, materiality was asserted by plaintiffs in that case to support an inference of reliance. Id. at 1026. Although Judge Koh did comment that there was no evidence the plaintiff had “based her decision to obtain an iPhone, in whole or in part,” on the alleged misrepresentation, this remark was not referencing the materiality of the misrepresentation but rather the fact that the plaintiff had never seen the misrepresentation before obtaining her iPhone. Id. at 1021. Accordingly, to the extent Google suggests that a failure to show materiality would be fatal to RMC's claim, that suggestion is not supported by California law.

11 Notably, the California Supreme Court has interpreted the UCL statute to only “mean that named plaintiffs, but not absent ones, must show proof of ‘actual reliance' at the certification stage.” Walker v. Life Ins. Co. of the Sw., 953 F.3d 624, 630 (9th Cir. 2020) (emphasis added).

Even if the Court were to entertain Google's materiality argument, Google focuses on incorrect measures of materiality. “A misrepresentation is judged to be ‘material' if ‘a reasonable man would attach importance to its existence or nonexistence in determining his choice of action in the transaction in question.” Tobacco II, 46 Cal.4th at 327. Accordingly, a proper materiality assessment in this case should evaluate whether a reasonable advertiser would “attach importance” to Google's representation that the selected cities and regions in the LT Settings Screen would be “where [the advertiser would] like [its] ad to appear.” MSJ Opp. 6. Google's argument, however, misses the mark by focusing instead on after-the-fact data that only 5% of RMC's clicks originated outside of its selected areas. MSJ 19-20. The inherent materiality of Google's misrepresentations cannot turn on facts and data that are only available after a reasonable advertiser has decided to advertise with Google.

The Court further notes that “materiality is generally a question of fact unless the ‘fact misrepresented is so obviously unimportant that the jury could not reasonably find that a reasonable man would have been influenced by it.'” Tobacco II, 46 Cal.4th at 327 (emphasis added). As with Google's arguments regarding reasonable reliance, the Court finds that PRMC has presented sufficient evidence to create, at the least, genuine issues of fact as to whether the LT Settings Screen statement would be material to a reasonable advertiser. See supra Section III.B.1 (citing evidence that advertisers were surprised by charges for out-of-area clicks).

Accordingly, Google is not entitled to summary judgment based on materiality.

3. Injury

Finally, Google argues that summary judgment is warranted because RMC has received the “benefit of the bargain” and, therefore, has suffered no injury under the UCL. MSJ 20-21. Specifically, because the out-of-area clicks nonetheless account for a potential customer's engagement with RMC's advertisement, RMC has received what it paid for, i.e., “convert[ing] an impression-the consumer's viewing of the ad-into a valuable interaction with the consumer.” Id. at 21. RMC again rejects Google's characterization, asserting that it did not want out-of-area clicks when it agreed to advertise with Google and that it receives no benefit from engagement with out-of-area consumers. MSJ Opp. 23-24.

Google's “benefit of the bargain” argument has no support under California law. In Kwikset Corp. v. Superior Ct., 51 Cal.4th 310, 332 (2011), the California Supreme Court considered a similar “benefit of the bargain” theory where plaintiffs purchased locks advertised as “Made in U.S.A.” and received locks that were not made in the United States. Similar to Google's contentions here, defendant Kwikset had argued that “consumers who receive a fully functioning product have received the benefit of their bargain, even if the product label contains misrepresentations that may have been relied upon by a particular class of consumers.” Id. The California Supreme Court rejected this argument, reasoning that “the observant Jew who purchases food represented to be, but not in fact, kosher; the Muslim who purchases food represented to be, but not in fact, halal; the parent who purchases food for his or her child represented to be, but not in fact, organic, has in each instance not received the benefit of his or her bargain.” Id. (emphasis added). Likewise, RMC-who paid for clicks represented to originate from Florida, Georgia, Louisiana, and North Carolina, but in fact originated from states such as New York, Texas, and California-did not receive the benefit of its bargain and, therefore, suffered economic injury in the amount of the extra money it paid for those clicks. Id. at 330.

Google again relies on inapposite California case law to support its theory. MSJ 21 (citing Animal Legal Def. Fund v. Mendes, 160 Cal.App.4th 136 (2008)). In Animal Legal Defense Fund, the California Court of Appeal rejected a theory of injury from plaintiffs who purchased dairy products that they would otherwise not have purchased if they knew that some of the cows were raised in cruel conditions. Id. However, Animal Legal Defense Fund was a California Court of Appeal opinion that pre-dated the California Supreme Court's decision in Kwikset; accordingly, to the extent that these two cases may be in tension with each other (the Court does not believe they are), the Court follows the state high court's holding in Kwikset. Cf. Hinojos v. Kohl's Corp., 718 F.3d 1098, 1107 (9th Cir. 2013) (noting that the “‘benefit of the bargain' rationale was explicitly rejected in Kwikset”). More critically, the dairy advertising in Animal Legal Defense Fund did not contain any affirmative representations regarding the conditions in which it raised its cows and, therefore, those conditions could not be considered part of the purchase agreement. Id. at 146-47 (“Any assumptions regarding treatment of the dairies' cows were not alleged to have been expressed by the consumers to anyone, including the retailers, much less the dairy owners and their contractors.”). By contrast, Google affirmatively represented to its advertisers on the LT Settings Screen that the advertisers' selected regions would be where their ads would appear. Google cannot now claim advertisers nonetheless received the “benefit of the bargain” when they paid for clicks from ads that were shown to users outside of their selected regions.

In sum, the Court finds that Google's motion for summary judgment on RMC's Location Targeting UCL claim is not meritorious. Google has challenged RMC's evidence on elements that are largely not required under California law for UCL fraud claims, such as reasonable (as opposed to actual) reliance and materiality. In any event, the Court finds that RMC has identified genuine issues of fact and introduced sufficient evidence to survive summary judgment. Accordingly, Google's motion for summary judgment on the Location Targeting UCL claim is DENIED.

IV. GOOGLE'S MOTION TO STRIKE PLAINTIFFS' EXPERT

The Court has previously decided several Daubert motions in connection with class certification. It granted in part and denied in part Plaintiffs' motions, striking the report of Professor Ronald Wilcox as untimely but refusing to strike the various expert reports and declarations of Dr. Lawrence Wu. ECF Nos. 654, 659. As for Google's motion, the Court declined to strike the opinions of Saul Solomon regarding proposed damages models for Plaintiffs' Location Targeting claims, but it deferred ruling on Google's objections to Dr. James L. Gibson and to Solomon's opinions regarding proposed damages models for Plaintiffs' Smart Pricing claims. ECF No. 661. The Court now decides those deferred issues.

A. Dr. James L. Gibson

Google raises three objections to the opinions of Dr. Gibson: (1) he is not qualified to offer an expert opinion; (2) his methodology is unreliable; and (3) the sample he relies on to form his opinion is too small. MTS 6-11.

1. Qualifications

Google first argues that Dr. Gibson is unqualified to opine on the advertising data at issue because his expertise is in political science, not big data or e-commerce. MTS 6. Plaintiffs counter that Dr. Gibson has ample experience conducting empirical analyses of large datasets, enabling him to opine here even if his primary field of study is not in digital advertising. Pls.' Opp. to MTS (“MTS Opp.”) 4-5, ECF No. 455.

Ninth Circuit precedent “contemplates a broad conception of expert qualifications.” Hangarter v. Provident Life Ins. & Accident Ins. Co., 373 F.3d 998, 1015 (9th Cir. 2004) (citation and internal quotation marks omitted). Heeding this principle, courts in this district have routinely held that an expert need not have the best qualifications in her field to survive a Daubert challenge. E.g., Zeiger v. WellPet LLC, 526 F.Supp.3d 652, 669 (N.D. Cal. 2021) (“Daubert does not require that the expert with the best possible qualifications testify ....”); United States v. Pac. Gas & Elec. Co., 2016 WL 3268994, at *2 (N.D. Cal. June 15, 2016) (“[E]xperts are not required to be the best or only individuals qualified in their fields.”); Regal Elecs., Inc. v. Pulse Eng'g, Inc., 2005 WL 6019703, at *2 (N.D. Cal. Nov. 14, 2005) (“It is not the province of the Court to select the best or most particularly qualified experts for the parties.”). They have also explained that “courts should not exclude expert testimony simply because . . . the proposed expert does not have the specialization that the court considers most appropriate.” In re Pac. Fertility Ctr. Litig., 2021 WL 842739, at *3 (N.D. Cal. Mar. 5, 2021) (quoting Schroeder v. Cnty. of San Bernardino, 2019 WL 3037923, at *3 (C.D. Cal. May 7, 2019). Thus, all Plaintiffs must do to qualify Dr. Gibson as an expert is to show that he has sufficient qualifications. Zeiger, 526 F.Supp.3d at 669. The Court concludes that they have done so.

Dr. Gibson has extensive experience with the empirical analysis of data. Throughout his career, he has published numerous books and articles based on empirical analysis, many of which were published in prominent political science journals. Mazzeo Decl., Ex. 23 (“7/9/15 Gibson Rep.”) ¶¶ 2-3 & Apps. A-B, ECF No. 424-31; Mustokoff Decl., Ex. 2 (“6/14/18 Gibson Decl.”) ¶¶ 2-4, ECF No. 454-8. He is also impeccably credentialed, having earned a PhD in political science, received grants from the National Science Foundation for his empirical work, and won several awards for his political science research. 7/9/15 Gibson Rep. ¶ 3; 6/14/18 Gibson Decl. ¶ 3.

Google does not directly challenge Dr. Gibson's considerable background in empirical political science. Instead, it contends that Dr. Gibson's political science credentials are not the right qualifications for this case. However, the fact that Dr. Gibson's training and research are in political science does not prevent him from serving as an expert here. At core, the opinions provided by Dr. Gibson are simply about identification: identifying clicks that had not been Smart Priced, identifying types of clicks, and identifying clicks from outside of designated geographical areas. See 7/9/15 Gibson Rep. Navigating datasets to perform these tasks requires expertise in data analysis, something Dr. Gibson unquestionably possesses. But contrary to Google's assertions, such identification does not require any specialized knowledge about e-commerce, especially since the identifying criteria were supplied by Google or derived from Plaintiffs' theories of liability. Id. ¶¶ 17-19, 22, 25, 27, 32, 38, 42, 47, 49. While a data scientist with a background in e-commerce may or may not be more qualified to testify about the claims at issue than Dr. Gibson, the fact that a proposed expert may not have the most appropriate specialization is not grounds for exclusion. Pac. Fertility Ctr., 2021 WL 842739, at *3; see also Zeiger, 526 F.Supp.3d at 669 (permitting an animal nutritionist to opine on safe levels of certain chemicals in pet food over objections that only a toxicologist would be qualified to opine).

The suggestion that Dr. Gibson is unqualified because he lacks experience with big data similarly fails to persuade. Dr. Gibson has experience working with datasets containing millions of datapoints. Evans Decl., Ex. 1 (“3/5/18 Gibson Dep.”) at 182:10-183:2, ECF No. 441-2. Although the datasets at issue in this case are larger than the previous datasets he has worked with, Dr. Gibson explains that there are available tools that allow for the easy scaling of analyses from smaller datasets to very large ones. 6/14/18 Gibson Decl. ¶¶ 8-9.

Google does not dispute Dr. Gibson's prior experience, nor does it explain how larger datasets pose challenges that Dr. Gibson is supposedly incapable of handling. Rather, it points to two isolated passages from Dr. Gibson's depositions, neither of which support its argument. In the first passage, Google asked Dr. Gibson whether he considered himself to be “one of the leading experts in the analysis of big databases.” 3/5/18 Gibson Dep. at 143:8-9. Dr. Gibson responded that he did not, id. at 143:10, an answer which Google took to mean that he disclaimed any expertise in big data. MTS 6. But Dr. Gibson testified only that he was not a leading expert. He did not state that he had no expertise in large datasets, and the fact that he does not possess the best possible qualifications does not prevent the Court from admitting his expert opinions. See Zeiger, 526 F.Supp.3d at 669.

The second deposition passage demonstrates even less. There, Dr. Gibson explained how he would analyze an extremely large dataset of 500 billion clicks by using distributed computing and certain software tools. 3/5/18 Gibson Dep. at 129:2-131:19. From Google's perspective, the response was too general and indicative of Dr. Gibson's lack of expertise. MTS 7 (quoting 3/5/18 Gibson Dep. at 130:3-131:15). The Court disagrees. It is true that Dr. Gibson could have provided a more fulsome and detailed response. As he made clear though, he was answering “off the top of [his] head” and would need to “do a proper investigation to be able to really answer [the] question.” 3/5/18 Gibson Dep. at 130:15-20. The measure of Dr. Gibson's qualifications is not the nimbleness with which he can answer a deposition question; it is the totality of his education, experiences, and accomplishments. When compared against the body of his research and credentials, the generality of Dr. Gibson's extemporaneous response to a technical question carries little persuasive weight.

In sum, the Court finds that Dr. Gibson is fully qualified to offer his expert opinions, and Google's arguments do not convince the Court otherwise.

2. Methodology

Google also contends that Dr. Gibson failed to use a reliable methodology when forming his opinions. Specifically, it asserts: (1) that Dr. Gibson did not account for the strategic choices that advertisers could make on Google's AdWords platform, rendering his identification of search bundled campaigns unreliable; (2) that Dr. Gibson's method identifies (a) non-Smart Priced clicks without regard to whether they should have been Smart Priced, and (b) identifies AFMA and mGDN clicks without determining whether the correct Smart Pricing multiplier was applied to those clicks; (3) that Dr. Gibson relied on other unsupported assumptions; and (4) that Dr. Gibson's opinions should be excluded to the extent they are based on a data sample that Dr. Gibson heavily criticized.¹² MTS 9-10. Plaintiffs respond that Dr. Gibson's methodology is rooted in record evidence and their theory of liability in the case. MTS Opp. 9-10. They also argue that Dr. Gibson's criticism of the representativeness of a data sample provided by Google does not undermine his opinions about whether putative class members can be identified using Google's data. Id. The Court addresses each of Google's methodology arguments in sequence.

12 Google also argues that Dr. Gibson's approach is not a common methodology but instead requires repeated individual analyses. MTS 10. Such argument goes to the substance of class certification requirements and have no bearing on the reliability of Dr. Gibson's methods.

First, the Court finds that Dr. Gibson's method for identifying search-bundled clicks is reliable because he expressly developed his method based on the deposition testimony of Google employees. 7/9/15 Gibson Rep. ¶¶ 22, 25, 27.

Second, Dr. Gibson did not need to develop methods for identifying whether clicks should have been Smart Priced or whether AFMA or mGDN clicks should have been Smart Priced at a different level. Despite Google framing the lack of such methods as an issue of reliability, its arguments are in fact disputes about the merits of Plaintiffs' claims, i.e., whether and how Google's contracts require Google to Smart Price clicks. If Google is correct, that would mean Plaintiffs' theories of liability fail, not that Dr. Gibson's methods are unreliable. Raising merits arguments in Daubert briefing is “wholly inappropriate,” and such arguments are not grounds for exclusion of expert testimony. Brown v. Google, LLC, 2022 WL 17961497, at *4 n.4 (N.D. Cal. Dec. 12, 2022). It is enough for purposes of Daubert that Dr. Gibson's methodology is “consistent with plaintiffs' class definition.” Id. at *5. Plaintiffs' class definition sweeps in all non-Smart Priced clicks, and that is what Dr. Gibson's methodology accomplishes. See CC Mot. 16. Likewise, Dr. Gibsons methodology accords with Plaintiffs' definition of AFMA and mGDN clicks. Compare 7/9/15 Gibson Rep. ¶¶ 38-41 (AFMA click methodology) and ¶¶ 42-45 (mGDN click methodology), with CC Mot. 7-8, 16 (class definition for AFMA and mGDN clicks). Whether or not Dr. Gibson's opinions will ultimately support class certification, the fact that Dr. Gibson's methodology tracks with Plaintiffs' class definitions is sufficient for his opinions to survive Google's Daubert challenge to his methodology.

Third, Dr. Gibson's opinions are not based on unsupported assumptions. Google takes issue with his purported assumption at deposition that Google's databases contain a variable specifying whether an advertiser turned on the Autobid feature in the AdWords platform. MTS 8. Yet, Google does not identify any point in Dr. Gibson's reports or declarations where he relies on that assumption to form his opinions, and to the extent Dr. Gibson sought to account for Autobid in his analysis, he did so by looking to advertisers' change history logs in accordance with testimony from Google employees. 7/9/15 Gibson Rep. ¶¶ 22-24. An assumption that is not relied upon by Dr. Gibson is plainly not a ground for excluding his opinions. In any case, “arguments that [an expert] based on his opinions on an improper assumption go to the weight, not the admissibility of the testimony.” Perez v. Rash Curtis & Assocs., 2019 WL 1491694, at *3 (N.D. Cal. Apr. 4, 2019).

Lastly, Dr. Gibson's criticisms of data samples that Google created do not make his opinions unreliable to the extent that he bases his opinions on those samples. Dr. Gibson criticized the construction and representativeness of Google's data samples. Mustokoff Decl., Ex. 3 (“12/20/18 Gibson Decl.”) ¶¶ 23-24, ECF No. 454-9. His opinions, though, relate to whether the underlying data stored by Google could be used to identify putative class members without individualized analysis. Those opinions concern the nature of the data held by Google and are separate and apart from Dr. Gibson's critiques about how Google used the data to create samples. Thus, Dr. Gibson's critiques of Google's data samples do not undermine his opinions.

3. Sample Size

As a final argument, Google asserts that the data samples that Dr. Gibson relied on were too small, claiming that he reviewed only the advertising click data for Cabrera and former plaintiff Woods. MTS 11. However, Google is factually mistaken. Dr. Gibson also reviewed data from over 400 million clicks that Google had sampled, confirming the conclusions that he drew from Cabrera's and Woods' data. 3/15/18 Gibson Dep. at 112:24-113:14. These millions of clicks are a sufficient sample size to support Dr. Gibson's opinions.

* * *

To conclude, the Court finds that Dr. Gibson is qualified under Daubert and has formed his opinions based on sufficient facts and reliable methods. The Court therefore DENIES Google's motion to strike Dr. Gibson's reports.

B. Saul Solomon

Google advances a litany of objections against Solomon's Smart Pricing opinions, but those challenges can be distilled down to two points: (1) that Solomon's proposed damages model improperly calculates damages based on the cost of clicks rather than accounting for the dynamics of AdWords auctions; and (2) that Solomon considered only a small sample of data-the clicks for Woods and Cabrera-without testing his methodology on any other class members. MTS 12-16, 23-24. Plaintiffs reply that Solomon's approach is consistent with Google's external representations and internal calculations, and that damages need only be a reasonable approximation based on the best evidence available. MTS Opp. 11-16. They also argue that Google's sample size objections ring hollow because Solomon relied on Dr. Gibson's analysis of the 400 million clicks from Google's data samples. Id. at 18.

The Court begins by observing that Google's argument about the importance of auction dynamics is, once again, a dispute about the merits of Plaintiffs' claims. Plaintiffs' theory of liability is that Smart Pricing should have been applied to the post-auction cost of a click. See MTS Opp. 12-13; FAC ¶¶ 26-29, 31 (allegations that Smart Pricing applies to the final cost or price of a click); Evans Decl., Ex. 11 (“7/10/15 Solomon Rep.”) at 9-12, ECF No. 441-10 (discussing public statements by Google that describe Smart Pricing as applying to the final price or cost of a click). Google may disagree with Plaintiffs' theory of liability, but that does not make Solomon's opinions unreliable. For Solomon's opinions to survive Google's objection, they need only be consistent with Plaintiffs' theory. Brown, 2022 WL 17961497, at *5. That is the case here, where Solomon's “overcharge” model calculates damages by applying click cost multipliers to the final click costs.¹³ 7/10/15 Solomon Rep. at 29-34.

13 Google does not challenge Solomon's alternate model of Smart Pricing damages, which he terms as “net profits.”

As for Google's second objection, the Court has previously rejected that argument in its order regarding Solomon's Location Targeting opinions. 9/26/22 Order 6, ECF No. 661. Google's argument is no more persuasive in the Smart Pricing context, and the Court rejects it for the same reasons articulated in its prior order. Solomon relies on the opinions of Dr. Gibson, who analyzed large data samples, and Solomon is not required to perform full calculations of classwide damages at this stage. Id. That is enough to satisfy Daubert.

The Court DENIES Google's motion to strike Solomon's Smart Pricing opinions.

V. PLAINTIFF'S MOTION FOR CLASS CERTIFICATION

A. Smart Pricing Class

Plaintiffs proposed the following definition for the Smart Pricing Class:

All persons and entities located within the United States who, between August 22, 2006 and February 20, 2013, advertised through Google's AdWords program and paid for clicks on their Google AdWords advertisement(s), where such clicks were not Smart Priced because they (1) originated from a property on Google's Display Network and Google applied no Smart Pricing measurement, or (2) were AFMA Clicks, Search Bundled Clicks, or mGDN Clicks.

CC Mot. 16 & n.4. In turn, Plaintiffs define the named clicks, as follows:

AFMA Clicks : “[C]licks on ads from mobile applications on the Display Network for which a click cost multiplier greater than 0.08 was applied, between August 22, 2006 and February 20, 2013.”

Search Bundled Clicks : “[C]licks on ads on the Display Network where the advertiser's settings allowed its ads to show on both the Search and Display Networks and did not set a Display Network bid different from the Search Network bid, between June 1, 2009 and December 13, 2012.”

mGDN Clicks : “[C]licks on ads on properties of a desktop publisher (i.e., desktop websites) on the Display Network from high end mobile devices (e.g., smart phones or tablets), between August 22, 2006 and November 6, 2011.”

CC Mot. 7-8.

While Plaintiffs present the Smart Pricing Class as a single, uniform class, each of the four different categories of clicks-(1) Non-Smart Priced Clicks, (2) AFMA Clicks, (3) Search Bundled Clicks, and (4) mGDN Clicks-reflect different theories of liability based on different factual arguments. As such, the Court treats the Smart Pricing Class as four subclasses, referring to them as the “Non-Smart Priced Clicks Subclass,” the “AFMA Clicks Subclass,” the “Search Bundled Clicks Subclass,” and the “mGDN Clicks Subclass.” Where the criteria for class certification can be analyzed consistently across all four subclasses, the Court will not distinguish between them in its discussion below. But where there are relevant differences for class certification, the Court will specify which subclass it is analyzing.

1. Rule 23(a) Requirements

a. Numerosity

Plaintiffs have shown that their proposed class meets the numerosity requirement. Although there is no exact numerical cut-off for when a proposed class is “so numerous that joinder of all the class members is impracticable,” courts often will presume that this requirement is satisfied when the class exceeds forty members. Fed. R. Civ. P.23(a)(1); see, e.g., Arroyo v. Int'l Paper Co., 2019 WL 1508457, at *2 (N.D. Cal. Apr. 4, 2019). In this case, Google has admitted that there are more than 50 AdWords advertisers falling in each of the four subclasses. Mazzeo Decl., Ex. 36 at 17, ECF No. 426-37 (Request for Admission Nos. 1-4). As such, Plaintiffs have established numerosity.

b. Commonality

The Court finds that Plaintiffs have demonstrated commonality as well. Commonality requires there to be “questions of law or fact common to the class.” Fed.R.Civ.P. 23(a)(2). A common question is one that “is capable of classwide resolution,” meaning that answering the question “will resolve an issue that is central to the validity of each one of the claims in one stroke.” Wal-Mart, 564 U.S. at 350. The key factor is whether the common question generates “common answers.” Id. However, Plaintiffs need not show that all questions are common; they need only identify a “single significant question of law or fact.” Stockwell v. City & Cnty. of San Francisco, 749 F.3d 1107, 1111 (9th Cir. 2014).

Here, Plaintiffs propose that the interpretation of the AdWords Agreement is a common question because it is a form contract that bound all putative class members during the class period. CC Mot. 19. Google does not directly contest commonality, but in its arguments against predominance, it contends that interpreting the AdWords Agreement does not present a common question because the endeavor requires extrinsic evidence, and such extrinsic evidence is not common to the class. CC Opp. 10. In its view, although Plaintiffs' proffered extrinsic evidence consists of public statements, each putative class member would have been exposed to different statements. Id. at 12-13. As a consequence, it would be necessary to conduct individualized investigations to determine which pieces of extrinsic information each putative class member did or did not view. Id.

While Google's argument might hold merit in the context of bargained-for contracts, when interpreting standardized form contracts like the AdWords Agreement, the subjective understanding of the signatories is not a factor. Contrary to Google's assertion that Plaintiffs' breach of contract claim requires numerous individual determinations, it is an “axiom of contract law[] that when there is a standardized agreement like the form contract at issue in this case, the agreement is interpreted wherever reasonable as treating alike all those similarly situated, without regard to their knowledge or understanding of the standard terms of the writing.” Williams v. Apple, Inc., 338 F.R.D. 629, 638 (N.D. Cal. 2021) (alteration in original) (internal quotations omitted) (quoting Bally v. State Farm Life Ins. Co., 335 F.R.D. 288, 302 (N.D. Cal. 2020)). Indeed, California courts have long recognized this axiom and have adopted an objective standard for interpreting form contracts. Rather than inquiring into a party's mindset or knowledge, “courts interpret form contracts to mean what a reasonable buyer would expect them to mean.” Emps. Cas. Ins. Co. v. Foust, 29 Cal.App.3d 382, 386 (1972); see also Gray v. Zurich Ins. Co., 65 Cal. 2d 263, 271 (1966) (similar). The Restatement, a source “entitled to great consideration as an argumentative authority” under California law, Lake Almanor Associates, 178 Cal.App.4th at 1200 (citation omitted), says much the same. It provides that standardized agreements should be “interpreted wherever reasonable as treating alike all those similarly situated, without regard to their knowledge or understanding of the standard terms of the writing.” Restatement (Second) of Contracts § 211(2). And it also directs courts to apply an objective standard by “seek[ing] to effectuate the reasonable expectations of the average member of the public” when construing and applying a standardized contract. Id. § 211 cmt. e (emphasis added). It is for these reasons that “claims arising from interpretations of a form contract appear to present the classic case for treatment as a class action.” Kleiner v. First Nat'l Bank of Atl., 97 F.R.D. 683, 691 (N.D. Cal. 1983); see also McGhee v. Bank of Am., 60 Cal.App.3d 442, 449 (1976) (“Controversies involving widely used contracts of adhesion present ideal cases for class adjudication ....”).

Despite these well-established maxims, determining how to interpret the AdWords Agreement is complicated by the fact that the Court previously found the relevant clause of the AdWords Agreement to be ambiguous, 8/24/12 Order 14, making it necessary to consider extrinsic evidence. Brown v. Goldstein, 34 Cal.App. 5th 418, 432 (2019) (“Extrinsic evidence is admissible . . . to interpret an agreement when a material term is ambiguous.”) (citation omitted). That is so because the usual approach to interpreting an ambiguous clause with extrinsic evidence is to evaluate how that evidence bears on “the parties' understanding and [the] intended meaning of the words used in their written agreement.” In re Facebook PPC Advertising Litig., 709 F.Supp.2d 762, 769 (N.D. Cal. 2010) (quoting Brawthen v. H & R Block, Inc., 28 Cal.App.3d 131, 136 (1972)) (emphasis removed). This subjective approach is in tension with the objective approach typically used to interpret form contracts.

Google suggests that the differing approaches can be reconciled by recognizing that they are deployed in different scenarios. It submits that, if a contract term is not ambiguous, the objective approach applies because no extrinsic evidence is needed, and the term can be interpreted from the lens of a reasonable person. But if the term is ambiguous and extrinsic evidence is necessary, Google maintains that courts must use extrinsic evidence to probe the signatories' subjective understandings, and that, in turn, requires individual investigation into what extrinsic information a person was exposed to. CC Opp. 12.

Google's position finds some support in the case law. For example, in Monaco v. Bear Stearns Co., 2012 WL 10006987, at *7-8 (C.D. Cal. Dec. 10, 2012), the court found that interpretation of a form contract could not be resolved on a classwide basis. Acknowledging the plaintiffs' argument that form contracts normally should be uniformly interpreted, Monaco concluded that such interpretive rule did not apply when the relevant terms of the form contract were ambiguous and required reference to extrinsic evidence of intent. Id. at *6-7. Likewise, Abbit v. ING USA Annuity, 2015 WL 7272220, at *9 (S.D. Cal. Nov. 16, 2015), held that interpretation of standard contract language could not be answered with common evidence when it “require[d] extrinsic evidence and individualized determinations as to [class members'] knowledge and understanding of this contract language.” Google's out-of-circuit authority lends support as well. See Krueger v. Nw. Mut. Life Ins. Co., 2011 WL 2938273, at *6 (N.D. Fla. July 21, 2011) (“Even the most common of contractual questions-those arising, for example, from the alleged breach of a form contract do not guarantee predominance if individualized extrinsic evidence bears heavily on the interpretation of the class members' agreements.”) (citation omitted); Adams v. Kansas City Life Ins. Co., 192 F.R.D. 274, 282 (W.D. Mo. 2000) (“By allowing extrinsic evidence of the parties' dealings, the breach of contract claims become individualized and not reasonably susceptible to class action treatment.”).

Other courts, however, have reached the opposite conclusion. In Rodman v. Safeway, Inc. (Rodman I), 2014 WL 988992, at *8 (N.D. Cal. Mar. 10, 2014), the court concluded that, even though it did not find the contract at issue to be ambiguous, if it needed to consider extrinsic evidence to resolve ambiguity, it could do so “without endangering predominance.” The court in Vedachalam v. Tata Consultancy Servs., Ltd., 2012 WL 1110004, at *15 (N.D. Cal. Apr. 2, 2012), considered the question even more directly, explaining that extrinsic evidence “[is] susceptible to proof on a class-wide basis.” There, the extrinsic evidence consisted of common documents that some class members signed but others did not. Id. Nonetheless, and consistent with the admonition not to consider a party's knowledge or understanding when interpreting a form contract, the fact that class members interacted with the extrinsic evidence differently did not defeat class certification. And in Menagerie Productions v. Citysearch, 2009 WL 3770668, at *10 (C.D. Cal. Nov. 9, 2009), the court determined that a form contract could be interpreted on a classwide basis if the relevant extrinsic evidence, “such as representations on [a defendant's] website,” were also available on a classwide basis.

The upshot of the Court's survey is that there is no clear answer as to whether courts interpreting form contracts with the assistance of extrinsic evidence must necessarily make individualized determinations about what extrinsic evidence a particular party knew or was exposed to. On balance, though, the Court agrees with those decisions holding that individualized determinations are not necessarily required. Courts have consistently and unequivocally stated that form contracts should be interpreted without regard for an individual party's knowledge. See, e.g., Williams, 338 F.R.D. at 638; Gray, 65 Cal. 2d at 271. In doing so, they have not indicated that there is an exception when extrinsic evidence is considered. To ask about whether a class member read or was exposed to a particular piece of extrinsic evidence, then, would run counter to this well-established canon of construction. See Restatement (Second) of Contracts § 211 cmt. b (noting that customers signing form contracts are bound to the contracts' terms even though they “do not in fact ordinarily understand or even read the standard terms”) (emphasis added).

In fact, a careful reading of Google's case law to the contrary reveals that they are consistent with this approach. The Restatement explains that form contracts should be interpreted uniformly for all those “similarly situated.” Restatement (Second) of Contracts § 211(2). However, in Monaco, different class members were told different information about their loan agreements by different brokers employed by the defendant. 2012 WL 10006987, at *6. So too in the remainder of Google's authorities. See Abbit, 2015 WL 7272220, at *2 (“[A]nnuities [were] sold by a variety of individuals and organizations, including independent agents, retail brokerdealers, marketing organizations, and banks.”); Kreuger, 2011 WL 2938273, at *5 (“[C]lass members purchased annuities from various sales agents across the country.”); Adams, 192 F.R.D. at 278-79 (noting that the insurance policies at issue were sold over the course of “hundreds of individual meetings between agents and customers” where the agents “did not utilize a uniform sales script” and instead varied their presentations “depending on the particular concerns expressed by particular customers”). Thus, in the cases cited by Google, the circumstances of contract formation differed from class member to class member, rendering them differently situated. Put another way, the very universe of extrinsic evidence available differed from class member to class member because the representations made by one agent to one class member would not have been available to any other class member. That scenario is in marked contrast to the facts in this case, where the same universe of extrinsic evidence-publicly available documents and statements from Google's website (see FAC, Exs. B-G; Mazzeo Decl., Exs. 8, 10-11, ECF Nos. 426-9, -11 to -12¹⁴)-was available to all putative class members. Indeed, it is on this basis that the Ninth Circuit affirmed certification in Rodman, because “[n]one of the offered extrinsic evidence . . . went to individual conversations and representations made to individual class members.”¹⁵ Rodman v. Safeway (Rodman II), 694 Fed.Appx. 612, 614 (9th Cir. 2017).

14 Exhibit 11 to the Mazzeo Declaration appears to be the same document as Exhibit 63 to the Kaplan Declaration. Like with Exhibit 63, the Court here considers only those excerpts of Exhibit 11 for which URLs are given. See supra note 4.

15 Google briefly argues that the circumstances of contract formation did differ between class members because some would have relied on advertising agencies. CC Opp. 12-13; Suppl. CC Opp. 20. However, for the reasons the Court discusses in addressing the same argument in the context of the Location Targeting Class, the Court finds that Google has failed to substantiate its argument with appropriate evidence. See infra Section V.B.2.a.iii. This argument is therefore not a basis for denying class certification.

In conclusion, the Court finds that interpretation of the AdWords Agreement presents common questions. The Agreement may be interpreted without individualized inquiry into extrinsic evidence by considering how a reasonable person, standing in the shoes of an advertiser, would interpret the Agreement given the extrinsic evidence available to her.

c. Typicality

To satisfy Rule 23's typicality requirement, class representatives must show that their claims are “reasonably coextensive with those of absent class members,” though their claims “need not be substantially identical.” Just Film, Inc. v. Buono, 847 F.3d 1108, 1116 (9th Cir. 2017). But when “there is a danger that absent class members will suffer [because] their representative is preoccupied with defenses unique to it,” a class representative is not typical. Hanon v. Dataprods. Corp., 976 F.2d 497, 508 (9th Cir. 1992) (citation omitted).

The Court finds that Plaintiffs' Smart Pricing claims are reasonably coextensive with those of absent class members. Their claims, like the claims of all class members, arise from the same contract (the AdWords Agreement) and the same conduct by Google (alleged failure to properly Smart Price). Google suggests that Cabrera's claim is dissimilar to those of class members because Cabrera did not rely on any extrinsic evidence before signing the AdWords Agreement while some class members did rely on such evidence. CC Opp. 22. But as discussed in its commonality analysis, the Court already concluded that interpretation of a form contract like the AdWords Agreement does not depend on individual knowledge or reliance, so it rejects Google's typicality argument here on the same grounds.

Google also contends that Cabrera is atypical since he is subject to unique defenses. However, none of those potential defenses threatens to preoccupy Cabrera to the detriment of the class. Google first asserts that Cabrera lacks Article III standing to assert his Smart Pricing claim. CC Opp. 21. While this may have been a colorable argument when Google filed its opposition, the Ninth Circuit has since ruled that Cabrera has standing, so it is no longer a live issue in this matter. See 1/4/21 Ninth Cir. Mem. Second, Google reiterates its summary judgment argument that Cabrera has not suffered actual damage as required by California contract law. CC Opp. 22. As the Court explained when addressing this argument in the context of summary judgment, California law permits Cabrera, the promisee of the AdWords Agreement, to sue for damages suffered by RMC, the third-party beneficiary, under certain circumstances. Supra Section III.A.2. Although his right to do so depends on whether RMC is a creditor or donee beneficiary, the record is devoid of any evidence that RMC is a donee beneficiary such that Cabrera cannot sue for RMC's damages. To defeat typicality, Google “must show some degree of likelihood a unique defense will play a significant role at trial.” In re Sony Vaio Computer Notebook Trackpad Litig., 2013 WL 12116137, at *11 (S.D. Cal. Sept. 25, 2013) (quoting Beck v. Maximus, Inc., 457 F.3d 291, 300 (3d Cir. 2006)). It is not enough to argue that there might be a colorable unique defense against Cabrera when Google has offered no evidence supporting that defense. Thus, the Court cannot conclude that either of the purported unique defenses threatens to occupy Cabrera to such an extent that he would be atypical of the class.

Google also makes a final argument, this time with respect to both Cabrera and RMC, that variations in sophistication between advertisers preclude a finding of typicality. Suppl. CC Opp. 24. This argument is properly addressed as a challenge to predominance, so it does not defeat typicality.

Therefore, the Court finds that Plaintiffs are typical of the Smart Pricing Class.

d. Adequacy

Rule 23(a)'s final requirement seeks to ensure that the class representative would “fairly and adequately protect the interests of the class.” Fed.R.Civ.P. 23(a)(4). “To determine whether the representation meets this standard, we ask two questions: (1) Do the representative plaintiffs and their counsel have any conflicts of interest with other class members, and (2) will the representative plaintiffs and their counsel prosecute the action vigorously on behalf of the class?” Staton v. Boeing Co., 327 F.3d 938, 957 (9th Cir. 2003).

Google does not argue that either RMC or Cabrera have a conflict of interest with other class members, nor does it cast any doubt that Plaintiffs and Plaintiffs' counsel would prosecute the action vigorously on behalf of the class. Rather, Google argues that Plaintiffs are inadequate representatives because they purportedly destroyed evidence and made false testimony. Suppl. CC Opp. 4-7. Google points to deleted text messages between Cabrera and the successor of his previous company, Martin Peen, as well as Cabrera's testimony regarding his online advertising expertise and sophistication. Id. The Court addresses each in turn.¹⁶

16 Google does not challenge the adequacy of proposed class counsel, and the Court finds that counsel is adequate.

Spoliation : Google first addresses the deletion of text messages between Cabrera and Peen in which Peen declined Cabrera's request to assist in the litigation. Suppl. CC Opp. 4-5. These communications, Google contends, support the argument that Cabrera is not an adequate representative. Id. The Court disagrees. It finds that Google's spoliation contentions are not likely to play a significant role at trial, so they do not render Plaintiffs inadequate.

Although the Court does not have the benefit of full briefing on a spoliation motion and does not now issue any opinion on spoliation, it is at least satisfied that “the relative strength of the [spoliation] argument demonstrates that [it] will not play a significant role at trial such that it precludes certification.” Brust v. Regents of the Univ. of Cal., 2008 WL 11512299, at *6 (E.D. Cal. Oct. 24, 2008). In particular, sanctions for spoliation may not issue unless “a reasonable trier of fact could find that the destroyed evidence was relevant.” In re Arris Cable Modem Consumer Litig., 327 F.R.D. 334, 357 (N.D. Cal. 2018). Here, however, the deleted communications with Peen appear to have minimal relevance on the litigation, given that Peen had never used AdWords, would not be a class member, and ultimately agreed to assign Training Option's claim to Cabrera. Silverman Decl., Ex. 7 (“Peen Dep.”) 113:7-114:7.

Google's cited case authorities do not otherwise convince the Court that the spoliation issues raised here are substantial enough to disqualify Plaintiffs from serving as representatives. Even though courts have occasionally denied certification where class representatives faced spoliation allegations, the lost evidence in those cases were central to the plaintiffs' claims and often involved tangible and irreplaceable objects. See, e.g., Victorino v. FCA U.S. LLC, 2018 WL 2455432, at *8 (S.D. Cal. June 1, 2018) (finding plaintiff inadequate where spoliated evidence included cylinders and a connecting hose from the allegedly defective vehicle); Arris, 327 F.R.D. at 358 (finding plaintiff inadequate where he had lost the allegedly defective modem at issue in the case). Nor is Google's citation to Akaosugi v. Benihana Nat. Corp., 282 F.R.D. 241 (N.D. Cal. 2012) any more persuasive. Suppl. CC Opp. 5. In Akaosugi, Judge Alsup found a named plaintiff to be inadequate where the plaintiff had wiped 6,500 documents from a USB drive before production, previously copied company records without authorization, and subsequently attempted to conceal his wrongdoing. 282 F.R.D. at 257. As discussed above, the purportedly spoliated communications here fall short of the misconduct in Akaosugi, both in terms of culpability and relevance to Plaintiffs' claims.

Accordingly, the Court finds Cabrera and RMC to be adequate class representatives, notwithstanding Google's spoliation allegations. Google submits only that the allegedly spoliated evidence is relevant to show Peen's impressions of Plaintiffs' claims, an issue that is unlikely to play a significant enough role at trial to call into question Cabrera's adequacy as representative.

Credibility : Next, Google argues that Cabrera would be an inadequate representative because he has allegedly made false or misleading statements about his internet marketing and coaching expertise. Specifically, Google claims that he advertised himself as a “marketing genius” and “business coach” but testified in deposition that he was only “average” in marketing. Suppl. CC Opp. 6-7. In response, Plaintiffs have supplied further context for those remarks, such as by explaining that the purported statements were puffery made by Cabrera's business partner, not Cabrera. Suppl. CC Reply 4.

It is true that credibility may be a “relevant consideration with respect to the adequacy analysis,” but credibility issues will only risk defeating certification if the issues are “so sharp as to jeopardize the interests of absent class members” and are “directly relevant to the litigation.” Nevarez v. Forty Niners Football Co., LLC, 326 F.R.D. 562, 583 (N.D. Cal. 2018). Here, Cabrera's personal advertising sophistication-credible or not-is not directly relevant to the central issues of this litigation, either for the contract or the UCL claim. As such, Cabrera's adequacy is not impacted by Google's purported credibility concerns.

* * *

To conclude, the Court holds that Plaintiffs have demonstrated that the Smart Pricing Class meets Rule 23(a)'s requirements. The Smart Pricing Class is numerous, at least one common question applies, Plaintiffs are typical, and Plaintiffs and their counsel are adequate representatives.

2. Rule 23(b)(3) Requirements

In addition to demonstrating that the Smart Pricing Class satisfies the four requirements of Rule 23(a), Plaintiffs must also show that one of the three grounds under Rule 23(b) applies. Plaintiffs move to certify the Smart Pricing Class under Rule 23(b)(3), so they must demonstrate that (1) “questions of law or fact common to class members predominate over any questions affecting only individual members,” and (2) “that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.” Fed.R.Civ.P. 23(b)(3).

a. Predominance

The predominance requirement “asks whether the common, aggregation-enabling, issues in the case are more prevalent or important than the non-common, aggregation-defeating, individual issues.” Tyson Foods, Inc. v. Bouaphakeo, 577 U.S. 442, 453 (2016) (internal quotation marks and citation omitted). This inquiry “tests whether proposed classes are sufficiently cohesive to warrant adjudication by representation.” Id. (quoting Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 623 (1997)). “Predominance is not, however, a matter of nose-counting.” Ruiz Torres v. Mercer Canyons Inc., 835 F.3d 1125, 1134 (9th Cir. 2016) (citation omitted). More important questions will be given correspondingly more weight in the predominance analysis over individualized questions that are less significant to the class claims. Id. In fact, “even if just one common question predominates, ‘the action may be considered proper under Rule 23(b)(3) even though other important matters will have to be tried separately.'” In re Hyundai & Kia Fuel Econ. Litig., 926 F.3d 539, 557 (9th Cir. 2019) (quoting Tyson Foods, 577 U.S. at 453).

Google argues that individualized issues predominate over common questions for three reasons: (1) interpretation of the AdWords Agreement requires extrinsic evidence that varies between class members; (2) determining injury requires individualized inquiry; and (3) variations in advertiser sophistication create individual questions. The Court addresses each in turn.

i. Use of Extrinsic Evidence

Google argues that interpretation of the AdWords Agreement requires individualized inquiries into the extrinsic evidence that any given advertiser was exposed to. CC Opp. 10-15. The Court already dealt with this argument in its analysis of commonality, and for the same reasons the Court found interpretation of the AdWords Agreement to be a common question, it also finds that individual issues surrounding the use of extrinsic evidence do not predominate. See supra Section V.A.1.b.

ii. Auction Dynamics

Similar to a line of reasoning it presented in summary judgment, Google asserts that Smart Pricing properly applies to bids rather than to post-auction click costs. CC Opp. 17-18. This, it contends, means that evaluating injury¹⁷ entails individualized analysis of auction dynamics. Id. Plaintiffs reply that there is no need to delve into auction dynamics because their theory of liability is that the AdWords Agreement required Google to apply Smart Pricing to post-auction costs, so no inquiry into auction dynamics is necessary. CC Reply 6-7.

17 In Olean, 31 F.4th at 669, the Ninth Circuit held that a class may be certified even if it “potentially includes more than a de minimis number of uninjured class members.” In so holding, the Olean court rejected a per se rule against certifying a class if it contains more than a de minimis number of uninjured class members, but affirmed that district courts retain discretion to decline certification on a case-by-case basis if individualized questions about injury predominate over common questions. Id. 669 & n.13.

As a threshold matter, the Court observes that the parties' dispute over the relevance of individual auction dynamics reflects a fundamental disagreement about the correct interpretation of the AdWords Agreement. Plaintiffs insist that the AdWords Agreement should be interpreted to mandate the Smart Pricing of post-auction costs. See FAC ¶¶ 26-29, 31; 7/10/15 Solomon Rep. at 9-12. While Google does not explicitly dispute Plaintiffs' interpretation in its argument about auction dynamics, its assertion that Smart Pricing only applies to bids cannot hold true unless the Court ultimately rejects Plaintiffs' interpretation. Thus, to the extent that Google has identified individualized questions of injury defeating predominance, those questions arise only if Plaintiffs' fail on their legal theory regarding the interpretation of the AdWords Agreement. By contrast, if Plaintiffs prevail on their legal theory, auction dynamics become immaterial to injury because injury would depend only on whether and by how much the post-auction cost had been discounted. In other words, Google's asserted individual inquiries are contingent on a question of law and have force only if Plaintiffs do not succeed in proving their contract interpretation.

The Ninth Circuit, however, has squarely rejected such contingent theories on class certification. In United Steel Workers International Union v. ConocoPhillips Co., 593 F.3d 802 (9th Cir. 2010), the Ninth Circuit reviewed a district court order denying class certification in a wage and hour dispute. The district court had concluded that plaintiffs failed to show predominance, reasoning that, “if Plaintiffs' ‘on duty' theory of liability fails, then common questions will no longer predominate over individual ones.” Id. at 804, 807. The Ninth Circuit held that this was an abuse of discretion. Id. at 809. It explained that Rule 23 does not “give[] a court any authority to conduct a preliminary inquiry into the merits of a suit” and that “[n]either the possibility that a plaintiff will be unable to prove his allegations, nor the possibility that the later course of the suit might unforeseeably prove the original decision to certify the class wrong, is a basis for declining to certify a class.” Id. at 808-09 (first quoting Eisen v. Carlisle & Jaquelin, 417 U.S. 156, 177-78 (1974); and then quoting Blackie v. Barrack, 524 F.3d 891, 901 (9th Cir. 1975)); see also Amgen, 568 U.S. at 466 (“Rule 23 grants courts no license to engage in free-ranging merits inquiries at the certification stage.”). According to the Ninth Circuit, by crediting the contingent argument that individualized questions would predominate if the plaintiffs' legal theory failed, the district court improperly judged the merits of the case. ConocoPhillips, 593 F.3d at 809. Put more generally, “[w]hat a district court may not do is to assume, arguendo, that problems will arise, and decline to certify the class on the basis of a mere potentiality that may or may not be realized.” Id. at 810. Instead, the proper approach is to accept a plaintiff's theory of relief for purposes of class certification, Keegan v. Am. Honda Motor Co., 284 F.R.D. 504, 530 (C.D. Cal. 2012) (citing ConocoPhillips, 593 F.3d at 808), so long as the plaintiff offers “sufficient information to form a reasonable judgment” that the theory is viable. ConocoPhillips, 593 F.3d at 810 (citation omitted); see also Wal-Mart, 564 U.S. at 351 (observing that “rigorous analysis” of Rule 23 may “entail some overlap with the merits of the plaintiff's underlying claim”). Should the “potentiality” of individualized problems be realized later, the Court can then address those problems via a motion to decertify. ConocoPhillips, 593 F.3d at 809.

Applying ConocoPhillips, the Court rejects Google's auction dynamics argument. Plaintiffs have put forth evidence supporting their position that the AdWords Agreement requires Google to Smart Price the post-auction cost of a click. Mazzeo Decl., Ex. 11; 7/10/15 Solomon Rep. at 9-15. This information is sufficient for the Court to form a reasonable judgment that Plaintiffs' theory is viable. supra Section III.A.1 (discussing the same evidence). Consequently, the Court must credit Plaintiffs' interpretation for purposes of class certification. As the Court explained above, the dynamics of any given auction need not be considered under Plaintiffs' theory, so individualized inquiries into those dynamics will not predominate.

iii. Return on Investment

Google also suggests that an advertiser is not injured if she receives a good return on investment (“ROI”) from a click even if Google failed to apply any Smart Pricing discount. CC Opp. 17, 19. And because any analysis of ROI necessitates individual inquiry, it argues, individual questions predominate. Id.

In support, Google points to evidence that non-Smart Priced clicks resulted in higher ROIs than Smart Priced clicks on average. Id. at 18-19 (citing 3/20/18 Wu Decl. ¶¶ 52, 60). But the price an advertiser pays for a click, i.e. the source of the advertiser's injury, is independent of the ROI the advertiser receives for any particular click. If an advertiser pays 50 cents for a click and receives a return worth one dollar, she would still receive one dollar's worth of return, regardless of whether she paid 50 cents or something less for the click. As this example illustrates, an advertiser is injured when she does not receive a discount because she pays more for the same return than she would have otherwise; the magnitude of the return is not relevant. And in any case, events that occur after an advertiser pays for a click-i.e., the ROI that she receives-are not germane to the injury inquiry because “the injury, for standing purposes, is the breach [of contract] itself.” In re Google Referrer Header Privacy Litig., 465 F.Supp.3d 999, 1011 (N.D. Cal. 2020) (quoting Alston v. Flagstar Bank, FSB, 609 Fed.Appx. 2, 3 (D.C. Cir. 2015)).

As such, the Court concludes that individualized inquiries about ROI will not predominate.

iv. Determining the Correct Level of Smart Pricing

Google next argues that Plaintiffs have failed to provide a classwide method for identifying the correct level of Smart Pricing that should have been applied to a click, or for identifying whether Smart Pricing should have been applied at all. CC Opp. 18-20; Suppl. CC Opp. 22-23. As a consequence, it says, individualized inquiries will be necessary to identify whether a class member was injured by a click because that click did not receive the correct discount, or because the click was not Smart Priced but should have been.¹⁸ Because the issue of the correct level of Smart Pricing varies as to each subclass, the Court treats each separately below.

18 Google also briefly argues that these purported issues run afoul of Comcast Corp. v. Behrend, 569 U.S. 27, 35 (2013), which requires that a damages model “measure only those damages attributable to [a plaintiff's] theory.” CC Opp. 19-20. However, the Court finds that Google's arguments are more properly characterized as raising questions about injury than as raising a mismatch between Plaintiffs' theory of liability and damages model.

Non-Smart Priced Clicks Subclass : Google asserts that Plaintiffs cannot determine whether a click that was not Smart Priced should have received a Smart Pricing discount. CC Opp. 18. In its view, Smart Pricing applies only to clicks for which an advertiser has (1) not selected certain bidding strategies and (2) that are less likely to result in a conversion, but Plaintiffs have not explained how they would identify clicks according to those two criteria. Id. at 18. Plaintiffs respond that Smart Pricing applies to post-auction click costs rather than bids, CC Reply 6-7, but did not otherwise address the two criteria that Google identified.

To begin, the Court can quickly dispose of the argument that Plaintiffs must account for advertisers' bidding strategies. Nowhere do Plaintiffs argue that the AdWords Agreement limits Smart Pricing to only certain bidding strategies, and any argument that runs counter to Plaintiffs' theory of legal liability cannot defeat class certification. See ConocoPhillips, 593 F.3d at 808-10.

The second half of Google's argument, however, has merit. Plaintiffs have consistently represented and argued that Smart Pricing does not apply to all clicks indiscriminately but is instead limited to clicks that are less likely to result in a conversion than a click on the Search Network (google.com and other search sites powered by Google). Compl. ¶¶ 25-26 & n.9; CC Mot. 4-5 (arguing and citing evidence showing that Smart Pricing applies to clicks that are less likely to convert than those on the Search Network). The Ninth Circuit, too, has explained that Smart Pricing is calculated by comparing the “the conversion rate for [a] lower-quality website [with] the conversion rate for the same ad on google.com.” Pulaski & Middleman, LLC v. Google, Inc., 802 F.3d 979, 982 (9th Cir. 2015). As a result, even applying Plaintiffs' theory of liability as ConocoPhillips requires, Plaintiffs may not simply assume that every non-Smart Priced click represents a breach. They must be able to identify whether clicks convert less than the advertisements on the Search Network, because it is only those clicks that must be Smart Priced.

Although Plaintiffs have provided a method for identifying the set of clicks that were not Smart Priced, they have not offered any systematic method for identifying which of those clicks should have been Smart Priced. In the absence of such a method, the Court cannot be confident that common questions predominate,¹⁹ or if the only way to identify clicks that should have been Smart Priced is to conduct extensive individualized inquiries. As a consequence, the Court finds that Plaintiffs have not carried their burden to show predominance as to the Non-Smart Priced Clicks Subclass.

19 Indeed, Plaintiffs have offered very little detail about the data Google keeps and whether that data could answer these questions. Plaintiffs' choice to use click cost multipliers to calculate Smart Pricing damages implies that, at some level, those multipliers carry information about the likelihood of conversion because the amount of the discount should be tied to the likelihood of conversion. See supra Section III.A.3. But Plaintiffs' experts work with several different multipliers-“Impression Click Cost Multipliers,” “Query Click Cost Multipliers,” “Raw Bid Multipliers,” “Production Bid Multipliers,” and “Manual Bid Multipliers”-without ever explaining how those multipliers differ or how Google calculates those multipliers. See 7/9/15 Gibson Rep., App. E at 3, 10-11. And their experts also appear to use click cost multipliers and bid multipliers interchangeably, again with little explanation for how those multipliers are derived and what they mean. 7/10/15 Solomon Rep. at 31-32. Moreover, one way that Plaintiffs identify Non-Smart Priced Clicks is by looking for clicks where the click cost multiplier is missing. 7/9/15 Gibson Rep. ¶ 34. So even if multipliers do carry information on the likelihood of conversion, it is not clear how one could determine if a click with a missing multiplier would be less likely to convert.

AFMA Clicks Subclass : Google argues that there is no single multiplier that can appropriately be applied category-wide to all AFMA Clicks. CC Opp. 19. According to Google, multipliers are uniquely determined based on a variety of factors, including characteristics of the advertiser, advertisement, and web property on which the advertisement appears. Id. (citing 3/20/18 Wu Decl. ¶ 72). Plaintiffs disagree, maintaining that multipliers are set at the webproperty level and do not require analysis of individual advertisers or advertisements. Suppl. CC Reply 14-15. They also assert that they have evidence showing 0.08 is the correct multiplier across the entire category of AFMA Clicks. CC Mot. 7-8; see also 7/9/15 Gibson Rep. ¶ 38 (using the 0.08 multiplier as a criterion for identifying actionable clicks). To support their use of the 0.08 multiplier, they cite two emails in which Google employees estimate that the default multiplier for such clicks could be as low as 0.08. Mazzeo Decl., Ex. 32, ECF No. 424-41; Mazzeo Decl., Ex. 33, ECF No. 424-42; see also 7/10/15 Solomon Rep. at 33 & n.112 (observing that 0.08 is a default multiplier). Google responds that a default multiplier is used only in the first few weeks of an advertisement's publication and therefore cannot be used to measure injury across the entire category of AFMA Clicks. Suppl. CC Opp. 22 (citing Mazzeo Decl., Ex. 32).

The Court agrees with Google. One of the two emails cited by Plaintiffs makes clear that 0.08 was a default multiplier value, and only a potential value at that. Mazzeo Decl., Ex. 32. Further, the email explains that a default multiplier is temporary-used for an advertisement's first four weeks unless a lack of web traffic causes that advertisement to stay in the default state for a longer time-and does not apply “for the lifetime of the property.” Id. This illustrates that, far from acting as a standard barometer for evaluating all AFMA Clicks, the 0.08 multiplier is an interim designation in place for the brief period after an advertisement's initial publication.

Plaintiffs point to no evidence that default multipliers apply past the default period and can be used throughout the entire lifetime of an advertisement. Instead, they suggest that if 0.08 is the incorrect multiplier, they can simply substitute 0.08 with some other number derived from discovery. 7/10/15 Solomon Rep. at 33. But this misunderstands the problem. The obstacle to class certification is not the value of the multiplier being used to determine injury. Rather, the obstacle is that Plaintiffs have not submitted evidence showing that there even exists a common multiplier that applies across the entire subclass. And they have not offered any other method for systematically determining the proper Smart Pricing multiplier for AFMA Clicks. Without being presented with such a method or with a common multiplier, the Court cannot conclude that common issues predominate. Accordingly, Plaintiffs have failed to meet their burden to show predominance as to the AFMA Clicks Subclass.²⁰

20 The Court notes that the parties also dispute whether multipliers are calculated at the webproperty or the advertisement level. However, even if Plaintiffs were correct that multipliers are calculated at the web-property level, that would not resolve the problem of identifying the correct multiplier once the default state expires.

The Court observes that Plaintiffs may be able certify a narrower subclass limited to AFMA Clicks within the default period, but because denial of class certification is without prejudice, the Court will not modify the class definition in the first instance.

Search Bundled Clicks Subclass : In contrast to their theory regarding AFMA Clicks, Plaintiffs contend that every Search Bundled Click-i.e., Display Network clicks on advertisements shown on both the Display and Search Networks for which an advertiser did not set different bids for each network-was uniformly overpriced by 6%. CC Mot. 8; see also Mazzeo Decl., Ex. 35, ECF No. 424-44 (internal Google email noting that Google was applying a 6% increase to all multipliers for Search Bundled Clicks); Mazzeo Decl., Ex. 38, ECF No. 424-48 (internal Google email stating that the 6% multiplicative factor was “likely not the correct value[]”). If Plaintiffs are correct, then every Search Bundled Click would represent an injury because every such click would be overpriced. That being the case, there would be no need for individualized questions about injury, so such questions do not predominate as to the Search Bundled Clicks Subclass.

mGDN Clicks Subclass : For mGDN Clicks, Plaintiffs assert that advertisers were injured whenever Google used a Smart Price multiplier greater than 0.42. CC Mot. 8; CC Reply 5-6. Once again, though, Plaintiffs have relied on a default multiplier to identify injury. Mazzeo Decl., Ex. 78, ECF No. 456-8 (the 0.42 multiplier is a default value used in an internal Google experiment); 7/10/15 Solomon Rep. at 33 & n.114 (acknowledging that 0.42 is a default multiplier). The use of a default multiplier here generates the same obstacles (for the same reasons) as for the AFMA Clicks Subclass. As a result, Plaintiffs have not met their burden to show predominance for the mGDN Clicks Subclass. Similar to the AFMA Clicks Subclass, Plaintiffs may be able to narrow this subclass to include only clicks within the default period, but the Court does not modify the class definition here.

v. Variations in Advertiser Sophistication

Google's last argument against predominance is that differences in the sophistication of advertisers create individualized issues. Suppl. CC Opp. 23-24. They contend that the most sophisticated advertisers frequently update their bid settings, thereby reducing the harm they suffer from misapplications of Smart Pricing. Id. What Google has just described is how sophisticated advertisers may more effectively mitigate their damages than less sophisticated ones. But that does not defeat predominance because “[d]amages calculations alone . . . cannot defeat certification.” Pulaski, 802 F.3d at 988 (alteration in original) (citation omitted).

b. Superiority

Finally, Plaintiffs must show that “a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.” Fed.R.Civ.P. 23(b)(3). This requirement tests whether “classwide litigation of common issues will reduce litigation costs and promote greater efficiency.” Valentino v. Carter-Wallace, Inc., 97 F.3d 1227, 1234 (9th Cir. 1996). In making that determination, courts should consider the factors enumerated in Rule 23: “(A) the class members' interests in individually controlling the prosecution or defense of separate actions; (B) the extent and nature of any litigation concerning the controversy already begun by or against class members; (C) the desirability or undesirability of concentrating the litigation of the claims in the particular forum; and (D) the likely difficulties in managing a class action.” Fed.R.Civ.P. 23(b)(3).

Google does not contest this requirement, and the Court finds that Plaintiffs have shown a class action is superior to individual actions. Plaintiffs' expert calculated that former plaintiff Woods would be entitled to only a few hundred dollars for his Smart Pricing claim, and Cabrera would be entitled to less than ten dollars for the same claims. 7/10/15 Solomon Rep. at 37 (Woods' damages); 11/8/2018 Suppl. Solomon Rep. at 4-5, ECF No. 424-86 (Cabrera's damages). The low potential damages are far outweighed by the high costs of litigation in a case like this, so it is clear that class members do not have sufficient incentive to pursue individual litigation. See Vizcarra v. Unilever U.S., Inc., 339 F.R.D. 530, 556 (N.D. Cal. 2021). For such negative value suits, “a class action is the only realistic possibility for redress.” Hamilton v. WalMart Stores, Inc., 39 F.4th 575, 588 n.6 (9th Cir. 2022).

* * *

To summarize, the Court finds that Plaintiffs have satisfied the requirements of Rules 23(a) and 23(b)(3) as to the Search Bundled Clicks Subclass and therefore GRANTS certification of that subclass. But, while Plaintiffs have satisfied Rule 23(a) and the superiority requirement of Rule 23(b)(3) as to the Non-Smart Priced Clicks, AFMA Clicks, and mGDN Clicks Subclasses, they have failed to demonstrate that common questions predominate. The Court therefore DENIES certification of those three subclasses.

B. Location Targeting Class

Plaintiffs have defined the Location Targeting Class as follows:

All persons and entities located within the United States who, between January 1, 2004 and March 22, 2011, advertised through Google's AdWords program and paid for clicks on their Google AdWords advertisement(s), where such clicks did not originate from the location selected by the advertiser.

CC Mot. 16.

1. Rule 23(a) Requirements

a. Numerosity

Rule 23(a)(1)'s numerosity requirement is plainly satisfied here. Generally, courts will find the requirement to be satisfied when a proposed class exceed forty members. See, e.g., Arroyo, 2019 WL 1508457, at *2. With respect to the Location Targeting claim, Google has admitted that “more than 50 AdWords Advertisers used Location Targeting and paid for clicks on Ads appearing to Internet users physically located outside of the AdWords Advertisers' Targeted Locations.” Decl. Margaret E. Mazzeo (“Mazzeo Decl.”), Ex. 36 at 17 (Request for Admission No. 5). Accordingly, numerosity is satisfied.

b. Commonality

The proposed Location Targeting Class also meets the commonality requirement of Rule 23(a)(2). To satisfy this element, the claims of the proposed class “must depend upon a common contention . . . of such a nature that it is capable of classwide resolution-which means that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke.” Wal-Mart, 564 U.S. at 350. Even a “single significant question of law or fact” will satisfy the commonality requirement. Stockwell, 749 F.3d at 1111.

For the Location Targeting Class, Plaintiffs have identified the common question as “whether Google's statements on the Settings Screen were ‘likely to deceive' an AdWords advertiser into believing that Google would only charge for clicks on ads that originated from users physically located within the advertiser's selected location.” CC Mot. 19-20. This question encapsulates the UCL's “likely to deceive” liability inquiry, which is assessed under the objective “reasonable consumer” standard and supported by statements viewed by all class members. Id.; see also supra Section III.B. Google does not directly dispute the existence of a common question, though it does argue that individual inquiries outweigh any such common questions. See CC Opp. 22-32. Accordingly, the Court finds that Rule 23(a)'s commonality requirement is satisfied because there exists at least one significant question that will resolve issues common to all Location Targeting claims.

c. Typicality

“Typicality focuses on the class representative's claim-but not the specific facts from which the claim arose-and ensures that the interests of the class representative align with the interests of the class.” Stiner v. Brookdale Senior Living, Inc., 2023 WL 2722294, at *27 (N.D. Cal. Mar. 30, 2023) (citing Just Film, 847 F.3d at 1116). Rule 23(a)(3) requires only that the class representative's claims be “reasonably coextensive with those of absent class members; they need not be substantially identical.” Just Film, 847 F.3d at 1116.

The Court finds that RMC demonstrated that its claims are reasonably coextensive with, and its interests are aligned with, those of absent class members. The Location Targeting claim arises from substantially identical statements that all AdWords advertisers were exposed to when they set up their campaigns on the AdWords interface. See CC Mot. 11 (citing Mazzeo Decl., Ex. 44 (“Tang Dep”) at 189:18-191:2, ECF No. 424-54; Ex. 45, ECF No. 424-55 (“In what geographical locations do you want your ads to appear?”); Ex. 47 at 7, ECF No. 424-57 (“Highlight the cities and regions on the left where you'd like your ad to appear.”)). RMC's UCL claim is, therefore, coextensive with those of absent class members who had also specified certain locations to target but paid for clicks from users outside of those locations.

Google argues that RMC is not typical of the class “due to variations in [advertiser] sophistication.” Suppl. CC Opp. 19. In support, Google cites evidence that some advertisers were “obviously orders of magnitude larger and more experienced than Cabrera” and who “vary in their use of external support and sources of information.” Id. at 10-11. However, this suggestion-that RMC is atypical because more sophisticated class members would not have been deceived-flies in the face of the UCL's objective “reasonable consumer” standard, as well as the UCL's broad purpose to protect consumers from unfair business practices. See Pulaski, 802 F.3d at 985; Tobacco II, 46 Cal.4th at 320 (“[R]elief under the UCL is available without individualized proof of deception, reliance and injury.”); Nilon v. Nat.-Immunogenics Corp., 2014 WL 12570897, at *5 (S.D. Cal. Apr. 15, 2014) (rejecting UCL defendant's argument that “many factors and variables will influence each putative class member's decision” because “[i]n determining whether typicality is met, the focus should be on the defendants' conduct and the plaintiffs' legal theory, not the injury caused to the plaintiff”). Accordingly, Google's typicality argument regarding advertiser sophistication can be readily rejected under the UCL case law.²¹

21 Similarly, Google's secondary argument-that Plaintiffs' purportedly idiosyncratic “aversion to conversions coming from out-of-area users interested in his services” would render them atypical (CC Opp. 32)-has been rejected in this Circuit. Courts have recognized that “individual experience with a product is irrelevant because the injury under the UCL . . . is established by an objective test.” Astiana v. Kashi Co., 291 F.R.D. 493, 502 (S.D. Cal. 2013) (finding typicality satisfied where class representatives had different perceptions, knowledge, preferences, and reasons for purchasing the defendant's products) (internal quotation marks omitted).

Google also briefly recycles its standing argument as a typicality argument, contending that RMC is an atypical class member because it received the benefit of the challenged conduct and therefore suffered no injury. Suppl. CC Opp. 19. However, as the Court determined above at Section III.B.3, Google's “benefit of the bargain” theory is not persuasive.

In sum, the Court finds that RMC has satisfied Rule 23(a)'s typicality requirement.

d. Adequate Representation

Google raises the same arguments against adequacy that the Court already rejected in its analysis of the Smart Pricing Class. supra Section V.A.1.d. The Court therefore finds that RMC is also adequate to serve as the class representative for the Location Targeting Class, and that proposed class counsel are adequate as well.

* * *

Based on the foregoing reasons, the Court is satisfied that Plaintiffs have met their burden on the Rule 23(a) requirements. The Court finds that Location Targeting Class is sufficiently numerous, there is at least one common question of law or fact, RMC is typical, and both RMC and its counsel will fairly and adequately protect the interests of the class.

2. Rule 23(b)(3) Requirements

As Plaintiffs seek to certify the Location Targeting Class under Rule 23(b)(3), they must show that (1) questions of law or fact common to class members predominate over any questions affecting only individual members; and (2) a class action is superior to other available methods for fairly and efficiently adjudicating the controversy. Fed.R.Civ.P. 23(b)(3).

a. Predominance

“The predominance inquiry is mainly concerned with ‘the balance between individual and common issues.'” In re Hyundai & Kia, 926 F.3d at 560. However, this inquiry is not simply a “matter of nose-counting.” Ruiz Torres, 835 F.3d at 1134. Greater weight is given to more important questions such that a class may be certified even if only a single but critical common question predominates. In re Hyundai & Kia, 926 F.3d at 557.

To assess whether common questions predominate over individual ones, the Court begins by considering the questions common to the class. The Court has already found-and Google does not dispute-that RMC has at minimum satisfied Rule 23(a)'s commonality requirement by identifying a question common to all class members: “whether Google's statements on the Settings Screen were ‘likely to deceive' an AdWords advertiser into believing that Google would only charge for clicks on ads that originated from users physically located within the advertiser's selected location.” CC Mot. 19-20; see also supra Section V.B.l.b. Because the UCL does not call for individualized determinations, the key question of UCL liability would be the same for all class members and be susceptible to common proof. See, e.g., Tobacco II, 46 Cal.4th at 312; Arris, 327 F.R.D. at 365 (concluding that UCL fraud claim was susceptible to classwide proof).

Given these characteristics of the UCL statute, many courts have unsurprisingly remarked that UCL claims are “ideal for class certification because they will not require the court to investigate class members' individual interaction with the product.” Bradach v. Pharmavite, LLC, 735 Fed.Appx. 251, 255 (9th Cir. 2018) (quoting Tait v. BSH Home Appliances Corp., 289 F.R.D. 466, 480 (C.D. Cal. 2012)); see also Gold v. Lumber Liquidators, Inc., 323 F.R.D. 280, 290 (N.D. Cal. 2017) (noting that the UCL is “particularly amenable to class certification”); Bailey v. Rite Aid Corp., 338 F.R.D. 390, 407 (N.D. Cal. 2021) (“Courts routinely hold that if a plaintiff shows by a preponderance of the evidence that the questions of materiality and likelihood of deception can be resolved with common evidence based on the objective reasonable consumer standard, then common questions predominate over individual ones with respect to claims under the UCL, CLRA, and FAL.”) (emphasis added).

In short, the common question RMC has identified is an important question-if not the most important-relating to the UCL fraud claim and is “apt to drive the resolution of the litigation.” Ruiz Torres, 835 F.3d at 1134. Therefore, Google faces a considerable burden to identify sufficient individualized inquiries that would undermine the predominance of a common question that is so central to the UCL class claim. To that end, Google advances several varieties of individualized inquiries in its initial and supplementary oppositions. Specifically, it contends that individualized inquiries exist as to (1) class members' injuries, CC Opp. 22-27, Suppl. CC Opp. 11-17; (2) the materiality of the “area-of-interest targeting” feature, CC Opp. 27-28; (3) whether certain class members were exposed to and relied on the Location Targeting statements on the AdWords Settings Screen, CC Opp. 29-32, Suppl. CC Opp. 8-11; and (4) choice of law, CC Opp. 32-33. It also asserts that (5) RMC's damages models run afoul of Comcast, 569 U.S. 27. The Court addresses each in sequence.

i. Injury

First, Google argues that-assuming “[w]hat advertisers care about are results, not necessarily where someone interested in their products or services may be located”-certain advertisers may not have been injured at all and may have even benefited from out-of-area clicks. CC Opp. 23. In support, Google proffers representative sample evidence that, in the aggregate, out-of-area clicks typically converted at a rate higher than in-area clicks. Id. To the extent Google is re-asserting the substance of its argument on summary judgment that class members received the “benefit of the bargain,” the Court has explained above that this theory is not supported under California law. See supra Section III.B.3 (citing Kwikset, 51 Cal.4th at 332).

Google attempts to recast its previous injury argument by repeatedly emphasizing that the unwanted product here (out-of-area ads) was not valueless and may even have performed better than what class members actually selected (in-area ads). See generally CC Opp. 24-27 (citing purported benefits in distinguishing Lambert v. Nutraceutical Corp., 870 F.3d 1170 (9th Cir. 2017); then distinguishing Pulaski, 802 F.3d at 989; and then distinguishing Arris, 327 F.R.D. at 334). This, however, is a distinction without a difference. The California Supreme Court noted in Kwikset that, “to an observant Jew who keeps kosher, [non-kosher meat] would be worthless.” 51 Cal.4th at 330. This observation would remain unchanged even if the non-kosher meat was nutritionally or culinarily superior to kosher meat, which is the equivalence of Google's contention here that out-of-area ads performed better than in-area ads. Id. at 329 (“For each consumer who relies on the truth and accuracy of a label and is deceived by misrepresentations into making a purchase, the economic harm is the same: the consumer has purchased a product that he or she paid more for than he or she otherwise might have been willing to pay if the product had been labeled accurately.”) (emphasis added). Likewise, the class members here have been injured, notwithstanding the purported better performance of the unwanted clicks; all advertisers paid more for the out-of-area clicks than what they otherwise might have been willing to pay.

The Ninth Circuit has also specifically rebuffed Google's attempt to defeat class certification using after-the-fact benefits received by consumers, holding instead that “UCL . . . restitution is based on what a purchaser would have paid at the time of purchase had the purchaser received all the information.” Pulaski, 802 F.3d at 989 (emphasis added); cf. Kwikset, 51 Cal.4th at 329. Accordingly, the fact that certain class members may have received a benefit or “conversion” from an out-of-area click would not change the amounts those members would be entitled to on restitution, which are measured at the time of those transactions. Google's after-the-fact benefits argument, therefore, does not raise a true individualized inquiry.²²

22 The Court does not read Google's briefs to be arguing that predominance may be defeated by individual inquiries into the amounts of damages. However, to the extent Google's argument on purported benefits relies on the different degrees of injury suffered by different class members, the Court reiterates the well-established rule in the Ninth Circuit that “damage calculations alone cannot defeat certification.” See, e.g., Yokoyama v. Midland Nat. Life Ins. Co., 594 F.3d 1087, 1094 (9th Cir. 2010).

The Court can appreciate that the nature of digital advertisements and “clicks” may present some difficulties in applying conventional injury principles. For example, given that advertising success is inherently accompanied by a degree of probability, it may be difficult to compare the value of an in-area click with the value of an out-of-area click, at least from an advertiser's perspective at the outset. However, the Ninth Circuit in Pulaski was confronted with an analogous quandary-also involving a UCL case relating to Google advertisements-where “the harm alleged [was] Google's placement of ads on lower-quality web pages without the advertisers' knowledge.” 802 F.3d at 989. There, the Ninth Circuit succinctly and clearly held that “restitution is available on a classwide basis once the class representative makes the threshold showing of liability under the UCL,” rejecting Google's contentions that class-wide restitution would include some advertisers who “derived direct economic benefits from ads placed on parked domains and error pages.” Id. at 984, 986. Here, confronted with a similar theory of injury based on a similar structure for advertising clicks, the Court cannot discern a meaningful basis to depart from the Ninth Circuit's holding in Pulaski.

Accordingly, the Court does not find that Google has raised any significant issues relating to the individualized assessments of after-the-fact benefits certain class members may have received from out-of-area clicks, and certainly not any issue sufficiently predominant to outweigh the common question of UCL liability.

ii. Materiality

Google also contends that there would be individual inquiries into materiality, namely whether it was material to class members that Google would charge advertisers for clicks originating outside of their selected locations. CC Opp. 27-28. In support of this proposition, Google relies on the fact that-after it allowed advertisers to opt out of the query parsing practice that gave rise to certain out-of-area clicks-only a slim minority of AdWords accounts opted out of the practice. Id.

As the Court explained above at Section III.B.2, a failure to establish materiality is not necessarily fatal for a UCL fraud claim because the core inquiry is whether reasonable members of the public are likely to be deceived. This framework has led courts to remark that, in “UCL fraud cases, questions of materiality . . . do not necessarily undermine predominance.” DZ Rsrv. v. Meta Platforms, Inc., 2022 WL 912890, at *7 (N.D. Cal. Mar. 29, 2022). And it makes sense why this is so; asking whether a reasonable consumer would have been deceived is an objective question that “can be proved through evidence common to the class[, so it] is a common question for purposes of Rule 23(b)(3).” Amgen, 568 U.S. at 467 (internal quotation marks and brackets omitted); see also Milan v. Clif Bar & Co., 340 F.R.D. 591, 599 (N.D. Cal. 2021) (“[T]he UCL, FAL, and CLRA . . . require only an objective showing that members of the public were likely to have been deceived by [the] product claims.”). Given the well-recognized objective nature of the materiality analysis, the Court finds Google's materiality argument to be suspect from the outset.

In support of its theory of individualized materiality, Google proffers data it collected after it allowed advertisers to opt out of query parsing. CC Opp. 27. This data shows that few advertisers opted out of query parsing once Google offered the option to do so (Evans Decl., Ex. 41 ¶¶ 7-9, ECF No. 443-24), which, according to Google, raises individualized issues. To the contrary, Google's reliance on this data actually supports RMC's position that materiality is a common question, because it did not arise out of individualized case studies; it was a study conducted across all AdWords accounts. In other words, even if Google is correct that its data shows LT Settings Screen statement were not material, “individual issues would not predominate; instead, [RMC's] claims would [simply] fail on a classwide basis.” Arris, 327 F.R.D. at 365. Google's quantitative data may be persuasive to a factfinder on the merits of RMC's claims, but for class certification, this data does not raise any individualized inquiries that would outweigh the common questions for RMC's UCL claim.

Accordingly, Google's evidence of “individualized issues” relating to the materiality of “area-of-interest targeting” do not disturb the predominance of common questions.

iii. Exposure and Reliance

Google further asserts that RMC has not satisfied its burden under Rule 23(b)(3) because individualized issues of reliance predominate. CC Opp. 29-32; Suppl. CC Opp. 8-11. Specifically, it contends that class members' reliance on the LT Settings Screen statements would vary depending on whether they viewed other curative disclosures and their level of sophistication.

Reliance and exposure to misrepresentations are related but not identical concepts under the UCL. With regards to reliance, the California Supreme Court in Tobacco II “interpreted [the UCL] statute to mean that named plaintiffs, but not absent ones, must show proof of ‘actual reliance' at the certification stage.” Walker v. Life Ins. Co. of the Sw., 953 F.3d 624, 630 (9th Cir. 2020) (citing Tobacco II, 46 Cal.4th 298). However, this presumption of absent class members' reliance may not hold if “there was no cohesion among the members because they were exposed to quite disparate information.” Id. at 631 (emphasis added) (quoting Stearns v. Ticketmaster Corp., 655 F.3d 1013, 1020 (9th Cir. 2011)). “In the absence of the kind of massive advertising campaign at issue in Tobacco II, the relevant class must be defined in such a way as to include only members who were exposed to advertising that is alleged to be materially misleading.” Id. at 634 (emphasis added) (quoting Mazza v. Am. Honda Motor Co., 666 F.3d 581, 596 (9th Cir. 2012), overruled on other grounds by Olean, 31 F.4th 651); see also Berger, 741 F.3d at 1069 (“[C]lass certification of UCL claims is available only to those class members who were actually exposed to the business practices at issue.”) (emphasis added).

Exposure : Here, RMC does not specifically allege or demonstrate the existence of a “massive advertising campaign” for Google's Location Targeting features. Accordingly, the Location Targeting Class must be defined to only include members who were exposed to the LT Settings Screen statements. RMC contends that it is. See Suppl. CC Reply 6 (“[T]he alleged misrepresentation appeared on the Setting Screen, which all LT Class members necessarily viewed when they signed up for location targeting.”). The Court agrees with RMC. The instructive and operative function of the LT Settings Screen statements creates a strong-if not, conclusive- inference that all advertisers would have seen and been exposed to the LT Settings Screen statements as they were setting up Location Targeting services. See Pulaski, 802 F.3d at 986 (rejecting disparate exposure argument where the UCL claim relied on statements in “the AdWords sign-up materials, all of which were presented to putative class members through the same online portal”) (emphasis added).

In its supplemental opposition, Google postulates that some class members might not have viewed or been exposed to the LT Settings Screen statement because they relied on advertising agencies to set up their AdWords campaigns. Suppl. CC Opp. 9. This argument arises from two paragraphs in Google's expert reports that remark broadly on survey results regarding the percentages of advertisers who use “external support” or “outsource” their advertising activities. Id. (citing Silverman Decl., Ex. 3 (“Wilcox Report”) ¶ 103, ECF No. 597-6; and then citing Evans Decl., Ex. 5 (“Kivetz Report”) ¶ 40, ECF No. 443-10). RMC responds that, even if class members used a third-party agency to implement Location Targeting services, agency law permits the principal to recover for fraud on its agent. Suppl. CC Reply 6-7. Although it is not particularly convinced by either party's arguments or analysis on this point, the Court finds that Google's evidence is too speculative to draw the conclusions it seeks and is not tailored to the Location Targeting Class but rather to some unspecified population of digital advertisers.

The Court does not find that Google's cited survey results support the proposed inference that class members had so relied on external agencies such that they were not exposed to Google's LT Settings Screen statement. For instance, the Wilcox Report only indicated that Google AdWords campaigns often use “external support,” which was defined to include campaign strategy support, those with more sophistication, and performance monitoring. Wilcox Report ¶ 61. Based on the Court's review, no survey response suggested that the advertisers' “external support” encompassed an external agency setting up advertisers' Location Targeting services for them such that advertisers would not have been exposed to the LT Settings Screen statement. Likewise, the Kivetz Report simply summarized the findings from a 2013 Digital Marketing Spending Survey that did not appear to be limited to Google AdWords advertisers. Kivetz Report ¶ 40. Moreover, the Kivetz Report only noted that the survey found certain advertisers “outsource their digital advertising activities,” without any indication as to whether “outsource” involved delegating tasks such as setting up Location Targeting services. Id. Simply put, the Court cannot infer from the survey responses (i.e., “external support” or “outsource”) whether advertisers would or would not have employed a third-party advertising agency that would have precluded exposure to the LT Settings Screen statements.²³

23 For similar reasons, the Court also declines to engage with RMC's theory of imputing fraud committed against an agent to fraud on the principal. Suppl. CC Reply 6-7. There is insufficient evidence before the Court as to the prevalence and use of third-party advertising agencies amongst the Location Targeting Class members, but there is even less evidence as to the nature of those third parties' relationships that would support class-wide inferences as to agency findings.

Accordingly, the Court finds that the Location Targeting Class is defined to only capture members who were exposed to the LT Settings Screen statements. Google's proffered evidence of potential third-party agencies do not support its proposed inferences that class members relied on those agencies such that they were not exposed to LT Settings Screen statements.

Reliance : As the Court indicated above, reliance is only relevant in the class certification context to the extent courts must ensure that class members were exposed to the misrepresentations at issue. See Walker, 953 F.3d at 631 (“Rule 23(b)(3)'s predominance inquiry does concern itself with exposure, but for the limited purpose of satisfying the UCL's standing requirement of reliance.”). To the extent Google suggests that there exist individualized inquiries of absent class members' reliance on the LT Settings Screen statements, that theory is foreclosed by the California Supreme Court's holding in Tobacco II, 46 Cal.4th at 324 (declining to impose “section 17204's standing requirements on absent class members in a UCL class action”).

However, Google nonetheless contends that the Court must consider individual inquiries as to whether class members viewed “curative disclosures.” In support, Google relies primarily on Mazza v. American Honda Motor Co., 666 F.3d 581 (9th Cir. 2012), to argue that a UCL class cannot be certified when the misleading representations were “dispelled by disclosures in other materials.” CC Opp. 29. Mazza, however, is factually distinguishable and does not support Google's broad proposition that any misrepresentation can be cured with disclosures elsewhere. First, the Ninth Circuit in Mazza declined to infer reliance because “it [was] likely that many class members were never exposed to the allegedly misleading advertisements, insofar as advertising of the challenged system was very limited.” Mazza, 666 F.3d at 595 (emphasis added). This is plainly not the case here, where the challenged LT Settings Screen statements were the very instructions that advertisers would have seen as they were selecting which locations to target with their ads. Second, the defendant in Mazza did not affirmatively “dispel” (as Google suggests) the misrepresentations using other disclosures. Rather, the Ninth Circuit simply observed that the misleading advertising materials did not make any affirmative representations regarding the existence of the product limitations at issue. See Mazza, 666 F.3d at 596 (noting that defendant's “advertising materials do not deny that [the challenged undisclosed] limitations exist”) (emphasis added). This too has limited relevance to the instant case, where the ubiquitous LT Settings Screen statements affirmatively instructed advertisers to select the “geographical locations” or the “cities and regions . . . where [they would] like [their] ad to appear.”²⁴ See Mazzeo Decl., Ex. 45 (“In what geographical locations do you want your ads to appear?”); Ex. 47 at 7 (“Highlight the cities and regions on the left where you'd like your ad to appear.”). Both points were central to the Ninth Circuit's holding in Mazza, but neither are present here.

24 To the extent Google relies on the single remark from Mazza that the “relevant class must also exclude those members who learned of the [] allegedly omitted limitations before they purchased or leased the [product],” 666 F.3d at 596, the close temporal and spatial proximity of the LT Settings Screen statements to the actual process of setting up Location Targeting would effectively render this argument untenable.

In addition to Mazza, Google also cites a string of cases that purportedly establishes the role of curative disclosures, CC Opp. 30-31, all of which are readily distinguishable from the present facts. See Philips v. Ford Motor Co., 2016 WL 7428810, at *23 (N.D. Cal. Dec. 22, 2016) (involving solely Magnusson-Moss Act and Song-Beverly Act claims, not UCL claims); Schellenbach v. GoDaddy.com, LLC, 321 F.R.D. 613, 623 (D. Ariz. 2017) (finding reliance was not susceptible to common proof because plaintiff “cannot show that all subclass members were subjected to a uniform omission”); Backhaut v. Apple Inc., 2015 WL 4776427, at *7 (N.D. Cal. Aug. 13, 2015) (denying UCL class certification because named plaintiffs failed to show “lost money or property”); In re Google Inc. Gmail Litig., 2014 WL 1102660, at *1 (N.D. Cal. Mar. 18, 2014) (asserting various state wiretap claims but no UCL claim); Decker v. Mazda Motor of Am., Inc., 2013 WL 12129281, at *7 (C.D. Cal. Mar. 29, 2013) (involving fraudulent omission of an automobile design defect that was repeatedly disclosed and published in technical updates).

Nor is there any basis for considering variations in class members' sophistication as part of the Court's Rule 23(b)(3) analysis. As the Court discussed above at Section V.B.1.c, Google's suggestion that UCL reliance is predicated on the sophistication of advertisers is inconsistent with the UCL's objective “reasonable consumer” standard. See also, e.g., Skinner v. Ken's Foods, Inc., 53 Cal.App. 5th 938, 948 (2020) (noting that a reasonable consumer “need not be ‘exceptionally acute and sophisticated,' nor must they ‘necessarily be wary or suspicious of advertising claims'”).

Accordingly, pursuant to Tobacco II and subsequent Ninth Circuit decisions, the Court does not find that issues arising from absent class members' reliance or exposure to curative disclosures are proper individualized UCL inquiries and, therefore, do not defeat the predominance of common questions.

iv. Choice of Law

Next, Google argues that, because the UCL is a California law that applies only to California class members, class members from other states would be subject to different consumer protection laws, resulting in individualized inquiries into those different consumer protection laws. CC Opp. 32-33 (citing Mazza, 666 F.3d at 590). RMC responds that California law applies to all putative class members, regardless of their state of residency, because all members were party to the AdWords agreement containing a California choice-of-law clause. CC Reply 18-19.

“A federal court sitting in diversity must look to the forum state's choice of law rules to determine the controlling substantive law.” Zinser v. Accufix Rsch. Inst., Inc., 253 F.3d 1180, 1187 (9th Cir. 2001). The California Supreme Court has recognized two different frameworks for selecting what jurisdiction's law should apply: (1) “[w]hen the parties have an agreement that another jurisdiction's law will govern their disputes, the appropriate analysis for the trial court to undertake is set forth in Nedlloyd [Lines B.V. v. Superior Ct., 3 Cal.4th 459 (1992)], which addresses the enforceability of contractual choice-of-law provisions”; and (2) “when there is no advance agreement on applicable law, but the action involves the claims of residents from outside California, the trial court may analyze the governmental interests of the various jurisdictions involved....” Wash. Mut. Bank, FA v. Superior Ct., 24 Cal.4th 906, 914-15 (2001).

Because there is indeed an “advance agreement on applicable law” here (the AdWords Agreement), the Court must apply the multi-step Nedlloyd framework to determine the enforceability of the choice-of-law clause. Wash. Mut. Bank, 24 Cal.4th at 915. To begin, the Court must analyze the choice-of-law clause to determine whether “the various claims of putative class members fall within its scope.” Id. at 916. If the class claims fall within the scope of the choice-of-law clause, the Court then determines whether the chosen state has a “substantial relationship to the parties or their transaction.” Id. If such a nexus exists, the Court must then ensure that the chosen state's law is not contrary to a “fundamental policy of California.” Id. If the clause passes these hurdles, “the [C]ourt shall enforce the parties' choice of law.” Id.

The Court first turns to the language of the choice-of-law provision, which reads: “THE AGREEMENT MUST BE CONSTRUED AS IF BOTH PARTIES JOINTLY WROTE IT AND GOVERNED BY CALIFORNIA LAW EXCEPT FOR ITS CONFLICTS OF LAWS PRINCIPLES.” AdWords Agreement ¶ 9. Google asserts that the scope of this clause only encompasses California contract law and leaves out non-contractual California consumer protection laws such as the UCL. CC Opp. 33 n.12. The California Supreme Court, however, does not interpret choice-of-law clauses so narrowly, explaining that the specific choice-of-law phrase used here (“governed by California law”) is an especially “broad one signifying a relationship of absolute direction, control, and restraint.” Nedlloyd, 3 Cal.4th at 469; see also Wash. Mut. Bank, 24 Cal.4th at 916 (“[W]hen a rational businessperson . . . provides that disputes arising from the agreement shall be governed by the law of an identified jurisdiction, the logical conclusion is that he or she intended that law to apply to all disputes arising out of the transaction or relationship.”). Here, the AdWords Agreement expressly purports to “govern [advertisers'] participation in Google's advertising program(s) . . . and/or [advertisers'] online management of any advertising campaigns,” which would appear to include the Location Targeting features on its face. AdWords Agreement 1 (emphasis added). Given such broad language in both the contract's agreed-upon scope and its choice-of-law clause, the Court finds that the Location Targeting UCL claim falls within the scope of the AdWords Agreement relationship.

Google cites several cases in support of its proposition that choice-of-law contract provisions do not import the selected jurisdiction's consumer protection laws. CC Opp. 33 n.12. However, these cases all involved clauses in contracts that did not cover the injury at issue or clauses that selected California law for only narrow applications. See, e.g., Darisse v. Nest Labs, Inc., 2016 WL 4385849, at *8 (N.D. Cal. Aug. 15, 2016) (declining to enforce choice-of-law clause where the contract at issue expressly did not cover the purchase in dispute); Frenzel v. AliphCom, 76 F.Supp.3d 999, 1009 (N.D. Cal. 2014) (declining to enforce choice-of-law clause in the terms of use for defendant's website while the UCL claim related to plaintiff's purchase of defendant's product); In re Sony Gaming Networks & Customer Data Sec. Breach Litig., 903 F.Supp.2d 942, 965 (S.D. Cal. 2012) (declining to apply choice-of-law clause where, “[b]y its own terms, [] the provision dictates only that California law applies to the construction and interpretation of the contract”) (emphasis added).

Moreover, the Ninth Circuit has held that, “[i]f a state law does not have limitations on its geographical scope, courts will apply it to a contract governed by that state's law, even if parts of the contract are performed outside of the state.” Gravquick A/S v. Trimble Navigation Int'l Ltd., 323 F.3d 1219, 1223 (9th Cir. 2003). Google has not identified-nor has the Court been able to identify-any “express requirement [in the UCL] limiting its protection to [consumers] located in California.”²⁵ Id. This is consistent with the various courts that have rejected arguments like Google's with some regularity, finding that the UCL applied where parties had selected California law to govern their contract. See, e.g., Maldonado v. Apple, Inc, 2021 WL 1947512, at *7 (N.D. Cal. May 14, 2021) (rejecting argument that choice-of-law clause did not include “state statutory law” and finding instead that “the parties contracted to apply the law of California to disputes arising from their [] contracts and these particular UCL . . . claims do”); Underwood v. Future Income Payments, LLC, 2018 WL 4964333, at *12 (C.D. Cal. Apr. 26, 2018) (finding that the UCL applies where the contract selected California law and defendants operated out of California). Accordingly, Google's narrow reading of the choice-of-law clause to exclude nonCalifornia class members from the UCL claim is unpersuasive.

25 Although the UCL (like most California law) is subject to a general presumption against extraterritoriality, see Sullivan v. Oracle Corp., 51 Cal.4th 1191, 1207 (2011), Google does not argue that extraterritoriality would bar the application of UCL in this case. In any event, “[h]onoring [a] choice of law [clause] does not give extraterritorial application to the statute, even if the contract was performed partially outside of California, so long as the contract itself was properly governed by California law.” Gravquick, 323 F.3d at 1222 (citing Foreman v. George Foreman Assocs., Ltd., 517 F.2d 354, 356 (9th Cir. 1975) (rejecting argument that enforcing a choice-of-law clause “gives extraterritorial effect to California law”)); see also Leonard v. Coll. Network, Inc., 2004 WL 2944050, at *6 (N.D. Cal. Dec. 17, 2004) (same).

Having found that the UCL claim falls within the scope of the choice-of-law clause, the Court proceeds to consider whether the selected jurisdiction has a “substantial relationship to the parties or their transaction” and whether the selected law is “contrary to a fundamental policy of California.” Wash. Mut. Bank, 24 Cal.4th at 915-16. As to the first requirement, California plainly has a substantial relationship to the parties, given that Google is headquartered in Mountain View, California. Compl. ¶¶ 21-23. As to the second, it is self-evident that California law is not contrary to itself. Accordingly, the Court is satisfied that California has a substantial relationship to the parties in this case and there is no conflict with fundamental California policy.

Because the choice-of-law clause in the AdWords Agreement satisfies the multi-part Nedlloyd test, the Court finds that the clause is enforceable and that the UCL governs Google's challenged location targeting practices under the Agreement. Accordingly, the certification of the nationwide Location Targeting Class would not require individual inquiries into the several states' consumer protection laws.

v. Damages Models

Finally, as part of their Rule 23(b)(3) obligation to demonstrate that common questions predominate, plaintiffs seeking class certification must present a model of damages that “measure[s] only those damages attributable to [their] theory.” Comcast, 569 U.S. at 35; see also id. (“If the model does not even attempt to do that, it cannot possibly establish that damages are susceptible of measurement across the entire class for purposes of Rule 23(b)(3).”).

The Ninth Circuit has not construed Comcast as requiring plaintiffs to proffer a class-wide method for calculating damages at the class certification stage, Pulaski, 802 F.3d at 987-88 (collecting circuit cases); instead, it has understood Comcast as requiring plaintiffs to demonstrate a logical connection between their damages model and their theory of liability. Id. (“[P]laintiffs must be able to show that their damages stemmed from the defendant's actions that created the legal liability.”). The Ninth Circuit has also consistently reaffirmed that “differences in damage calculations do not defeat class certification.” Id. at 988.

With respect to their obligations under Comcast, RMC offers three different methods for calculating damages that are attributable to their theory of UCL fraud liability: (1) the “partial refund” method; (2) “Smart Pricing” restitution (“SPR”); and (3) restitutionary disgorgement of profits. CC Mot. 28-34. The “partial refund” method calculates damages as the full amount class members paid for out-of-area clicks. Id. at 29; see also 11/8/2018 Suppl. Solomon Rep. ¶ 5. The alternative SPR method calculates damages by starting with the amount class members paid for out-of-area clicks (i.e., the “partial refund” amount) and deducting “the amount Google's Smart Pricing algorithms estimate that a rational advertiser might have been willing to pay for these mistargeted clicks.” 11/8/2018 Suppl. Solomon Rep. ¶ 5. This method purports to track the restitution method approved by the Ninth Circuit in Pulaski by accounting for the residual value class members may have received from paying for an out-of-area click. See CC Mot. 31-32. Finally, RMC proposes that damages may also be calculated by restitutionary disgorgement of profits, which is “the amount of net profit Google made directly from the Out-of-Area Clicks” estimated by Plaintiffs' expert. 11/8/2018 Suppl. Solomon Rep. ¶ 5.

The Court finds that RMC's proposed theories of restitution satisfy Comcast's directive for damages models to measure only those attributable to plaintiffs' theory of liability, i.e., that Google's Settings Screen statement deceived class members into believing that it would only charge for in-area clicks. Even the most generous model of restitution-the partial refund model-would only award class members for their payments on clicks originating from outside the locations they had specifically selected, which is precisely what RMC claims they were deceived into paying. Whether an advertiser should receive as damages the full amount that she paid or whether some portion of that amount should be deducted as value received (i.e., SPR Model) or as Google's costs (i.e., restitutionary disgorgement of profits) does not alter the fact that the base restitution amount is derived from and tracks RMC's theory of liability regarding out-ofarea clicks. This nexus between RMC's restitution models and its theory of liability also distinguishes the case here from others where courts have rejected damages models under Comcast because the damages models were not consistent with the liability case. See, e.g., Arris, 327 F.R.D. at 369-70 (“[T]his Court and other courts in this district have found that damages models fail under Comcast where the model is not consistent with the liability case.”) (collecting cases). “At this stage, Plaintiffs need only show that such damages can be determined without excessive difficulty and attributed to their theory of liability,” Just Film, 847 F.3d at 1121, and the Court finds that RMC's proffered restitution models have done so to Comcast's satisfaction.

Google criticizes RMC's damages models for their failure to “segregate those advertisers who benefited from [the out-of-area clicks] from any who might have been harmed.” CC Mot. 25. It also contends that the Court may not certify a class that includes uninjured individuals. CC Mot. 25; Suppl. CC Mot. 11-17. With respect to the first point, the Court has explained that any benefits that certain class members eventually received from out-of-area clicks do not impact UCL restitution, which is “based on what a purchaser would have paid at the time of purchase had the purchaser received all the information.” Pulaski, 802 F.3d at 989 (emphasis added). The purported failure to segregate class members who had received benefits from out-of-area clicks, therefore, cannot defeat certification.

As to Google's second point regarding uninjured individuals, the Ninth Circuit has expressly “overrule[d] the statement . . . that ‘no class may be certified that contains members lacking Article III standing,'” remarking that this requirement “does not apply when a court is certifying a class seeking injunctive or other equitable relief” and where the class representative has standing. Olean, 31 F.4th at 682 n.32; cf. TransUnion LLC v. Ramirez, 141 S.Ct. 2190, 2208 (2021) (expressly declining to “address the distinct question whether every class member must demonstrate standing before a court certifies a class”) (italics in original). In short, even if the Court were to agree with Google that some class members were not injured by the LT Settings Screen statement, such a finding would not compel the Court to decline class certification.

Google's remaining objections to RMC's proposed theories of restitution are also unavailing. For instance, to the extent Google argues that the “partial refund” method arbitrarily ascribed a value of $0 to out-of-area clicks, that argument is a “merits argument about the proper amount of damages, not a mismatch between Plaintiffs' damages model and theory of liability.” Arris, 327 F.R.D. at 370. Google asserts the same argument to oppose RMC's restitutionary disgorgement of profits, though it also suggests that RMC's disgorgement model is actually non-restitutionary. CC Opp. 26. With respect to the latter point, the Court does not consider RMC's theory of disgorgement of profits to be non-restitutionary because RMC is only seeking disgorgement of monies that class members themselves have paid to Google for out-of-area clicks. See Korea Supply Co. v. Lockheed Martin Corp., 29 Cal.4th 1134, 1148 (2003) (“Under the UCL, an individual may recover profits unfairly obtained to the extent that these profits represent monies given to the defendant or benefits in which the plaintiff has an ownership interest.”). And finally, Google objects to the SPR method of calculating restitution by summarily asserting that, “because the conversion rate for out-of-area clicks is generally higher per dollar spent,” this method “would generally yield a higher value for out-of-area clicks.” CC Mot. 26 n.9. This, however, is unsupported by any evidence suggesting Smart Pricing would result in a higher value than what class members actually paid. In any event, this argument also appears to be contesting the proper amount of damages, as opposed to the threshold question of whether the SPR method is congruent with RMC's theory of UCL fraud liability.

Accordingly, the Court finds that RMC's proposed restitution calculations satisfy its obligations under Comcast to provide a damages model that measures only those damages attributable to its theory of liability.

* * *

Having considered the questions common to the Location Targeting Class, as well as the individualized inquiries that Google has identified, the Court finds that the central and common question of UCL liability predominates over potential individualized inquiries such as injury and exposure to Google's misrepresentations.

b. Superiority

RMC also must demonstrate that “a class action is superior to other available methods.” Fed.R.Civ.P. 23(b)(3). Rule 23 instructs courts to consider four factors when assessing superiority: “(A) the class members' interests in individually controlling the prosecution or defense of separate actions; (B) the extent and nature of any litigation concerning the controversy already begun by or against class members; (C) the desirability or undesirability of concentrating the litigation of the claims in the particular forum; and (D) the likely difficulties in managing a class action.” Fed.R.Civ.P. 23(b)(3).

The Court finds that RMC has also satisfied this requirement, as a class action is the superior method of handling the claims of individual advertisers. As an example, the restitutionary award for RMC on the Location Targeting UCL claim would only reach a maximum of $88.73. No reasonable advertiser is likely to pursue these claims on his or her own, given the cost and other resources required to litigate against a company like Google, as evidenced by the longevity of this very case. These factors all “vividly point[] to the need for class treatment.” Just Film, 847 F.3d at 1123.

* * *

Having reviewed the evidence presented and considered the arguments asserted regarding the certification of the Location Targeting Class, the Court finds that Plaintiffs have satisfied the four requirements of Rule 23(a). Furthermore, the Court finds that questions common to the Location Targeting Class predominate over any questions affecting only individual members, and that a class action is superior to other methods of adjudicating the parties' dispute. Accordingly, the Court finds that the Location Targeting Class may be certified under Rule 23(b)(3).

VI. CONCLUSION

For the reasons above, the Court DENIES Google's motion for summary judgment and motion to strike expert reports. The Court also GRANTS IN PART and DENIES IN PART Plaintiffs' motion for class certification. Specifically, the Court certifies the Location Targeting Class and Search Bundled Clicks Subclass. However, Plaintiffs have failed to meet their burden of showing that common questions predominate as to the Non-Smart Priced Clicks, AFMA Clicks, and mGDN Clicks Subclasses, so the Court declines to certify those subclasses. The Court's denial as to those three subclasses is without prejudice, and Plaintiffs may renew their motion for class certification after addressing the issues identified in this Order.

IT IS SO ORDERED.