Justin Lytle v. Nutramax Laboratories, Inc.

Opinion

FOR PUBLICATION

UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT

JUSTIN LYTLE and CHRISTINE No. 22-55744 MUSTHALER, D.C. No. Plaintiffs-Appellees, 5:19-cv-00835- v. FMO-SP

NUTRAMAX LABORATORIES, ORDER AND INC. AND NUTRAMAX AMENDED LABORATORIES VETERINARY OPINION SCIENCES, INC,

Defendants-Appellants.

Appeal from the United States District Court for the Central District of California Fernando M. Olguin, District Judge, Presiding

Argued and Submitted October 18, 2023 Pasadena, California

Filed April 22, 2024 Amended August 23, 2024

Before: A. Wallace Tashima and Holly A. Thomas, Circuit Judges, and Jed S. Rakoff, * District Judge.

* The Honorable Jed S. Rakoff, United States District Judge for the Southern District of New York, sitting by designation. 2 LYTLE V. NUTRAMAX LABORATORIES, INC.

Opinion by Judge Rakoff

SUMMARY **

Class Action

The panel affirmed the district court’s grant of class certification in a consumer class action concerning the marketing of the pet health product Cosequin. Plaintiffs claim that Nutramax violated the California Consumers Legal Remedies Act (“CLRA”) by marketing Cosequin as promoting healthy joints in dogs, when in fact Cosequin provides no such health benefit. The district court certified a class of California purchasers of certain Cosequin products who were exposed to the allegedly misleading statements. Nutramax challenged the district court’s reliance upon the proposed damages model of Plaintiffs’ expert, Dr. Jean- Pierre Dubé, to find that common questions predominated as to injury. The panel held that, contrary to Nutramax’s contention, there was no general requirement that an expert actually apply to the proposed class an otherwise reliable damages model in order to demonstrate that damages are susceptible to common proof at the class certification stage. Rather, class certification plaintiffs may rely on an unexecuted damages model to show that damages are susceptible to common proof. The panel concluded that the

** This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader. LYTLE V. NUTRAMAX LABORATORIES, INC. 3

district court did not abuse its discretion in finding that Dr. Dubé’s proposed model was sufficiently sound and developed to satisfy this standard at the class certification stage. The panel rejected Nutramax’s contention that the district court incorrectly concluded that the element of reliance was susceptible to common proof. The district court properly found that classwide reliance may be established under the CLRA through proof that a misrepresentation is material. While the presumption of reliance is rebuttable, the district court did not abuse its discretion in concluding that Nutramax failed to rebut the presumption here.

COUNSEL

Mark R. Sigmon (argued), Milberg Coleman Bryson Phillips Grossman PLLC, Raleigh, North Carolina; Greg F. Coleman and Adam A. Edwards, Milberg Coleman Bryson Phillips Grossman PLLC, Knoxville, Tennessee; Michael H. Pearson and Daniel L. Warshaw, Pearson Simon & Warshaw LLP, Sherman Oaks, California; Matthew D. Schultz, Levin Papantonio Rafferty Proctor Buchanan O'Brien Barr & Mougey PA; Pensacola, Florida; for Plaintiffs-Appellees. David R. Carpenter (argued), Nicole M. Baade, Sean A. Commons, and Amy P. Lally, Sidley Austin LLP, Los Angeles, California; Joshua A. Glikin, Shulman Rogers PA, Potomac, Maryland; for Defendants-Appellants. 4 LYTLE V. NUTRAMAX LABORATORIES, INC.

ORDER

The opinion filed on April 22, 2024, and appearing at 99 F.4th 557, is hereby amended. The amended opinion will be filed concurrently with this order. With these amendments, the panel has unanimously voted to deny the petition for panel rehearing. Judge H.A. Thomas has voted to deny the petition for rehearing en banc, and Judge Tashima and Judge Rakoff so recommend. The full court was advised of the petition for rehearing en banc. A judge requested a vote on whether to rehear the matter en banc. The matter failed to receive a majority of the votes of the nonrecused active judges in favor of en banc consideration. Fed. R. App. P. 35(a). The petition for panel rehearing and the petition for rehearing en banc, Dkt. No. 52, are DENIED.

OPINION

RAKOFF, District Judge:

This is a putative consumer class action concerning the marketing of the pet health product Cosequin. Plaintiffs- Appellees claim that Nutramax Laboratories, Inc. and Nutramax Laboratories Veterinary Sciences, Inc. (collectively, “Nutramax”) violated the California Consumers Legal Remedies Act (“CLRA”), Cal. Civil Code §§ 1750–1784, by marketing Cosequin as promoting healthy joints in dogs, when in fact Cosequin provided no such health benefits. Below, the district court certified a class of California purchasers of certain Cosequin products who LYTLE V. NUTRAMAX LABORATORIES, INC. 5

were exposed to the allegedly misleading statements. Nutramax now appeals that grant of class certification on two grounds. First, Nutramax challenges the district court’s reliance upon the proposed damages model of Plaintiffs’ expert, Dr. Jean-Pierre Dubé, to find that common questions predominated as to injury. Nutramax claims this was error because the proposed model had not actually been applied to the proposed class. We conclude that, contrary to Nutramax’s contention, there is no general requirement that an expert actually apply to the proposed class an otherwise reliable damages model in order to demonstrate that damages are susceptible to common proof at the class certification stage. Rather, we hold that class action plaintiffs may rely on a reliable though not-yet-executed damages model to demonstrate that damages are susceptible to common proof so long as the district court finds that the model is reliable and, if applied to the proposed class, will be able to calculate damages in a manner common to the class at trial. We further conclude that the district court did not abuse its discretion in finding Dr. Dubé’s proposed model was sufficiently sound and developed to satisfy this standard at the class certification stage. Second, Nutramax contends that the district court incorrectly concluded that the element of reliance was susceptible to common proof. We disagree. The district court properly found that classwide reliance may be established under the CLRA through proof that a misrepresentation is material. While the presumption of reliance is rebuttable, the district court did not abuse its discretion in concluding Nutramax had failed to rebut the presumption here. 6 LYTLE V. NUTRAMAX LABORATORIES, INC.

Accordingly, for the reasons set forth more fully below, we affirm the district court’s grant of class certification. I. Nutramax develops and sells pet health supplements. Plaintiffs-Appellees Justin Lytle and Christine Musthaler are two dog owners who purchased a product produced by Nutramax, Cosequin, for their dogs. In this action, Plaintiffs allege that Nutramax marketed Cosequin as a health supplement that would improve their dogs’ joints and mobility when, in fact, there is no evidence that Cosequin provides any such health benefit. After the close of fact and expert discovery, Plaintiffs sought to certify the following class pursuant to Federal Rule of Civil Procedure (“FRCP”) 23(b)(3):

All persons residing in California who purchased during the limitations period the following canine Cosequin products for personal use: Cosequin DS Maximum Strength Chewable Tablets; Cosequin DS Maximum Strength Plus MSM Chewable Tablets; and Cosequin DS Maximum Strength Plus MSM Soft Chews.

Plaintiffs initially asserted that numerous statements made in marketing materials for Cosequin and/or on the packaging of the three products listed above were false and misleading. However, at class certification Plaintiffs narrowed their claims to four statements that appeared on Cosequin’s packaging:

(1): “Joint Health Supplement”; LYTLE V. NUTRAMAX LABORATORIES, INC. 7

(2): “Use Cosequin to help your pet Climb stairs, Rise and Jump!”; (3): “Supports Mobility for a Healthy Lifestyle”; and (4): “Mobility, Cartilage and Joint Health Support.”

Plaintiffs argued that these statements were false and misleading because Cosequin does not, in fact, improve dogs’ joint health. According to Plaintiffs, the only two peer- reviewed, double-blinded, randomized controlled trials that have been conducted on Cosequin’s efficacy have concluded that Cosequin confers no more benefit to canine joint health than a placebo. Plaintiffs’ veterinary expert Dr. Steven Budsberg opined that “the strongest available evidence . . . has consistently found no reliable evidence that Cosequin provides any efficacy in supporting, maintaining, or improving joint health,” and Plaintiffs’ biostatistics expert Dr. Richard Evans concluded that “there is no evidence that [glucosamine and chondroitin, the active ingredients in Cosequin, have] a greater prophylactic effect than placebo control on maintaining joint health in healthy pet dogs.” Nutramax, for its part, presented contrary evidence about the products’ efficacy, including non-randomized control trials and the testimony of Nutramax’s own veterinary expert, Dr. Marcellin-Little. Nutramax further argues that the two studies relied on by Plaintiffs involved dogs with osteoarthritis, even though Cosequin is not marketed to treat that condition. The district court found that Plaintiffs had adequately demonstrated for class certification purposes the reasonable likelihood that a jury could find that Cosequin provided no benefit to dogs’ joint health. 8 LYTLE V. NUTRAMAX LABORATORIES, INC.

The four challenged statements appeared on the labels of various Cosequin products in various combinations over time. Although the parties offer differing characterizations about the extent and significance of these variations, the parties agree that the first statement (“Joint Health Supplement”) appeared on all product labels or packages throughout the class period. Further, approximately 90% of sales during the class period—including those to the named plaintiffs—were attributable to products that contained the “Joint Health Supplement” statement. The district court ultimately found that the presence of the Joint Health Supplement statement on every challenged product was sufficient to support class certification. Before the district court, Nutramax raised a host of challenges to Plaintiffs’ class certification motion, contesting each prerequisite of FRCP 23(a) and 23(b)(3). Nutramax also belatedly filed two evidentiary objections to Plaintiffs’ expert witnesses Bruce Silverman and Dr. Dubé . The district court rejected each of these objections and certified the class as requested by Plaintiffs. Nutramax then promptly sought and obtained permission to file the instant interlocutory appeal of the district court’s class certification order pursuant to FRCP 23(f). On appeal, Nutramax limits its challenge to just two of the district court’s findings: (1) that damages are susceptible to common proof, and (2) that causation/reliance is susceptible to common proof. Nutramax also persists in its evidentiary objection to Dr. Dubé insofar as it relates to the first question of whether damages are susceptible to common proof. The parties’ briefing and the district court’s rulings on these two issues are briefly described below. LYTLE V. NUTRAMAX LABORATORIES, INC. 9

A. To establish damages on a classwide basis, Plaintiffs put forward the testimony of Dr. Dubé, a professor of marketing at the University of Chicago Booth School of Business who, as the district court found, has “extensive experience in marketing data and analytics.” To measure classwide damages, Dr. Dubé proposed to conduct a “conjoint survey” (or “conjoint analysis”). Simply put, a conjoint survey allows a researcher to test the economic value a consumer places on a given product feature, such as a particular statement on a package, by showing the product to individual survey participants with and without certain features, and then using survey responses to calculate the economic value consumers place upon the feature. Dr. Dubé explained that, in conducting a conjoint analysis, a researcher is able to control for other variables such as package size and the competing products by modifying the underlying choice- tasks presented to participants. At the time of class certification, Dr. Dubé had not yet actually applied his conjoint analysis to a representative sample in the case, that is, he had not actually surveyed any class members or calculated what class members’ damages might be. Rather, in his report, Dr. Dubé opined that, “[b]ased on [his] analysis of market data and other marketing documents made available through discovery,” he believed a conjoint analysis “is well-suited to the facts of this case and will successfully isolate the economic damages associated with the Challenged Claims.” Dr. Dubé acknowledged, however, that he would not know whether the class actually suffered any damages until he actually executed his survey. Nutramax’s class certification opposition criticized Dr. Dubé for failing to actually conduct (i.e., “execute”) the 10 LYTLE V. NUTRAMAX LABORATORIES, INC.

survey and thus complete his analysis. Before the district court, Nutramax also argued that Dr. Dubé’s model was under-developed in other respects, although Nutramax apparently did not argue this was a basis to exclude his testimony under Federal Rule of Evidence (“FRE”) 702 or FRCP 23, but instead argued that his opinion violated FRCP 16, 26, 37 and the district court’s scheduling order. The district court rejected Nutramax’s argument that it could not rely on Dr. Dubé’s unexecuted damages model or that the model should be excluded under the standard set forth in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). The district court began by noting Dr. Dubé’s credentials, finding him to be well qualified for the purposes offered. The district court further observed that conjoint surveys such as that proposed by Dr. Dubé “are a well-established method for measuring class-wide damages in CLRA mislabeling cases,” and also noted that conjoint analyses previously prepared by Dr. Dubé had been upheld as reliable in similar cases. Most importantly, the court below, citing other district court precedent, held that “[a] plaintiff is not required to actually execute a proposed conjoint analysis to show that damages are capable of determination on a class-wide basis.” B. Nutramax also opposed class certification on the ground that individual questions predominated with respect to the element of reliance because (1) substantial variation in the labels precluded a finding of predominance, (2) Plaintiffs had failed to show common exposure to the challenged statements, and (3) Plaintiffs had not shown the challenged statements factored into the typical consumer’s purchasing decision or would have been material to such consumers. LYTLE V. NUTRAMAX LABORATORIES, INC. 11

Although reliance and materiality are separate elements, under the CLRA, a plaintiff can create a presumption of reliance by demonstrating a material misrepresentation was made to the entire class. See Stearns v. Ticketmaster Corp., 655 F.3d 1013, 1022 (9th Cir. 2011). To establish materiality (and relatedly, reliance for class certification purposes), Plaintiffs put forward the expert testimony of Silverman, a former advertising executive, who testified that a reasonable consumer would find the product labels misleading. As noted, Plaintiffs also put forward two expert witnesses, Dr. Budsberg and Dr. Evans, to testify to the falsity of the challenged statements and demonstrate that Cosequin did not offer joint health benefits. Plaintiffs also argued that the survey conducted by Nutramax’s own expert, Dr. Carol Scott, actually supported a finding of materiality. 1 In response, Nutramax put forward two experts. First, Nutramax put forward Dr. Scott to present the results of a consumer survey about the purchasing decisions of Cosequin customers. Dr. Scott’s survey found that purchasers “consulted a variety of information sources prior to purchasing [Cosequin] for the first time, with one’s veterinarian being most frequently mentioned (i.e., 57% of respondents . . . ), followed by website ratings or reviews

1 In this regard, Plaintiffs homed in on one particular finding of Dr. Scott’s survey that they claimed actually supports a finding of materiality. Dr. Scott showed respondents a label from Product #2, containing only the Joint Health Supplement statement, and asked: “Based on this package, what do you think[] Cosequin will do for your dog? That is, why should you give your dog Cosequin?” The “most frequently given response” (by nearly 80% of respondents) was “improve/help/maintain mobility, flexibility, joint health/support.” In the district court and on appeal, Plaintiffs argue that this finding by Dr. Scott further demonstrates that the challenged statements would have been misleading to a reasonable consumer. 12 LYTLE V. NUTRAMAX LABORATORIES, INC.

(26.9% of respondents).” Only about a quarter of respondents identified the packaging as a source of information they consulted prior to deciding to purchase Cosequin for the first-time. Dr. Scott’s survey found that the most common motivations for purchasing Cosequin were advice from a veterinarian, advice from someone else, or research on websites, whereas only a small fraction of respondents cited the packaging for their motive. Second, Nutramax put forward Dr. Olivier Toubia, who conducted a consumer choice survey somewhat similar to that proposed by Dr. Dubé. Dr. Toubia showed one group of individuals the actual product packaging and another group a modified version that (supposedly) removed the challenged claims. The results of this study, according to Dr. Toubia, showed that removing the statements did not materially impact the price a consumer would be willing to pay. The district court reviewed the evidence presented by the parties and ultimately credited the evidence presented by Plaintiffs that the challenged statements would be materially misleading to a reasonable consumer. In a footnote, the district court explained that it was “unpersuaded” by Dr. Scott’s and Dr. Toubia’s expert reports, noting that they contained “flaws that undercut their persuasiveness.” In particular, the district court cited the fact that Dr. Toubia’s survey failed to remove the “Joint Health Supplement” statement from the packaging, even though that was at the core of Plaintiffs’ claim. The district court did not explain the basis for its rejection of Dr. Scott’s survey, but ultimately concluded Nutramax’s evidence was “outweighed by the common evidence presented by plaintiffs.” LYTLE V. NUTRAMAX LABORATORIES, INC. 13

II. Before a class may be certified, the district court must conduct a “rigorous analysis” to determine if the prerequisites of FRCP 23 have been satisfied. Olean Wholesale Grocery Coop., Inc. v. Bumble Bee Foods LLC, 31 F.4th 651, 664 (9th Cir. 2022) (en banc). The plaintiffs must “actually prove—not simply plead—that their proposed class satisfies each requirement of Rule 23.” Halliburton Co. v. Erica P. John Fund, Inc., 573 U.S. 258, 275 (2014). The plaintiffs bear “the burden of establishing that the prerequisites of Rule 23 are satisfied by a preponderance of the evidence.” Olean, 31 F.4th at 665. At issue here is the predominance requirement: that “questions of law or fact common to class members predominate over any questions affecting only individual members.” FRCP 23(b)(3). This requirement presupposes satisfaction of the commonality requirement of FRCP 23(a)(2), which itself tests “the capacity of a classwide proceeding to generate common answers apt to drive the resolution of the litigation.” Alcantar v. Hobart Serv., 800 F.3d 1047, 1052 (9th Cir. 2015) (quoting Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350 (2011)). But the predominance inquiry goes further and “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). At the same time, it is critical to keep in mind that class certification is different from summary judgment. “A court . . . is merely to decide [whether a class action is] a suitable method of adjudicating the case.” Edwards v. First Am. Corp., 798 F.3d 1172, 1178 (9th Cir. 2015). With 14 LYTLE V. NUTRAMAX LABORATORIES, INC.

respect to the predominance inquiry specifically, a district court must evaluate “‘the method or methods by which plaintiffs propose to use the [class-wide] evidence to prove’ the common question in one stroke.” Olean, 31 F.4th at 666 (quoting In re Hydrogen Peroxide Antitrust Litig., 552 F.3d 305, 312 (3d Cir. 2008)). “In determining whether the ‘common question’ prerequisite is met, a district court is limited to resolving whether the evidence establishes that a common question is capable of class-wide resolution, not whether the evidence in fact establishes that plaintiffs would win at trial.” Olean, 31 F.4th at 666–67. We review the “decision to certify a class and any particular underlying Rule 23 determination involving a discretionary determination” for an abuse of discretion. Olean, 31 F.4th at 663 (cleaned up). We review de novo “the district court’s determination of underlying legal questions.” Id. “A district court applying the correct legal standard abuses its discretion only if it (1) relies on an improper factor, (2) omits a substantial factor, or (3) commits a clear error of judgment in weighing the correct mix of factors.” Sali v. Corona Reg’l Med. Ctr., 909 F.3d 996, 1004 (9th Cir. 2018) (cleaned up). We review evidentiary rulings for an abuse of discretion. City of Pomona v. SQM N. Am. Corp., 750 F.3d 1036, 1043 (9th Cir. 2014). We review factual findings underlying a class certification ruling for clear error. Sali, 909 F.3d at 1002. We review the question of whether an expert’s damages model “is capable of showing class-wide impact, thus satisfying one of the prerequisites of Rule 23(b)(3) of the Federal Rules of Civil Procedure, for an abuse of discretion.” Olean, 31 F.4th at 663. LYTLE V. NUTRAMAX LABORATORIES, INC. 15

III. A. Nutramax’s principal argument on appeal is that the “rigorous analysis” required by FRCP 23 categorically prohibits a class-action plaintiff from relying on an unexecuted damages model to demonstrate predominance (at least where that model is the only evidence of classwide injury). 2 In addressing this argument, the parties focus on the statement in the Supreme Court’s decision in Comcast Corp. v. Behrend—repeated in our en banc decision in Olean Wholesale Grocery Cooperative, Inc. v. Bumble Bee Foods LLC—that a class action plaintiff must “establish[] that damages are capable of measurement on a classwide basis.” 569 U.S. 27, 34 (2013) (emphasis added). Plaintiffs contend that this phrase sanctions precisely what Dr. Dubé has done here: submit an expert report opining that damages can be measured on a classwide basis and setting forth a reliable method for doing so. Nutramax, by contrast, argues that to satisfy the predominance requirement, “[P]laintiffs must proffer admissible, affirmative evidence that classwide injury and damages in fact are capable of classwide measurement.” In Nutramax’s view, the only way to carry this burden is to actually put forward common evidence showing that classwide damages exist.

2 On appeal, the parties both appear to assume that the predominance requirement would be satisfied only if damages are capable of measurement on a classwide basis. For purposes of this appeal, we adopt this assumption. We note, however, that we have held that individual questions of damages do not necessarily defeat class certification, as the district court here expressly acknowledged. See Vaquero v. Ashley Furniture Indus., Inc., 824 F.3d 1150, 1155 (9th Cir. 2016); Pulaski & Middleman, LLC v. Google, Inc., 802 F.3d 979, 987 (9th Cir. 2015). 16 LYTLE V. NUTRAMAX LABORATORIES, INC.

Nutramax’s argument rests upon a misapprehension of the temporal focus of the class certification inquiry. As explained below, class action plaintiffs are not required to actually prove their case through common proof at the class certification stage. Rather, plaintiffs must show that they will be able to prove their case through common proof at trial. Given, moreover, that the Federal Rules contemplate that certification will be made “[a]t an early practicable time,” FRCP 23(c)(1)(A), we see no reason why plaintiffs may not, in appropriate circumstances, satisfy this burden through a proffer of a reliable method of obtaining evidence that will come into existence once a damages model is executed, even when the results are not yet available at the class certification stage. We thus hold that class action plaintiffs may rely on an unexecuted damages model to demonstrate that damages are susceptible to common proof so long as the district court finds, by a preponderance of the evidence, that the model will be able to reliably calculate damages in a manner common to the class at trial. Contrary to Nutramax’s contention, there is no requirement that the evidence relied upon by Plaintiffs to support class certification be presented in an admissible form at the class certification stage. See Sali, 909 F.3d at 1004 (“Inadmissibility alone is not a proper basis to reject evidence submitted in support of class certification.”); B.K. v. Snyder, 922 F.3d 957, 974 (9th Cir. 2019) (“At this ‘tentative, preliminary, and limited’ stage we have held strictly admissible evidence is not required, and we have indicated that plaintiffs can meet their evidentiary burden in part through allegations when the allegations are detailed and supported by additional materials” (internal citation omitted)). As we explained in Sali v. Corona Regional Medical Center, “an inquiry into the evidence’s ultimate LYTLE V. NUTRAMAX LABORATORIES, INC. 17

admissibility should go to the weight that evidence is given at the class certification stage.” 909 F.3d at 1006. 3 Of course, if it is unlikely that a particular piece of common proof will be available or admissible at trial, that possibility weighs against a finding that common questions (and common answers) will predominate. But “[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 which apparently satisfies” Rule 23. Id. at 1004–05 (quoting Blackie v. Barrack, 524 F.2d 891, 901 (9th Cir. 1975)). Nor is there a requirement that class action plaintiffs actually prove that classwide damages exist in order to obtain class certification. Rather, we have repeatedly found class treatment to be appropriate, in analogous contexts, based upon a showing that damages could be calculated on a classwide basis, even where such calculations have not yet been performed. See Just Film, Inc. v. Buono, 847 F.3d 1108, 1121 (9th Cir. 2017) (“At this stage, Plaintiffs need only show that such damages can be determined without excessive difficulty and attributed to their theory of liability, and have proposed as much here.”); Lambert v. Nutraceutical Corp., 870 F.3d 1170, 1182 (9th Cir. 2017) (“Uncertainty regarding class members’ damages does not prevent certification of a class as long as a valid method has been proposed for calculating those damages.”), rev’d and remanded on other grounds, 139 S. Ct. 710 (2019); Leyva v.

3 The district court correctly cited Sali for this exact principle. Nutramax inexplicably asserts the district court committed legal error by doing so without ever discussing Sali or explaining why, in Nutramax’s view, the rule announced in that case is inapplicable. 18 LYTLE V. NUTRAMAX LABORATORIES, INC.

Medline Indus. Inc., 716 F.3d 510, 514 (9th Cir. 2013) (finding evidence “that damages could feasibly and efficiently be calculated once the common liability questions are adjudicated” was sufficient to satisfy predominance). For example, in Lambert v. Nutraceutical Corp., we reversed a district court’s denial of class certification and concluded the plaintiffs had satisfied the predominance requirement by proposing a valid damages model, even where the plaintiffs did not yet have all of the data necessary to perform their damages calculations. 870 F.3d at 1183–84. We explained that the “precise [data] is unnecessary for class certification” because “the question is only whether [plaintiff] has presented a workable method.” Id. at 1184. Requiring that class action plaintiffs actually prove classwide injury at this stage would improperly conflate the class certification inquiry with the merits. To be sure, courts may not avoid resolving questions pertinent to class certification merely because they overlap with the merits of the plaintiffs’ case, see Wal-Mart, 564 U.S. at 351; Comcast, 569 U.S. at 33–34, but this does not “grant[] courts . . . license to engage in free-ranging merits inquiries at the certification stage.” Amgen Inc. v. Conn. Ret. Plans & Tr. Funds, 568 U.S. 455, 466 (2013). Rather, such “[m]erits 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.” Id. While acknowledging this rule, Nutramax argues that the theoretical possibility Dr. Dubé’s model, once executed, might show no injury at all “goes to the heart of the Rule 23 inquiry.” But Nutramax fails to convincingly explain why this is so. The focus of the predominance inquiry “is whether the method of proof would apply in common to all class members,” “not whether the method of proof would or could LYTLE V. NUTRAMAX LABORATORIES, INC. 19

prevail.” In re Celexa & Lexapro Mktg. & Sales Pracs. Litig., 915 F.3d 1, 12 (1st Cir. 2019); see Stockwell v. City & Cnty. of San Francisco, 749 F.3d 1107, 1112 (9th Cir. 2014) (“[W]hether class members could actually prevail on the merits of their claims is not a proper inquiry in determining the preliminary question whether common questions exist.” (internal quotation marks omitted)). As the Supreme Court explained in Tyson Foods, Inc. v. Bouaphakeo, “[w]hen . . . ‘the concern about the proposed class is not that it exhibits some fatal dissimilarity but, rather, a fatal similarity—[an alleged] failure of proof as to an element of the plaintiffs’ cause of action—courts should engage that question as a matter of summary judgment, not class certification.’” 577 U.S. at 457 (quoting Richard A. Nagareda, Class Certification in the Age of Aggregate Proof, 84 N.Y.U. L. Rev. 97, 107 (2009)); see also Amgen, 568 U.S. at 468 (“A failure of proof on the common question of materiality ends the litigation and thus will never cause individual questions of reliance or anything else to overwhelm questions common to the class.”); Alcantar, 800 F.3d at 1053 (“A common contention need not be one that ‘will be answered, on the merits, in favor of the class.’” (quoting Amgen, 568 U.S. at 459)). “To hold otherwise would turn class certification into a mini-trial, when the purpose of class certification is merely to select the method best suited to adjudication of the controversy fairly and efficiently.” Alcantar, 800 F.3d at 1053 (cleaned up). The theoretical possibility that Dr. Dubé’s model, when executed, will reveal no damages thus does not undermine predominance, because that result would nonetheless be 20 LYTLE V. NUTRAMAX LABORATORIES, INC.

common to the class. 4 Nor does the possibility Dr. Dubé’s analysis might reveal damages with respect to some, but not all, of the challenged statements undermine predominance, because the very structure of the conjoint survey allows for an overcharge to be associated with each individual statement and label, allowing the amount each class member is entitled to recover to be easily assessed based solely on the product the class member purchased. And the possibility that an ascertainable portion of the class may be unable to recover—those not exposed to a statement with any attributable overcharge—does not in itself demonstrate class certification was improper. See Olean, 31 F.4th at 669, 680– 81 (holding that the possibility some class members suffered no injury does not, by itself, defeat class certification); Just Film, 847 F.3d at 1120 (“To gain class certification, Plaintiffs need to be able to allege that their damages arise from a course of conduct that impacted the class. But they need not show that each members’ damages from that conduct are identical.”). Nutramax’s argument in favor of its proposed categorical rule—requiring that a damages model always be executed prior to class certification—rests almost entirely upon its misinterpretation of the Supreme Court’s decision

4 Nutramax argues that this is not so because “Plaintiffs’ complaint asserted claims based not only on the labels, but also representations on the Cosequin® website and in other media; and Plaintiffs pled alternative theories of damages beyond the price-premium theory.” But if that were sufficient to defeat predominance, the rule announced in Tyson Foods would be rendered meaningless. In almost every case it will be possible to point to some individual proof that could substitute for the (purportedly) deficient aggregate proof. Here, however, Plaintiffs have narrowed the case to their challenge of these specific label claims using aggregate proof, and that is what the district court properly focused on. LYTLE V. NUTRAMAX LABORATORIES, INC. 21

in Comcast Corp. v. Behrend, 569 U.S. 27 (2013), and our recent en banc decision in Olean. We agree that these cases demand some assessment of the adequacy of Plaintiffs’ common proof at class certification, but nothing in these cases requires that Plaintiffs actually prove damages at the class certification stage or prohibits Plaintiffs from relying on an unexecuted but reliable damages model. In Comcast, plaintiffs put forward four theories of liability, but the district court certified a class as to only one of them. 569 U.S. at 31. However, the plaintiffs’ damages expert—whose model provided the only means of proving classwide damages—conducted an indivisible analysis which “did not isolate damages resulting from any one [of the four] theor[ies] of [liability].” Id. at 32. On appeal, the defendants argued class certification was improper because plaintiffs had “failed to attribute damages” to the sole remaining theory in the case, but the Third Circuit affirmed the district court’s grant of class certification over this objection, reasoning that it was an improper “attac[k] on the merits of the methodology [that had] no place in the class certification inquiry.” Id. The Supreme Court reversed. The Court explained that plaintiffs’ “model failed to measure damages resulting from the particular antitrust injury on which petitioners’ liability in this action is premised.” Id. at 36. This disconnect was fatal, because it meant plaintiffs could not “establish[] that damages are capable of measurement on a classwide basis.” Id. at 34. In other words, where an expert’s damages model is untethered from plaintiff’s theory of liability such that it has no possibility of demonstrating the amount of damages in a particular case, Comcast holds that a plaintiff may not rely upon it to show that damages are “capable of measurement on a classwide basis.” Id. Notably, the 22 LYTLE V. NUTRAMAX LABORATORIES, INC.

plaintiffs in Comcast “never challenged” the “need to prove damages on a classwide basis” in order to demonstrate predominance. Id. at 42 (Ginsburg, J., dissenting). Nutramax contends that, as in Comcast, permitting Plaintiffs here to rely on an unexecuted damages model would “reduce Rule 23(b)(3)’s predominance requirement to a nullity,” and would “not establish that the requirements of Rule 23 are satisfied ‘in fact.’” This is an overreading of Comcast. That decision has generally been construed to stand for the unremarkable proposition that “plaintiffs must be able to show that their damages stemmed from the defendant’s actions that created the legal liability.” Leyva v. Medline Indus. Inc., 716 F.3d 510, 514 (9th Cir. 2013). Indeed, even after Comcast, we have repeatedly reaffirmed that class treatment may be appropriate even where damages must be assessed on an individualized basis. See Vaquero v. Ashley Furniture Indus., Inc., 824 F.3d 1150, 1155 (9th Cir. 2016); Pulaski & Middleman, LLC v. Google, Inc., 802 F.3d 979, 987 (9th Cir. 2015). Contrary to Nutramax’s contention, the possibility that Dr. Dubé’s analysis might reveal damages with respect to some, but not all, of the challenged statements does not create “the problem that manifested in Comcast,” because, as explained above, the structure of the conjoint analysis allows a damages figure to be associated with each challenged statement. The same was not true in Comcast, where plaintiffs’ damages model, again, did not isolate damages between any of plaintiffs’ theories of liability. See Comcast, 569 U.S. at 32. Nutramax’s reliance on Olean is similarly unavailing. In Olean, plaintiffs brought an antitrust class action against the major U.S. packaged tuna suppliers, alleging that they engaged in unlawful price fixing. 31 F.4th at 661–62. The appeal centered around whether the district court properly LYTLE V. NUTRAMAX LABORATORIES, INC. 23

granted class certification based upon plaintiffs’ expert evidence showing antitrust impact. Id. We analyzed each of the plaintiffs’ damages models and the defendants’ objections to those models, and ultimately concluded that the district court did not abuse its discretion in certifying the class. See id. at 670–85. Nothing in Olean requires that an expert actually execute a damages model before it can be relied on. In describing the legal standard, Olean made clear that the focus is on “‘the method or methods by which plaintiffs propose to use the [class-wide] evidence to prove’ the common question in one stroke,” and “a district court is limited to resolving whether the evidence establishes that a common question is capable of class-wide resolution, not whether the evidence in fact establishes that plaintiffs would win at trial.” Olean, 31 F.4th at 666–67 (quoting In re Hydrogen Peroxide Antitrust Litig., 552 F.3d at 313). While Olean suggested that a district court must “[w]eigh[] conflicting expert testimony” and “consider[] factors that may undercut the model’s reliability,” id. at 666, 683, this does not categorically require an expert to execute the expert’s model before it may be relied upon. Rather, as explained in more detail below, the fact that an expert’s model has not yet been executed is simply one factor that must be considered. Nutramax cites snippets of Olean to argue that “[e]vidence is ‘capable of resolving a common issue’ where each class member could rely on it at trial and the evidence ‘could reasonably sustain a jury verdict in favor of the plaintiffs, even though a jury could still decide that the evidence was not persuasive.’” But this suggests that Olean required the same showing necessary to avoid summary judgment, which it plainly did not. Indeed, Nutramax’s reading of the “could reasonably sustain a jury verdict” 24 LYTLE V. NUTRAMAX LABORATORIES, INC.

language in Olean would mean that we took as mandatory the Supreme Court’s suggestion that class certification might sometimes overlap with the merits inquiry and would make a determination of the merits required, which would be contrary to settled law. It is settled that the question at class certification is not whether plaintiffs have put forward evidence capable of sustaining a jury verdict, but rather whether plaintiffs have shown enough to satisfy FRCP 23. To require an actual weighing, at class certification, of whether plaintiffs’ evidence could sustain a jury verdict would collapse the class certification and summary judgment inquiries in precisely the manner Tyson Foods warns against. 577 U.S. at 457 (noting that failures of proof that are common to the class should be engaged “as a matter of summary judgment, not class certification”); see also Miles v. Kirkland’s Stores Inc., 89 F.4th 1217, 1224 n.2 (9th Cir. 2024) (noting a “merits question that should be left for summary judgment or trial, not [resolved] at class certification”). Finally, Nutramax contends that where it is uncertain that a damages model will show any injury at all, permitting a class to be certified and sending notice to class members would be an inefficient use of resources. Plaintiffs respond that executing a conjoint analysis is extremely expensive and time consuming, so that requiring Plaintiffs to do so before the precise contours of the class have been established would risk wasting resources, not save them. We regard these competing speculations as largely irrelevant to determining what Rule 23 does or does not require. And since the determination of class certification is largely within the discretion of the district court, it is worth noting that the vast majority of district courts in our circuit to consider the question have found that a damages expert need not fully LYTLE V. NUTRAMAX LABORATORIES, INC. 25

execute his or her proposed conjoint analysis before it can be relied upon at class certification. See, e.g., Gunaratna v. Dennis Gross Cosmetology LLC, 2023 WL 5505052, at *19 (C.D. Cal. Apr. 4, 2023); Bailey v. Rite Aid Corp., 338 F.R.D. 390, 408 n.14 (N.D. Cal. 2021); Testone v. Barlean’s Organic Oils, LLC, 2021 WL 4438391, at *17 (S.D. Cal. Sept. 28, 2021); Guido v. L’Oreal, USA, Inc., 2014 WL 6603730, at *13–14 (C.D. Cal. July 24, 2014). 5 This common practice weighs against imposing the categorical rule Nutramax requests.

5 While there are district court cases that have found a proposed conjoint analysis to be insufficiently detailed or thorough to support a finding of predominance, they also do not support the categorical rule Nutramax proposes. Rather, they reflect the unremarkable proposition that an underdeveloped expert model is far less likely to be able to establish that a particular element is susceptible to common proof. See In re ConAgra Foods, Inc., 302 F.R.D. 537, 552 (C.D. Cal. 2014) (rejecting proposed conjoint survey where expert simply “offer[ed] a basic description of the manner in which hedonic regression and conjoint analysis operate, and assert[ed] that the exact specifications [the expert’s analysis would] use will be solidified as discovery progresses”); Miller v. Fuhu Inc., 2015 WL 7776794, at *21-22 (C.D. Cal. Dec. 1, 2015) (concluding that expert’s testimony that “it is possible and practical to design and conduct” a conjoint analysis was insufficient to establish predominance where expert had not designed such a survey). Indeed, in ConAgra, the district court subsequently granted class certification based upon a “proposed conjoint analysis” once the expert had provided a more fully developed methodology. See In re ConAgra Foods, Inc., 90 F. Supp. 3d 919, 954 (C.D. Cal. 2015); see also Vizcarra v. Unilever United States, Inc., 2023 WL 2364736, at *16–18 (N.D. Cal. Feb. 24, 2023) (noting the court had previously rejected un-executed conjoint analysis based on methodological flaws, but granting class certification based on revised analysis that corrected those flaws, even where the survey had not been fully executed). 26 LYTLE V. NUTRAMAX LABORATORIES, INC.

B. Having concluded that there is no categorical prohibition on a district court relying on an unexecuted damages model to certify a class, we must nevertheless determine whether, on the facts of this case, it was error for the district court to grant certification. In particular, Nutramax argues that the district court erred in concluding that Dr. Dubé’s opinions were sufficiently reliable to satisfy FRE 702 or FRCP 23. Nutramax argues that because Dr. Dubé had not yet obtained all of the data necessary to fully execute his model, the district court could not have found that his opinion satisfied the requirements of FRE 702 that his testimony be “based on sufficient facts or data.” Fed. R. Evid. 702(b). But here again, we think Nutramax confuses a class certification proceeding with a summary judgment motion. As applied to class certification where the issues are commonality and predominance, the Rule 702(b) question concerns whether the data suffices to show that a common question predominates over individual issues, not whether the subsequently executed model applied to a more complete dataset would then meet the requirements of 702(b) as applied to a summary judgment motion. By the same token, we do not agree with Nutramax that the “rigorous analysis” required at the class certification stage means that every expert opinion offered at that stage must be subjected to a full evidentiary hearing to see if each such opinion meets the requirements of Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). 1. The manner and extent to which the Daubert framework applies at the class certification stage is an unsettled question. A leading treatise has suggested that there is at LYTLE V. NUTRAMAX LABORATORIES, INC. 27

least some divergence among the Circuits on this question, with some employing a “full” Daubert inquiry and others employing a more limited one. See 3 Newberg and Rubenstein on Class Actions § 7:24 (6th ed. 2022). 6 Our own precedent has somewhat oscillated between these two approaches. In Ellis v. Costco Wholesale Corp., we affirmed a district court’s application of the Daubert standard at the class certification stage. 657 F.3d 970, 982– 83 (9th Cir. 2011); see also 3 Newberg and Rubenstein on Class Actions § 7:24 (6th ed. 2022) (characterizing Ellis as adopting a full Daubert test). Similarly, in Olean, we stated in dicta that “[i]n a class proceeding, defendants may challenge the reliability of an expert’s evidence under Daubert,” although the defendants in Olean did not actually raise a Daubert challenge. 31 F.4th at 665 n.7. In Sali v. Corona Regional Medical Center, however, we noted that while “a district court should evaluate admissibility under the standard set forth in Daubert,” whether testimony is admissible under that standard is “not . . . dispositive,” but instead “should go to the weight that evidence is given at the class certification stage.” 909 F.3d at 1006. We cited with approval the Eighth Circuit’s decision in In re Zurn Pex Plumbing Products Liability Litigation, which is the leading decision endorsing a more limited Daubert inquiry. Id. at 1004; see Cox v. Zurn Pex, Inc. (In re Zurn Pex Plumbing Prods. Liab Litig.), 644 F.3d 604, 613 (8th Cir. 2011) (“The main purpose of Daubert exclusion is to protect juries from being swayed by dubious scientific testimony. That interest

6 The Supreme Court originally granted certiorari in Comcast to resolve this question, but ultimately resolved the case on other grounds once it became apparent the question was not properly presented. Comcast, 569 U.S. at 39–40 (Ginsburg, J., dissenting). 28 LYTLE V. NUTRAMAX LABORATORIES, INC.

is not implicated at the class certification stage where the judge is the decision maker.”). We think that, at least for purposes of this case, the distinction between a “full” and “limited” Daubert inquiry is a function of what aspect of FRCP 23 is being addressed. Here, the question under FRE 702 is whether the model that the plaintiffs’ expert is offering on the issues of commonality and predominance is reliable for FRCP 23 purposes. Accordingly, such Daubert factors as peer review of the proffered model may be highly relevant, while others, such as known error rate, may be more applicable to the later- executed results of the test. Daubert itself stressed that its suggested factors were simply illustrative and needed to be applied flexibly, and this surely means applying them only to the extent helpful to the issue at hand. See Daubert, 509 U.S. at 593–94 (noting that “[t]he inquiry envisioned by Rule 702 is . . . a flexible one” and disclaiming any intent “to set out a definitive checklist or test”). Thus, for example, whether a “full” or “limited” Daubert analysis should be applied may depend on the timing of the class certification decision. If discovery has closed and an expert’s analysis is complete and her tests fully executed, there may be no reason for a district court to delay its assessment of ultimate admissibility at trial. By contrast, where an expert’s model has yet to be fully developed, a district court is limited at class certification to making a predictive judgment about how likely it is the expert’s analysis will eventually bear fruit. This still requires determining whether the expert’s methodology is reliable, so that a limited Daubert analysis may be necessary, but the more full-blown Daubert assessment of the results of the application of the model would be premature. LYTLE V. NUTRAMAX LABORATORIES, INC. 29

Here, we are satisfied that the district court’s limited Daubert analysis was sufficient for the immediate purposes. As the district court expressly recognized, “the court considers only if expert evidence is useful in evaluating whether class certification requirements have been met,” and for that purpose a more limited Daubert inquiry may be sufficient. This is consistent with the Supreme Court’s general rule that “[m]erits 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, 568 U.S. at 466. In finding that the prerequisites of FRE 702 were met for FRCP 23 purposes, the district court relied on Dr. Dubé’s unchallenged credentials, Dr. Dubé’s review of documentary evidence and marketing data, the fact that “[c]onjoint surveys, like the one proposed by [Dr. Dubé], are a well-established method for measuring class-wide damages in CLRA mislabeling cases,” and the fact that Dr. Dubé had successfully performed conjoint analyses in prior cases similar to this. It is true Dr. Dubé has not collected all of the necessary data to perform his calculations in the instant case, but implicit in Dr. Dubé’s opinion—which the district court credited—is the conclusion that he would be able to obtain such information, and Nutramax offers no reason to think he would be unable to do so. Nor, as explained below, has Nutramax shown either that Dr. Dubé’s methodology is flawed or that there is a likelihood that he will improperly apply that method to the facts. In light of the foregoing, we cannot say that the district court abused its discretion in rejecting Nutramax’s Daubert challenge to Dr. Dubé’s opinion. 30 LYTLE V. NUTRAMAX LABORATORIES, INC.

2. Nutramax also argues that, even if an unexecuted damages model may in some circumstances support class certification, on the facts of this case Dr. Dubé’s model is too underdeveloped to satisfy the “rigorous analysis” required under FRCP 23. Nutramax contends that Dr. Dubé has not designed the survey questionnaire, has not determined the precise demographic makeup of the individuals to be surveyed, has not selected all of the parameters for his model, and lacks certain data needed to finalize his calculations. Plaintiffs respond that Dr. Dubé has in fact fully designed the conjoint analysis and the methodology behind it, including by identifying the target population, analyzing economic data to determine the structure of the market, and specifying the mathematical analysis he will perform on the survey results. While Plaintiffs acknowledge Dr. Dubé has not yet programmed the survey (i.e., written the questions), they cite to Dr. Dubé’s testimony describing this as merely “an implementation detail,” and argue that it makes sense not to finalize the survey questions until the exact scope of the class is known. In short, according to Plaintiffs, the survey is fully designed and all that remains is for it to be executed. As already discussed above, Plaintiffs may rely on an unexecuted damages model to demonstrate that damages are susceptible to common proof. To be sure, the fact the model has not been executed remains relevant. Olean makes clear that “[t]he determination whether expert evidence is capable of resolving a class-wide question in one stroke may include ‘[w]eighing conflicting expert testimony’ and ‘[r]esolving expert disputes’ where necessary to ensure that Rule 23(b)(3)’s requirements are met.” 31 F.4th at 666 (internal citation omitted) (quoting In re Hydrogen Peroxide, 552 LYTLE V. NUTRAMAX LABORATORIES, INC. 31

F.3d at 323–24). This assessment of a “model’s reliability” required by Olean goes beyond the Daubert analysis, and the fact that an expert’s model is sufficiently reliable to meet the standard of FRE 702 as applied to a FRCP 23 determination may not be sufficient to satisfy the standard. See Ellis v. Costco Wholesale Corp., 657 F.3d 970, 982 (9th Cir. 2011) (suggesting FRCP 23 requires a distinct analysis beyond an assessment of admissibility under Daubert); Olean, 31 F.4th at 665–66 nn. 7, 9. Rather, the district court must also probe the likelihood that the model will be capable of generating common answers. A district court may not, however, “decline certification merely because it considers plaintiffs’ evidence relating to the common question to be unpersuasive and unlikely to succeed.” Id. at 667. 7 In applying this test to an unexecuted damages model, the question a district court must ask is whether the model will likely be able to generate common answers at trial. The fact that a model is underdeveloped may weigh against a finding that it will provide a reliable form of proof. Merely gesturing at a model or describing a general method will not suffice to meet this standard. Rather, plaintiffs—or their expert—must chart out a path to obtain all necessary data

7 Olean gave the following examples of expert evidence that, while otherwise admissible under Daubert, might be unable to generate common answers: where “the expert evidence was inadequate to prove an element of the claim for the entire class [i.e., is not common to all]; where the damages evidence was not consistent with the plaintiffs’ theory of liability; where the evidence contained unsupported assumptions; or where the evidence demonstrated nonsensical results such as false positives, i.e., injury to class members who could not logically have been injured by a defendant’s conduct.” Id. at 666 n.9 (internal citations omitted). 32 LYTLE V. NUTRAMAX LABORATORIES, INC.

and demonstrate that the proposed method will be viable as applied to the facts of a given case. 8 Here, the district court recognized that Plaintiffs were required to “show that damages are capable of measurement on a class-wide basis,” while acknowledging they may do so without executing the model. On appeal, Nutramax raises a flurry of attacks on the reliability of Dr. Dubé’s model that were never presented to the district court. While Nutramax referenced the underdeveloped nature of Dr. Dubé’s model throughout its briefing below, the only such argument it developed with any thoroughness was the contention that Dr. Dubé lacked critical data needed to complete his analysis. Since we cannot say that the district court abused its discretion in failing to consider arguments with which it was not presented, see Van v. LLR, Inc., 61 F.4th 1053, 1066 n.9 (9th Cir. 2023) (“An issue cannot form part of the district court’s class certification decision if it was never raised at the class certification stage.” (cleaned up)), we focus our analysis on those matters considered by the district court. 9

8 Plaintiffs’ briefing on appeal seems to advance a rule that merely putting forward a viable method is sufficient. To the extent that is Plaintiffs’ position, we disagree. Even where a method is otherwise valid and reliable under FRE 702, it may nonetheless fail to produce common answers for any number of reasons, such as when the model does not apply in a manner common to the class. Hence, we underscore that the ultimate inquiry is whether a proposed model is likely to provide common answers at trial. 9 In particular, on appeal Nutramax relies heavily upon the rebuttal report of their expert, Dr. Toubia, to challenge the reliability of Dr. Dubé’s methodology. Had Dr. Toubia’s report been fairly presented to the district court as a basis for denying class certification, it might have been error for the district court to not address it. See Olean, 31 F.4th at 666 LYTLE V. NUTRAMAX LABORATORIES, INC. 33

With respect to the data Dr. Dubé had not yet collected, the district court acknowledged defendant’s argument but credited Dr. Dubé’s implicit conclusion that he would be able to obtain such data prior to trial. Nutramax has not convincingly demonstrated that the district court erred in reaching this conclusion. Nutramax’s other attacks on Dr. Dubé’s methodology, to the extent they were presented to the district court, fare no better. As the district court observed, conjoint analysis is a well-accepted technique that is frequently used to establish damages in CLRA actions. See, e.g., Briseno v. ConAgra Foods, Inc., 674 F. App’x 654, 657 (9th Cir. 2017) (describing conjoint analysis as a “well-established damages model”); Hadley v. Kellogg Sales Co., 324 F. Supp. 3d 1084, 1107 (N.D. Cal. 2018) (“[C]onjoint analysis is a well- accepted economic methodology.” (quoting In re Dial Complete Mktg. & Sales Pracs. Litig., 320 F.R.D. 326, 331 (D.N.H. 2017)). Where an expert’s proposed method is novel or untested, it makes sense to demand a greater degree of specificity and completeness before it is relied upon to certify a class. See In re New Motor Vehicles Canadian Exp. Antitrust Litig., 522 F.3d 6, 26–30 (1st Cir. 2008) (finding that “novelty and complexity of the theories advanced” by plaintiffs and their expert made certification of a class based upon incomplete model less appropriate). By contrast, here there is no dispute that a conjoint analysis is capable of measuring classwide damages, at least in the abstract, and

(suggesting a district court must “[r]esolv[e] expert disputes” at class certification). However, our review of the record reveals that Nutramax never attempted to use Dr. Toubia’s rebuttal report to attack Dr. Dubé’s model, and instead cited Dr. Toubia’s rebuttal report only a single time, and for an entirely unrelated proposition. 34 LYTLE V. NUTRAMAX LABORATORIES, INC.

the only real question at this stage is whether Dr. Dubé will properly apply the method to the facts of the case. Nutramax cites a variety of potential errors Dr. Dubé might commit in executing his damages model. For example, Nutramax argues “the precise wording of a questionnaire is critical” and could “bias[] the results,” and that “assumptions underlying [his] economic model” may not account for real- world factors. While unanswered questions such as these, and the attendant possibility of errors, are certainly relevant, Nutramax offers no reason to think that Dr. Dubé will commit any of these errors. Dr. Dubé’s qualifications are undisputed, he has successfully conducted conjoint analyses in the past, and Dr. Dubé testified he did not “envision anything particularly unique about this survey.” The speculative possibility that Dr. Dubé might slip up in executing his model, standing alone, is insufficient to defeat class certification. See Sali, 909 F.3d at 1004–05 (“Neither 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 which apparently satisfies [Rule 23].” (quoting Blackie v. Barrack, 524 F.2d 891, 901 (9th Cir. 1975)). 10 Accordingly, the record was sufficient to support the district court’s conclusion that Dr. Dubé’s model is capable of showing damages on a class wide basis.

10 This is especially true here given that “[c]lass wide damages calculations under the . . . CLRA are particularly forgiving” and “require[] only that some reasonable basis of computation of damages be used.” Nguyen v. Nissan N. Am., Inc., 932 F.3d 811, 818 (9th Cir. 2019) (quoting Lambert, 870 F.3d at 1183). LYTLE V. NUTRAMAX LABORATORIES, INC. 35

C. As the above discussion makes clear, it is not required that a plaintiff’s expert must execute their damages model prior to class certification provided it is shown that the model provides a reliable and adequate method for calculating damages. We do, however, think it important to make clear that a plaintiff may not avoid ultimate scrutiny of the admissibility of their experts’ final opinions simply by declining to develop those opinions in advance of class certification. Accordingly, on remand, Nutramax must be given the opportunity in advance of trial to test the sufficiency and reliability of Dr. Dubé’s model once it has been fully executed, including through a motion for summary judgment and/or a renewed Daubert motion. IV. The second broad issue raised by Nutramax on appeal is whether the district court erred in concluding that common questions predominated with respect to the element of reliance. California’s CLRA prohibits “unfair methods of competition and unfair or deceptive acts or practices.” Cal. Civ. Code § 1770(a). To bring a CLRA claim, a plaintiff must show (1) the defendant engaged in deceptive conduct and (2) the deception caused plaintiff harm. Stearns, 655 F.3d at 1022. However, under the CLRA, “[c]ausation, on a classwide basis, may be established by materiality. If the trial court finds that material misrepresentations have been made to the entire class, an inference of reliance arises as to the class.” Id. (quoting In re Vioxx Class Cases, 180 Cal. App. 4th 116, 129 (2009)). A misrepresentation is material “if a reasonable [person] would attach importance to its existence or nonexistence in determining his choice of action in the transaction in question[.]” Id. 36 LYTLE V. NUTRAMAX LABORATORIES, INC.

Because materiality (and, hence, in this case reliance) may be proved by reference to an objective, reasonable consumer standard, reliance under the CLRA is generally susceptible to common proof. See Noel v. Thrifty Payless, Inc., 17 Cal. App. 5th 1315, 1334 (2017) (“When the consumer shows the complained-of misrepresentation would have been material to any reasonable person, he or she has carried the burden of showing actual reliance and causation of injury for each member of the class.”), rev’d on other grounds, 7 Cal. 5th 955 (2019); Stearns, 655 F.3d at 1022; Bradach v. Pharmavite, LLC, 735 F. App’x 251, 254 (9th Cir. 2018) (“Under California law, class members in CLRA . . . actions are not required to prove their individual reliance on the allegedly misleading statements. Instead, the standard . . . is whether members of the public are likely to be deceived. For this reason, . . . [CLRA] claims are ideal for class certification because they will not require the court to investigate class members’ individual interaction with the product.” (internal quotation marks and citations omitted)). However, while materiality can support an inference of reliance, that does not necessarily mean that the inference will hold as to the entire class, such that common questions predominate. “If the misrepresentation or omission is not material as to all class members, the issue of reliance ‘would vary from consumer to consumer’ and the class should not be certified.” Stearns, 665 F.3d at 1022–23 (quoting Vioxx, 180 Cal. App. 4th at 129). The district court cited the correct legal standard, noting that materiality can be used to establish reliance under the CLRA, while also acknowledging that the presumption is rebuttable insofar as “reliance would vary from consumer to consumer.” Nutramax nonetheless contends the district court committed legal error by finding “the ‘objective test’ used LYTLE V. NUTRAMAX LABORATORIES, INC. 37

for materiality rendered materiality and thus causation inherently a common issue.” Read in context, we do not think the district court applied an incorrect, irrebuttable presumption as Nutramax suggests. Rather, the observation that materiality and causation are “inherently common issues” was made only after the district court considered the relevant evidence of materiality and concluded it supported an application of the presumption. We therefore review the district court’s findings on this issue for abuse of discretion. See Sali, 909 F.3d at 1002. 11 Plaintiffs presented ample evidence to show that the challenged statements would be materially misleading to a reasonable consumer. To demonstrate that a reasonable consumer would have understood the challenged statements to have promised that the product would improve a dog’s joint health, Plaintiffs cited the testimony of the named plaintiffs, the testimony of their advertising expert, Silverman, and the survey results of Nutramax’s own expert, Dr. Scott, which indicated that the near-universal understanding of Cosequin’s purpose was that it would “improve/help/maintain mobility, flexibility, joint health/support.” This understanding, caused by Nutramax’s packaging statements, was, according to Plaintiffs’ experts, false and misleading because there is no evidence that Cosequin improves a dog’s joint health. The district court credited all of this evidence and expressly found that it “outweighed” the contrary evidence presented by defendants’ experts, including Dr. Scott. Nutramax has not

11 We also do not think the district court committed legal error when it cited the rule that “a plaintiff is not required to show that the challenged statement is the ‘sole or even the decisive cause’ influencing the class members’ decisions to buy the challenged products.” As explained below, this statement is consistent with California law. 38 LYTLE V. NUTRAMAX LABORATORIES, INC.

shown that this conclusion was an abuse of discretion. The district court thus correctly found that Plaintiffs had demonstrated that the presumption of reliance applied. 12 Nutramax nonetheless argues that it put forward sufficient evidence to rebut the presumption, and the district court erred in concluding otherwise. Nutramax points to what it describes as “overwhelming evidence that veterinarians frequently recommend joint supplements, including Cosequin.” Nutramax similarly cites the results of Dr. Scott’s survey, which suggests approximately half of survey respondents decided to purchase Cosequin before going to a physical store or website. According to Nutramax, all of this demonstrates that individualized assessments will be needed to determine whether any given class member actually relied on the label. The fact some class members considered sources of information other than the packaging in making their purchasing decisions does not necessarily undermine reliance. To establish reliance under the CLRA, a misrepresentation need not be “the sole or even the decisive cause of the injury-producing conduct.” Moore v. Mars Petcare US, Inc., 966 F.3d 1007, 1020 (9th Cir. 2020) (quoting Kwikset Corp. v. Superior Ct., 246 P.3d 877, 888 (Cal. 2011)). For example, in Moore, we found that “[t]he fact that vets had prescribed each Plaintiff the pet food— rather than each discovering the pet food on their own—

12 Nutramax argues that Plaintiffs’ evidence of falsity showed only that Cosequin was unable to treat dogs with arthritis, not that it was totally ineffective at promoting joint health generally, but Nutramax made this argument in the court below and it was expressly rejected by the district court, which credited Plaintiffs’ evidence that “Cosequin has no effect on joint health.” LYTLE V. NUTRAMAX LABORATORIES, INC. 39

[did] not negate the allegation of actual reliance because the prescription requirement and advertising need not be the sole or even the decisive cause of the purchase.” Id. at 1020–21. In other words, Moore rejected the very argument now advanced by Nutramax, that there can be no reliance where a veterinarian recommends a product. Nutramax attempts to distinguish Moore on the grounds that it was decided on a motion to dismiss. However, at least one California appellate court has applied the rule beyond the motion to dismiss context. See Veera v. Banana Republic, LLC, 6 Cal. App. 5th 907, 919 (2016) (applying rule to deny motion for summary judgment). And the model jury instructions for a CLRA claim promulgated by the Judicial Council of California similarly indicate that, “[t]o prove reliance, [name of plaintiff] need only prove that the representation was a substantial factor in [his/her/nonbinary pronoun] decision” and “does not need to prove that it was the primary factor or the only factor in the decision.” Judicial Council of California, Model Civil Jury Instruction No. 4700. 13

13 The full instruction reads: [[Name of plaintiff]’s harm resulted from [name of defendant]’s conduct if [name of plaintiff] relied on [name of defendant]’s representation. To prove reliance, [name of plaintiff] need only prove that the representation was a substantial factor in [his/her/nonbinary pronoun] decision. [He/She/Nonbinary pronoun] does not need to prove that it was the primary factor or the only factor in the decision. 40 LYTLE V. NUTRAMAX LABORATORIES, INC.

To be sure, the fact that some consumers relied on other sources of information is relevant to the assessment of reliance. In an appropriate case, such evidence might be sufficient to demonstrate that the misrepresentation was not “a substantial factor” in a large percentage of consumers’ purchasing decisions. But we do not think it was an abuse of discretion for the district court to find to the contrary here. If one imagines a counterfactual where the packaging contained no suggestion that Cosequin benefited a dog’s joint health, it is entirely plausible that no consumer would have chosen to purchase the product. See Alvarez v. NBTY, Inc., 331 F.R.D. 416, 423 (S.D. Cal. 2019) (“Without [the challenged] statement, no consumer would have a reason to purchase the Products and would otherwise be purchasing a random bottle of supplements without any knowledge of what benefit, if any, the supplements provided.”). Further, while one of Plaintiffs’ theories of causation is that they would not have purchased the products had they known the truth, another theory is that they paid more than they otherwise would have as a result of the misleading statements. See Hinojos v. Kohl’s Corp., 718 F.3d 1098, 1108 (9th Cir. 2013) (“[T]he CLRA’s ‘any damage’ requirement is a capacious one that includes any pecuniary damage as well as opportunity costs and transaction costs that result when a consumer is misled by deceptive marketing practices.”). If, as Plaintiffs contend, Cosequin provided no benefits to joint health, it is still more plausible

If [name of defendant]’s representation of fact was material, reliance may be inferred. A fact is material if a reasonable consumer would consider it important in deciding whether to buy or lease the [goods/services].] LYTLE V. NUTRAMAX LABORATORIES, INC. 41

that the consumers would have paid less had they known the truth. This construction of the reliance requirement comports with the objectives of the CLRA more generally. The California legislature has declared that the CLRA “shall be liberally construed and applied to promote its underlying purposes, which are to protect consumers against unfair and deceptive business practices and to provide efficient and economical procedures to secure such protection.” Cal. Civ. Code § 1760. Consistent with this remedial objective, we have suggested that CLRA claims “are ideal for class certification because they will not require the court to investigate class members’ individual interaction with the product.” Bradach, 735 F. App’x at 254 (quotation omitted). Nutramax’s cramped interpretation of the reliance requirement, permitting the presumption to be overcome based upon marginally different interactions with a product containing a label that is otherwise materially misleading, would undermine the presumption of reliance and the capacity to bring CLRA class actions. None of the cases that Nutramax cites, finding that the presumption of reliance was overcome, is to the contrary. For example, in Stearns v. Ticketmaster Corp., the plaintiffs sued Ticketmaster for creating a website that misled individuals who purchased tickets into inadvertently signing up for an unrelated monthly subscription. 655 F.3d at 1017. While the district court acknowledged that this practice likely violated the CLRA, it nevertheless found class certification inappropriate because the plaintiffs’ proposed class was too broad. Id. at 1024. We upheld this holding and explained that materiality was not uniform across the class because many class members would not have been deceived 42 LYTLE V. NUTRAMAX LABORATORIES, INC.

by Ticketmaster’s marketing practice and had intentionally signed up for the subscription. Id. Again, in In re Vioxx Class Cases, the “[p]laintiffs suggest[ed] that Merck hid an increased risk of death, associated with Vioxx,” which plaintiffs used to support an inference of reliance. 180 Cal. App. 4th 116, 133 (2009) (cleaned up). While a risk of death is material in the abstract, the trial court found class treatment inappropriate because defendants had introduced “overwhelming evidence” that materiality varied on an individual basis. Id. For example, the record contained undisputed evidence that the drug actually did not increase the risk of death for all of the class members and that some individuals continued taking the drug even after learning of the risks in light of its substantial countervailing benefits. Id. at 133–34. Finally, in Fairbanks v. Farmers New World Life Ins. Co., the plaintiffs alleged that defendant’s marketing practices surrounding “universal life insurance” were misleading because the marketing suggested the policies were “permanent” when in fact the policies were not permanent and were systematically underfunded. 197 Cal. App. 4th 544, 553 (2011). The court observed that plaintiffs’ class certification motion “assume[d] that anyone who purchases universal insurance does so because . . . a universal policy (if sufficiently funded) can be permanent,” when in fact the record showed there were many other reasons an individual might purchase a universal insurance policy unrelated to its supposed permanence. Id. at 565. Therefore, not all class members would find the misrepresentations about the policy being “permanent” material or misleading. LYTLE V. NUTRAMAX LABORATORIES, INC. 43

The common theme unifying each of these cases is that a sizable portion of the class either were not misled by the statements or would not have found the misrepresentations to be material had they known the truth. Here, by contrast, Dr. Scott’s own survey indicated that the near-universal reason class members purchased Cosequin was because it would “improve/help/maintain mobility, flexibility, joint health/support.” Indeed, it is difficult to see why else consumers would purchase this “joint health supplement” other than to improve their dog’s joint health. For purposes of class certification, Plaintiffs have adequately demonstrated that a reasonable consumer would have been misled into believing Cosequin would improve their dogs’ joint health, when, in fact, Cosequin provided no such benefits, and that this misrepresentation would have been material as to the entire class. The district court thus did not abuse its discretion in concluding reliance may be proven on a class wide basis. V. For the foregoing reasons, we affirm the district court’s grant of class certification. AFFIRMED.