Opinion
5:14-cv-05659-EJD
09-30-2024
PAUL ORSHAN, et al., Plaintiffs, v. APPLE INC., Defendant.
ORDER GRANTING IN PART AND DENYING IN PART MOTIONS TO EXCLUDE EXPERT TESTIMONY AND TO CERTIFY CLASS
RE: ECF NOS. 194, 213, 226
EDWARD J. DAVILA UNITED STATES DISTRICT JUDGE
Before the Court is Plaintiffs' second motion for class certification of claims that Defendant Apple Inc. misled consumers about the available storage capacity of certain smartphone and tablet devices running its iOS 8 operating system. The Court denied Plaintiffs' initial class certification motion on adequacy, typicality, and predominance grounds. Now, with new class representatives and newly developed evidence, Plaintiffs again move for class certification. Apple opposes class certification and, relatedly, moves to exclude parts of the report by Plaintiffs' expert, Dr. Andreas Groehn. After the parties fully briefed both motions, the Court held a hearing on these matters. Following careful consideration of the parties' submissions and their arguments at hearing, the Court GRANTS IN PART and DENIES IN PART Apple's motion to exclude and GRANTS IN PART and DENIES IN PART Plaintiffs' motion for class certification.
After the hearing, Plaintiffs moved for leave to file supplemental authority, which Apple opposed to the extent Plaintiffs made arguments about that authority. The Court GRANTS Plaintiff's motion for leave to file supplemental authority. The Court considers Plaintiffs' new authority but does not consider either side's arguments about that authority. Separately, both sides requested judicial notice of certain facts and publications. ECF Nos. 203, 208. Neither side opposes judicial notice, and the Court finds both sides' requests proper under Federal Rule of Evidence 201. Therefore, the Court GRANTS both requests for judicial notice.
I. BACKGROUND
Plaintiffs are Paul Orshan, Deanna Ness, Kye Weasner, and Margaret Hart, each of whom purchased Apple devices advertised as having 16 gigabytes of storage. Decl. of Deanna Ness (“Ness Decl.”) ¶ 2, ECF No. 194-29; Decl. of Kye Weasner (“Weasner Decl.”) ¶ 2, ECF No. 194-30; Decl. of Margaret Hart (“Hart Decl.”) ¶ 2, ECF No. 194-31; Decl. of Paul Orshan (“Orshan Decl.”) ¶ 2, ECF No. 194-37. Three of the Plaintiffs-Ms. Ness, Mr. Weasner, and Ms. Hart-purchased their 16 GB Apple devices with iOS 8 preinstalled. Ness Decl. ¶ 3; Weasner Decl. ¶ 3; Hart Decl. ¶ 3. Mr. Orshan's device did not initially come with iOS 8, but he later upgraded to iOS 8. Orshan Decl. ¶ 3. All four Plaintiffs testified to believing, based on Apple's representations, that they would receive 16 gigabytes of storage for personal use; Plaintiffs claim that they would not have purchased their devices had they known iOS 8 took up a significant portion of the advertised 16 gigabytes of storage. Ness Decl. ¶¶ 4-5; Weasner Decl. ¶¶ 4-5; Hart Decl. ¶¶ 4-5; Orshan Decl. ¶¶ 4-5. Because Plaintiffs believe that Apple misrepresented the amount of storage capacity available in their 16 GB devices, they sued Apple under three California consumer protection laws: the Unfair Competition Law (“UCL”), the False Advertising Law (“FAL”), and the Consumer Legal Remedies Act (“CLRA”).
Plaintiffs now seek to certify three classes. First, they propose a nationwide “Upgrade Subclass” consisting of those who purchased 16 GB Apple devices without iOS 8 between March 11, 2011 and September 16, 2014, and who subsequently upgraded those devices to iOS 8 between September 17, 2014 and September 30, 2016. Renewed Mot. for Class Cert. (“Class Mot.”) 11, ECF No. 194. Second, they propose a nationwide “Preinstall Subclass” consisting of those who purchased 16 GB Apple devices with iOS 8 preinstalled between September 17, 2014 and September 30, 2016. Id. Finally, as an alternative to the nationwide Preinstall Subclass, Plaintiffs propose a “California Preinstall Subclass” consisting of Preinstall Subclass members who purchased their 16 GB Apple devices in California. Id.
II. MOTION TO EXCLUDE
The Court begins with Apple's motion to exclude certain parts of Dr. Groehn's expert report. Mot. to Exclude Testimony of Dr. Andreas Groehn (“Expert Mot.”), ECF No. 213. Specifically, Apple takes issue with Dr. Groehn's opinions about classwide damages and seeks to exclude all three of Dr. Groehn's proposed damage models.
In Daubert v. Merrell Dow Pharmaceuticals, Inc., the Supreme Court held that Federal Rule of Evidence 702 requires courts to serve as gatekeepers of expert testimony to ensure that such testimony “rests on a reliable foundation and is relevant to the task at hand.” 509 U.S. 579, 597 (1993). At the class certification stage, this requirement is somewhat relaxed, and courts do not always need to conduct a full Daubert analysis. Lytle v. Nutramax Lab'ys, Inc., ___ F.4th ___, No. 22-55744, 2024 WL 3915361, at *11 (9th Cir. Aug. 23, 2024). The level of scrutiny that applies at class certification “is a function of what aspect of [Federal Rule of Civil Procedure] 23 is being addressed.” Id. Where an expert offers opinions to show that damages can be measured on a classwide basis but has “has not [yet] collected all of the necessary data to perform his calculations,” courts perform only a limited Daubert analysis focused on the expert's methodology, leaving a “full-blown Daubert assessment” of the expert's results for another time. Id.
That is the case here. In opining about how he could measure classwide damages, Dr. Groehn had access to a dataset of subsidized prices for relevant Apple devices at the time those devices initially entered the market, Am. Expert Rep. of Dr. Andreas Groehn (“Am. Groehn Rep.”) tbl.4, ECF No. 194-1, but he did not have access to datasets showing how prices changed as newer generation devices were released, or showing how many units of each device were sold. Id. ¶¶ 92, 106. Since those missing datasets would help Dr. Groehn further refine his damage models, id., it is too early to conduct a full Daubert analysis.
Under the limited Daubert rubric appropriate for class certification, the Court considers two factors: (1) relevance to class certification, and (2) reliability of the expert's proposed methodology “in light of the criteria for class certification and the current state of the evidence.” Tait v. BSH Home Appliances Corp., 289 F.R.D. 466, 495 (C.D. Cal. 2012) (citation omitted); see also Lytle, 2024 WL 3915361, at *11. These two factors track those required by Daubert and Rule 702. Daubert, 509 U.S. at 597 (requiring judges to “ensur[e] that an expert's testimony both rests on a reliable foundation and is relevant to the task at hand”) (emphasis added); Fed.R.Evid. 702. But unlike in a full Daubert analysis, the question of relevance is tailored to class certification, and the required showing of reliability less demanding. Applying this limited Daubert rubric, the Court agrees with Apple that Dr. Groehn's first damage model (the “Rule-of-Three” Model) does not pass muster. But the Court finds that Dr. Groehn's remaining two damage models (the “Incremental Price Regression Model” and the “Hedonic Regression Model”) are reliable enough to consider at the class certification stage.
A. Rule-of-Three Model
The Rule-of-Three Model relies on a simple calculation to determine the value of storage: divide the price difference between two device models (e.g., the price difference between a 16 GB iPad 2 and 32 GB iPad 2) by the storage capacity difference between those same two devices models. Am. Groehn Rep. ¶ 83. This is the exact model that the Court already rejected in denying Plaintiffs' previous class certification motion. Id. ¶ 78.
The Court previously rejected the Rule-of-Three model because the model had two obvious flaws. First, the model assumed that each additional gigabyte of storage was worth the same amount as the previous gigabyte of storage even though market data showed that assumption to be plainly wrong. Order Den. Mot. for Class Cert. (“Prior Order”) 4-5, ECF No. 184. For example, consumers paid $100 more for 16 additional gigabytes of storage when moving from the 16 GB version of the iPad 2 to the 32 GB version. Id. But consumers still only paid $100 more for 32 additional gigabytes when moving from the 32 GB version to the 64 GB version. Id. Second, the model produced unexplained discrepancies, such as estimating that each gigabyte of storage in the iPad mini 3 without cellular connection was worth $2.08 but estimating that each gigabyte in iPad mini 3 with cellular connection was worth three times as much at $6.25. Id. at 5.
In offering the Rule-of-Three Model a second time, Plaintiffs have addressed only one of those two flaws. Plaintiffs corrected the second flaw, explaining that the discrepancy in iPad mini 3 prices was the result of a data entry error that Dr. Groehn has since corrected. Class Mot. 31 n.13. The first flaw remains. Dr. Groehn still has not justified the assumption that each additional gigabyte of storage is worth the same amount as the previous gigabyte. In fact, Dr. Groehn admits that this assumption is wrong. Dr. Groehn concedes that less consumer demand for additional storage as storage capacity increases, coupled with the economic concept of diminishing marginal utility, causes each additional gigabyte of storage to be worth less than the previous gigabyte. Am. Groehn Rep. ¶ 86.
Dr. Groehn tries to rationalize his use of this faulty assumption because using the assumption supposedly underestimates true damages. Id. ¶¶ 88-89. Dr. Groehn explains that the Rule-of-Three Model uses later gigabytes of storage (i.e., the 16th through 32nd gigabytes) to estimate the value of earlier gigabytes of storage (i.e., the 14th through 16th gigabytes, which are approximately the gigabytes that consumers allegedly did not have access to). Id. ¶¶ 87-89. According to him, since the later gigabytes are worth less than the 14th through 16th gigabytes, the model's estimated value for the 14th through 16th gigabytes is necessarily lower than the actual value. Id. But Daubert asks whether expert opinions are reliable and relevant, not whether they are conservative. See Fed.R.Evid. 702 (no discussion of conservatism as a criterion for admissible expert opinions). If a damage model could survive Daubert by simply underestimating true damages, an expert could avoid having a court exclude her opinions by picking an arbitrary damage figure that is comfortably below any reasonable amount of true damages even though such an opinion would be plainly unreliable. Thus, Plaintiffs cannot sneak Dr. Groehn's Rule-of-Three Model past Daubert on the ground that the model consistently underestimates true damages.
The Rule-of-Three Model makes an assumption that even Dr. Groehn admits is wrong. Therefore, it does not reliably estimate damages, and the Court GRANTS Apple's motion to exclude as to that model.
B. Incremental Price Regression and Hedonic Regression Models
Dr. Groehn's remaining two damage models are both based on regressions. Because Apple's criticisms of each are similar, the Court addresses both models together. While Apple makes potentially compelling arguments about the persuasiveness of Dr. Groehn's two regression models, the Court's role under Daubert is that of a gatekeeper tasked with ensuring a threshold level of reliability and relevance, not that of a factfinder charged with deciding how convincing an expert's opinion is. Primiano v. Cook, 598 F.3d 558, 565 (9th Cir. 2010) (citation omitted). From that angle, Apple's arguments go to the weight of Dr. Groehn's opinions, not their admissibility, and the Court finds that Dr. Groehn's two regression models survive Daubert at this stage.
1. Reliability
Apple's main argument for excluding Dr. Groehn's regression models is that those models are unreliable. However, Apple's reliability arguments are based on challenges to Dr. Groehn's results rather than challenges to Dr. Groehn's methodology. The Court found above that conducting a full Daubert analysis of Dr. Groehn's results is not appropriate right now because there is additional data that Dr. Groehn needs to finalize his proposed damage models. Apple's reliability arguments therefore do not justify excluding Dr. Groehn's regression models at class certification. And even if the Court were to consider Apple's results-based critiques, those critiques still would not justify excluding Dr. Groehn's regression models.
a. Inconsistent Results
Apple's primary argument is that Dr. Groehn's regression models are unreliable because they produce inconsistent results both internally within the same model and externally when comparing between the two models. Expert Mot. 8.
Apple's first contention about internal inconsistency appears to draw on one of the Court's concerns in its Prior Order. There, the Court excluded Dr. Groehn's Rule-of-Three Model partly because that model produced internally inconsistent results. The Rule-of-Three Model produced vastly different damage estimates for Apple devices where storage ostensibly should have been valued similarly; namely, it estimated that storage for the iPad mini 3 with cellular connection was worth $6.25 per gigabyte while storage for the iPad mini 3 and without cellular connection was worth $2.08 per gigabyte. Prior Order 5. That is not the case for Dr. Groehn's two regression models. Both the Incremental Price Regression and the Hedonic Price Regression Models produce roughly the same damage estimates for iPad versions with and without cellular connections. Am. Groehn Rep. tbl.9. And while there are some differences in the damage estimates between different generations of iPhones and iPads, Apple has not argued or otherwise suggested that damages should be the same across different device generations.
Nonetheless, Apple claims that separate analysis from its own expert, Dr. Justin McCrary, shows that Dr. Groehn's regression models are internally inconsistent. Expert Mot. 8-9. At a high level, Dr. Groehn's regression models estimate damages by comparing 16 GB Apple devices to their counterparts with greater storage capacities (e.g., comparing the 16 GB iPhone 6 with 64 GB and 128 GB versions of the iPhone 6). Expert Rep. of Prof. Justin McCrary (“McCrary Rep.”) ¶ 24, ECF No. 201-18. As Dr. McCrary observes, the results from Dr. Groehn's regression models can vary depending on the comparator devices chosen. For example, when Dr. Groehn used prices from the 16 GB, 32 GB, 64 GB, and 128 GB versions of the iPad Air to run his Incremental Price Regression Model, he estimated that the 16th gigabyte of storage was worth $7.36. Am. Groehn Rep. fig.25. But the Incremental Price Regression Model produces different results when using different comparator devices. Excluding the 32 GB version of the iPad Air from the regression leads to an estimated price of $5.28 for the 16th gigabyte of storage. McCrary Rep., Ex. 1. And excluding both the 32 GB and 64 GB versions generates an estimated price of $2.68 for the 16th gigabyte. Id. Dr. McCrary demonstrated similar outcomes when using different comparators with the Hedonic Regression Model. Id.
Apple purports to cite Dr. Groehn's report for these inconsistent results, but a review of Dr. Groehn's report shows that many of the allegedly inconsistent numbers do not come from his report. Instead, those numbers come from Dr. McCrary's report.
These observations about the choice of comparators are not fatal to Dr. Groehn's regression models. When the Court rejected the Rule-of-Three Model for producing internally inconsistent results, it did so because the model produced notably different results for two iPad models that should have had similar results. By contrast, the inconsistencies in the two regression models here arise only when different data (i.e., different comparator devices) are inputted into those models. Apple has not offered any reason for the Court to expect different data to produce similar results, especially when, as illustrated by the iPad Air example above, there are so few datapoints to begin with.
That is not to say that one should never expect reliable models to produce similar results with different data. For example, because Dr. Groehn's Rule-of-Three Model assumed that each gigabyte of storage had the same value, that model should be expected to yield the same result regardless of the comparator device chosen. The fact that the Rule-of-Three Model did not do so rendered the model unreliable, not because of some general rule requiring different data to produce the same result, but because the different results falsified one of the model's key assumptions.
To the extent Apple's criticism raises any doubts about Dr. Groehn's regression models, those are doubts related to Dr. Groehn's choice of data. These data choices are easy to understand-whether to include a particular Apple device in the regression model or not-so a jury could readily evaluate whether those choices add to or detract from Dr. Groehn's persuasiveness. As such, Apple's arguments about comparator devices go to weight, not admissibility, and they are better directed to the jury as factfinder, not to the Court as gatekeeper. See Southland Sod Farms v. Stover Seed Co., 108 F.3d 1134, 1143 n.8 (9th Cir. 1997) (when a jury is able to assess purported deficiencies in an expert's methodology, it is appropriate for juries to do so rather than the Court).
Apple's arguments about the inconsistencies between the Incremental Price Regression and Hedonic Regression Models similarly do not justify excluding those two models. Apple contends that, because the Incremental Price Regression and Hedonic Regression Models are both supposed to measure the same quantity, the fact that the two models yield disparate results when run on the same data shows that neither model is reliable. Expert Mot. 9-10. Again, while Apple might be able to convince a jury with that argument, Apple has not shown that Dr. Groehn's models are so fatally flawed as to require exclusion.
Usually, when two models produce different results, the question of which model is more accurate falls to the factfinder as part of a classic battle of the experts. After hearing each expert's thoughts about why one model is better than the other, the factfinder then weighs the experts' testimony and decides which model she trusts more. Of course, there is no battle of the experts here. Instead of two competing experts offering competing models, a single expert, Dr. Groehn, proposes both models at issue. And instead of there being testimony about how one model is better than the other, Dr. Groehn has refused to pick favorites. Dep. of Andreas Groehn 252:1024, ECF No. 201-19. As Apple notes, this is somewhat unusual. But it is not grounds for exclusion. Like in a classic battle of the experts, it is the factfinder's job, not the Court's, to decide which of the two models to put more trust in. One model might be better than the other, but that does not mean either (or both) models are unreliable. Indeed, it is natural for there to be “some minor disagreement between [] models.” In re LIBOR-Based Fin. Instruments Antitrust Litig., 299 F.Supp.3d 430, 478 (S.D.N.Y. 2018).
Of course, some deviations between two different models can be so significant that they call an expert's opinion into doubt even at the Daubert stage. For example, courts might exclude an expert's opinion where the expert proposes two different models that “produce directionally inconsistent results a substantial portion of the time.” Id. at 477. Or, courts might find experts to be unreliable where they offer different models to measure the same quantity and those models produce results magnitudes apart from each other. Cf. Immunex Corp. v. Sanofi, No. 17-cv-02613, 2018 WL 6252460, at *10 (C.D. Cal. Aug. 24, 2018) (in the context of evaluating prior art for claim construction, “if two experimental designs were to produce such diametrically opposed results as 0% inhibition and 100% inhibition between the same antibody pair in the same direction, that would be a sure sign that the methodology in at least one of those assays was flawed”). The Incremental Price Regression and Hedonic Regression Models, however, do not produce results diverging to that extent. See Am. Groehn Rep. tbl.9.
Therefore, the supposed discrepancies in and between Dr. Groehn's regression models do not require exclusion.
b. Differences From Real-World Pricing
Apple's secondary argument is that Dr. Groehn's regression models are unreliable because they produce results that are unrealistic when compared to real-world pricing. Once again, that argument goes to weight, not admissibility.
First, Apple claims that Dr. Groehn's models estimate storage to represent a disproportionately high amount of the total value of Apple's devices-77% or higher for some devices. McCrary Rep. ¶ 44. But in calculating those numbers, Dr. McCrary used the subsidized prices of Apple devices as a proxy for total value. Id.; Am. Groehn Rep. ¶ 74 & tbl.4 (noting the use of subsidized prices). As the term suggests, subsidized prices account for compensation that Apple received beyond the price paid by the consumer, meaning that subsidized prices do not accurately represent total value. The unsubsidized prices are a better reflection of Apple devices' total value, and such prices can be significantly greater than subsidized ones. Am. Groehn Rep. ¶ 74 (a 16 GB iPhone 6 is priced at $199 when subsidized and $649 when unsubsidized). Thus, Apple's arguments about storage as a percentage of an Apple device's total value miss the mark.
Second, Apple asserts that it prices its devices in increments of $100 between different tiers of storage, and that all its prices end in the digit “9.” McCrary Rep. ¶¶ 47, 53-55. From Apple's perspective, any damage model that does not adhere to these historical practices is unreliable because it produces unrealistic results. For instance, the Incremental Price Regression Model estimates $6.82 in damages for the iPhone 6. Id. ¶ 47. The actual (subsidized) market price of a 16 GB iPhone 6 was $199, so this result implies that, if Apple had made disclosures about the amount of storage available to users, the but-for price of the iPhone 6 would have been $192.18. Id. (subtracting $6.82 from $199). Since, according to Apple, it would never actually set such a price, Dr. Groehn's model must be unreliable. Expert Mot. 14.
Apple misunderstands the relevant measure of damages. Damages do not depend on the but-for price that Apple would be willing to set in a hypothetical world with accurate disclosures. Instead, damages are based on “the price a consumer would have been willing to pay for the product had it been labeled accurately.” Vizcarra v. Unilever U.S., Inc., 339 F.R.D. 530, 553 (N.D. Cal. 2021) (emphasis added) (citing Pulaski & Middleman, LLC v. Google, Inc., 802 F.3d 979, 988-89 (9th Cir. 2015)). Specifically, damages are “the difference between what was paid and what a reasonable consumer would have paid at the time of purchase without the fraudulent or omitted information.” Pulaski, 802 F.3d at 989 (emphasis added) (citing Kwikset Corp. v. Superior Ct., 51 Cal.4th 310, 329 (2011)).
Applying this “reasonable consumer” measure of damages instead of Apple's proposed measure of damages also makes good sense. For one, the more market power a defendant company has, the more easily it can argue that it would not have changed its prices even with additional disclosures correcting the alleged misrepresentations. In that case, Apple's proposed measure of damages-based on the price that a company would set rather than the price a reasonable consumer would pay-would effectively shield companies with greater market power from damages. In fact, Apple's argument would probably shield a majority of companies from damages, even those without much market power. Most companies choose prices ending in “9,” McCrary Rep. ¶ 47 & n.93, so if the Court were to accept the premise that damage models need to replicate real-world prices, that would be akin to requiring most damage models to generate but-for prices that end in “9.” For that to happen, damage models would need to value the product features that a defendant allegedly made misrepresentations about in perfect $10 or $100 increments. There is little reason to believe that, among the vast array of product features that a plaintiff could allege misrepresentations about, any significant number of those product features would have values at those perfect increments. As a consequence, Apple's argument would immunize most companies from damages in practice.
Accordingly, the fact that Dr. Groehn's regression models do not align perfectly with Apple's real-world pricing practices does not render his models unreliable.
2. Relevance
In addition to Apple's arguments about Dr. Groehn's reliability, Apple also claims that Dr. Groehn's regression models are not relevant because they do not translate Plaintiffs' damages theory into economic terms as is required at the class certification stage. Comcast Corp. v. Behrend, 569 U.S. 27, 38 (2013). In particular, Apple asserts that Dr. Groehn (1) wrongly applied the “fraud on the market” theory to damages, (2) wrongly assumed that consumers were harmed even if they purchased Apple devices before iOS 8 was released, and (3) failed to account for individual consumers' disparate expectations about storage space on their Apple devices. Expert Mot. 15-23.
The first two arguments are challenges to the merits of Plaintiffs' damages theory-they do not call into question whether Dr. Groehn's regression models adequately translate Plaintiffs' theory into economic analysis. Such merits issues “are not grounds for exclusion of expert testimony,” Cabrera v. Google LLC, No. 5:11-cv-01263, 2023 WL 5279463, at *15 (N.D. Cal. Aug. 15, 2023), and it is “wholly inappropriate” to raise merits arguments in Daubert briefing. Brown v. Google, LLC, No. 20-cv-3664, 2022 WL 17961497, at *4 n.4 (N.D. Cal. Dec. 12, 2022). Meanwhile, Apple's third argument misunderstands Plaintiffs' damages theory. Plaintiffs' price premium theory measures damages based on “what a reasonable consumer would have paid at the time of purchase.” Pulaski, 802 F.3d at 989 (emphasis added) (citation omitted). The reasonable consumer standard is objective. Lytle, 2024 WL 3915361, at *14; Cabrera, 2023 WL 5279463, at *29; Bailey v. Rite Aid Corp., 338 F.R.D. 390, 407 (N.D. Cal. 2021); see also Montera v. Premier Nutrition Corp., 111 F.4th 1018, 1033 (9th Cir. 2024) (same in the context of New York law). So, Plaintiffs' theory did not require Dr. Groehn to take individual consumer expectations into account.
As such, Dr. Groehn's regression models are not subject to exclusion on the basis of relevance.
In sum, the Court excludes Dr. Groehn's Rule-of-Three Model as unreliable but finds that the Incremental Price Regression and Hedonic Regression Models survive Daubert at the class certification stage. Thus, the Court GRANTS IN PART and DENIES IN PART Apple's motion to exclude.
To be clear, the Court expresses no opinion on whether Dr. Groehn may present both regression models to a jury. Even if both regression models eventually survive a full Daubert assessment, presenting both models to a jury might cause confusion in violation of Federal Rule of Evidence 403. See Daubert, 509 U.S. at 595. That is especially so if the jury would not hear testimony from Dr. Groehn explaining how the two regression models differ, or explaining why the jury might prefer one over the other. Ultimately, Plaintiffs may need to choose one regression model over the other, but that time is not now.
III. MOTION FOR CLASS CERTIFICATION
Having resolved Apple's motion to exclude, the Court turns to class certification. Federal Rule of Civil Procedure 23 governs class certification. Under Rule 23, all classes must satisfy four requirements: numerosity, commonality, typicality, and adequacy. Fed.R.Civ.P. 23(a).
Then, at least one of the three circumstances under Rule 23(b) must apply. Olean Wholesale Grocery Coop., Inc. v. Bumble Bee Foods LLC, 31 F.4th 651, 663 (9th Cir. 2022) (en banc). Here, Plaintiffs invoke Rule 23(b)(3), which adds two more requirements: predominance and superiority. In assessing these six class certification requirements, the Court must conduct a rigorous analysis, Olean, 31 F.4th at 664, and consider the merits of Plaintiffs' claims to the extent they are relevant to Rule 23's requirements. Amgen Inc. v. Conn. Ret. Plans & Tr. Funds, 568 U.S. 455, 466 (2013).
Previously, the Court found that Plaintiffs had established the numerosity, commonality, and superiority elements. Prior Order 11, 23. Apple does not argue that the record has developed in a way that requires reconsideration of those findings. See Opp'n to Class Mot. (“Class Opp'n”), ECF No. 201. So, the Court finds that Plaintiffs have met the numerosity, commonality, and superiority requirements for the same reasons as before and focuses in this Order on typicality, adequacy, and predominance.
A. Typicality and Adequacy
The typicality and adequacy requirements test whether the proposed class representatives and class counsel can satisfactorily represent absent class members. Typicality asks whether the class representatives' claims and defenses are similar to those of absent class members. Hanon v. Dataproducts Corp., 976 F.2d 497, 508 (9th Cir. 1992) (citation omitted). This is a permissive inquiry. A representative's claims and defenses need not be “substantially identical” to those of absent class members; it is enough that the representative's and absent class members' claims and defenses “are reasonably co-extensive.” Castillo v. Bank of Am., NA, 980 F.3d 723, 729 (9th Cir. 2020) (quoting Hanlon v. Chrysler Corp., 150 F.3d 1011, 1020 (9th Cir. 1998)). In turn, adequacy asks whether class representatives and class counsel can be expected to zealously prosecute claims on behalf of the class and to protect the class's interests. Ellis v. Costco Wholesale Corp., 657 F.3d 970, 985 (9th Cir. 2011). In determining adequacy, courts ask, “(1) do the named plaintiffs and their counsel have any conflicts of interest with other class members and (2) will the named plaintiffs and their counsel prosecute the action vigorously on behalf of the class?” Id. (quoting Hanlon, 150 F.3d at 1020).
1. Paul Orshan
Plaintiffs propose Paul Orshan as the sole class representative for the Upgrade Subclass. Although the Court previously rejected Apple's typicality and adequacy challenges to Mr. Orshan, Prior Order 12-14, Apple again challenges Mr. Orshan on those grounds. According to Apple, Mr. Orshan is subject to two unique defenses: (1) that he was not exposed to any representations made by Apple about storage capacity and (2) that he did not rely on any such representations. Class Opp'n 31-32. The existence of such unique defenses may defeat typicality. DZ Rsrv. v. Meta Platforms, Inc., 96 F.4th 1223, 1238 (9th Cir. 2024). But Rule 23 does not require complete identity of claims and defenses between class representatives and absent class members. Castillo, 980 F.3d at 729. The presence of unique defenses only poses an obstacle to class certification when those defenses “threaten to become the focus of the litigation.” Hanon, 976 F.2d at 508. Apple's unique defenses against Mr. Orshan are not so substantial that they risk becoming the focus of this case.
Apple's first defense-lack of exposure to Apple's representations-is not supported by Mr. Orshan's testimony. Apple points to testimony that Mr. Orshan did not speak to an Apple representative or employee before purchasing his 16 GB device. Class Opp'n 31; Dep. of Paul Orshan (“Orshan Dep.”) 85:23-25, ECF No. 201-3. But Mr. Orshan also testified that he researched his options on Apple's website and that he specifically asked to purchase a 16 GB device. Orshan Dep. 85:6-15, 86:4-14. By specifically asking for a 16 GB device, Mr. Orshan showed that he was necessarily aware that Apple labeled its devices by storage capacity, which in turn showed that he was exposed to at least some of Apple's representations about storage capacity. See also infra Section III.B.2. Apple's point about Mr. Orshan's failure to speak with an Apple representative only demonstrates that Mr. Orshan was not exposed to every possible representation about storage capacity. But a class representative need not have been exposed to every single relevant misrepresentation. Even if the exact representations that were received might differ between a class representative and absent class members, the question for typicality is whether the nature of the representations received by the representative and class members is the same. See Ellis, 657 F.3d at 984 (“Typicality refers to the nature of the claim or defense of the class representative, and not to the specific facts from which it arose or the relief sought.”) (quoting Hanon, 976 F.2d at 508); Wolin v. Jaguar Land Rover N. Am., LLC, 617 F.3d 1168, 1175 (9th Cir. 2010) (“Typicality can be satisfied despite different factual circumstances ....”). Like all absent class members, Mr. Orshan claims to have been injured by Apple's representations that its iPhones and iPads came with 16 gigabytes of storage. That is enough for typicality.
Similarly, Apple's argument that reliance issues create a unique defense does not render Mr. Orshan atypical. To establish reliance under the UCL, FAL, and CLRA, Mr. Orshan must show that Apple's misrepresentations about storage space were a cause of his decision to purchase a 16 GB Apple device, but he does not need to show that they were the “sole or even the decisive” cause of his decision. Kwikset, 51 Cal.4th at 327 (UCL and FAL) (emphasis added) (quoting In re Tobacco II Cases, 46 Cal.4th 298, 328 (2009)); see also Lytle, 2024 WL 3915361, at *15 (same for CLRA). Mr. Orshan testified that the alleged misrepresentations were one cause of his decision to purchase a 16 GB device, explaining that he decided on the 16 GB device because he “believed that the stated gigabytes of storage for me was going to be sufficient.” Orshan Dep. 86:12-14.
Apart from the two supposed unique defenses above, Apple also invokes the Court's prior finding that members of the Upgrade Subclass who upgraded wirelessly saw different representations about the size of iOS 8 than those who upgraded via iTunes. Class Opp'n 32 (quoting Prior Order 20). Apple claims that, due to this difference, Mr. Orshan can only represent wireless upgraders because he upgraded wirelessly. Id. The problem is that Plaintiffs have reoriented their theory of liability since the Court originally denied class certification. Now, Plaintiffs assert that Apple is liable for misleading Upgrade Subclass members about available storage capacity at the time of purchase rather than at the time of upgrading. The alleged misrepresentations that Mr. Orshan, wireless upgraders, and iTunes upgraders received at the time of purchase are all fundamentally the same-that Apple's devices had 16 gigabytes of storage.
The Court expresses no opinion on whether it was appropriate for Plaintiffs to shift their theory, or whether Plaintiffs' new focus on misrepresentations at the time of purchase can survive scrutiny on the merits.
Accordingly, the Court finds Mr. Orshan to be a typical and adequate representative for the Upgrade Subclass.
2. Deanna Ness and Margaret Hart
Plaintiffs propose Deanna Ness and Margaret Hart as class representatives for the nationwide Preinstall Subclass. Like with Mr. Orshan, Apple argues that Ms. Ness and Ms. Hart are atypical because they are subject to unique defenses regarding reliance. Class Opp'n 32-33. Also like with Mr. Orshan, those reliance arguments are not enough to disqualify either Ms. Ness or Ms. Hart.
While neither Ms. Ness nor Ms. Hart recalled the exact representations about storage space that they received from Apple, both remembered learning from Apple that their devices were supposed to have 16 gigabytes of storage. Dep. of Deanna Ness (“Ness Dep.”) 159:2-6, 165:2166:9, ECF No. 201-5; Dep. of Margaret Hart (“Hart Dep.”) 43:11-14, 118:6-119:15, ECF No. 201-6. Plaintiffs are “not required to . . . prove individualized reliance on specific misrepresentations or false statements where, as here, those misrepresentations and false statements were part of an extensive [marketing] campaign.” In re Tobacco II, 46 Cal.4th at 328; see also Dep. of Raja Bose 57:11-58:11, ECF No. 194-3 (noting that Apple differentiates between different version of the same iPhone or iPad model by storage capacity); infra Section III.B.2 (elaborating on the extensiveness of Apple's marketing). Thus, being unable to remember specific representations does not raise doubts about reliance. In addition, both Ms. Ness and Ms. Hart testified that they would have acted differently had they known that not all 16 gigabytes of storage were available for personal use, showing that Apple's representations about storage were one cause that contributed their decision to purchase 16 GB Apple devices. Ness Dep. 63:8-17; Hart Dep. 44:18-24, 48:3-49:14. In the end, a jury may not believe Ms. Ness's and Ms. Hart's testimony about reliance, but their testimony is enough to assure the Court that they will not be preoccupied with reliance issues to the point of distraction.
Apple also briefly argues that Ms. Ness was not injured because she never ran out of storage space. Class Opp'n 33. Even if true, the Court already rejected the premise that a consumer needed to run out of space to be harmed, Prior Order 21-22, so this argument does not render Ms. Ness atypical.
Apple also argues that Ms. Ness and Ms. Hart are inadequate because they have credibility issues. Class Opp'n 33-34. However, credibility concerns only pose class certification obstacles in “rare circumstances” when there is “admissible evidence so severely undermining plaintiff's credibility that a fact finder might reasonably focus on plaintiff's credibility.” Mendez v. R+L Carriers, Inc., No. 11-cv-2478, 2012 WL 5868973, at *14 (N.D. Cal. Nov. 19, 2012) (quoting Dubin v. Miller, 132 F.R.D. 269, 272 (D. Colo. 1990)). Such circumstances include when the named plaintiff is “a serial litigant who purchased stock solely to facilitate litigation” and when the named plaintiff “insist[s] that he was not really deceived by the alleged misrepresentation.” DZ Rsrv., 96 F.4th at 1239. Simple inconsistencies in testimony are not enough. Cruz v. Dollar Tree Stores, Inc., Nos. 07-cv-2050, 07-cv-4012, 2009 WL 1458032, at *7 (N.D. Cal. May 26, 2009). Yet, inconsistencies are all that Apple can point to here, and those inconsistencies can readily be explained by fading memories given the decade that has passed since the events relevant to this case.
Recent Ninth Circuit precedent places credibility issues under the rubric of typicality rather than adequacy, DZ Rsrv., 96 F.4th at 1239, but the standards for judging whether credibility issues prevent plaintiffs from representing a class are the same regardless.
As such, the Court finds that Ms. Ness and Ms. Hart are typical and adequate class representatives.
3. Kye Weasner
Plaintiffs propose Kye Weasner as a class representative for two subclasses-the nationwide Preinstall Subclass along with Ms. Ness and Ms. Hart, and the California Preinstall Subclass as the sole representative. In challenging Mr. Weasner, Apple essentially repeats its arguments against the other proposed class representatives, claiming that reliance and credibility issues prevent Mr. Weasner from filling a representative role. For reasons similar to those discussed above, the Court finds that reliance and credibility do not bar Mr. Weasner from acting as a class representative. Mr. Weasner testified that he received representations about his Apple device's storage and that he would have acted differently if he had known that not all 16 gigabytes of storage were available. Dep. of Kye Weasner 64:5-7, 66:14-21, 71:9-17, ECF No. 194-10. That is enough to assuage any concerns that he would be preoccupied with reliance issues. And Mr. Weasner's involvement in a single other class action, as well as minor inconsistencies in his testimony, do not raise significant credibility concerns. So, Mr. Weasner is typical and adequate.
4. Class Counsel
Plaintiffs propose a five-person team as class counsel. They propose William Anderson of Handley Farah & Anderson PLLC as lead counsel, to be supported by an executive committee consisting of Robert Shelquist of Lockridge Grindal Nauen P.L.L.P., Michael McShane of Audet & Partners, LLP, Charles LaDuca of Cuneo Gilbert & LaDuca, LLP, and Jon Herskowitz of Baron & Herskowitz. Class Mot. 16 n.7. Apple does not contest Plaintiffs' pick of lead counsel, and based on proposed counsel's resumes and litigation efforts to date, the Court finds that proposed class counsel are adequate due to their “qualifications, experience, and commitment to the prosecution of this case.” Cottle v. Plaid Inc., 340 F.R.D. 356, 371 (N.D. Cal. 2021). Although the Court has doubts about whether a case of this size is most efficiently litigated with five separate law firms, those doubts do not render class counsel inadequate. Therefore, the Court approves class counsel, but notes that, if and when class counsel move for attorneys' fees, the Court will closely scrutinize counsel's billing records regarding allocation of labor to ensure that this case is prosecuted efficiently and that class members are not paying for duplicative work.
B. Predominance
To satisfy the predominance requirement, Plaintiffs must show that “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) (citation omitted). This does not mean that common questions must outnumber individual questions. In assessing predominance, courts give more important questions greater weight than less important ones. Ruiz Torres v. Mercer Canyons Inc., 835 F.3d 1125, 1134 (9th Cir. 2016). Thus, even if there is only a single common question, class certification may be proper if that single common question is so significant that it predominates over all other individual questions. In re Hyundai & Kia Fuel Econ. Litig., 926 F.3d 539, 557 (9th Cir. 2019) (en banc).
In this case, Apple argues that five important individual questions overwhelm common issues: (1) choice of law, (2) cohesion, (3) materiality and reliance, (4) injury, and (5) damages.
1. Choice of Law
The first step of a predominance analysis is to consider whether choice-of-law issues will require a court to consider multiple states' laws. Stromberg v. Qualcomm Inc., 14 F.4th 1059, 1067 (9th Cir. 2021). This is because even if multiple states recognize similar claims, those claims may have different elements or be subject to different defenses. Such “variances in state law [may] overwhelm common issues and preclude predominance for a single nationwide class.” Mazza v. Am. Honda Motor Co., 666 F.3d 581, 596 (9th Cir. 2012), overruled on other grounds by Olean, 31 F.4th 651. Because Plaintiffs here brought state claims, the Court applies the choice-of-law rules from the forum state, California. Stromberg, 14 F.4th at 1067. Under California law, there are two choice-of-law tests, and courts must decide which to apply based on the circumstances. Wash. Mut. Bank, FA v. Superior Ct., 24 Cal.4th 906, 914 (2001). When there is a contractual choice-of-law clause, courts apply Nedlloyd Lines B.V. v. Superior Court, 3 Cal.4th 459 (1992), to determine if the claims at issue fall within the scope of that choice-of-law clause and if such clause is enforceable. Wash. Mut. Bank, 24 Cal.4th at 914-15. When there is no controlling choice-of-law clause, courts instead use the governmental interest test. Id.
a. Contract
Plaintiffs rely on the choice-of-law clause in the Software License Agreement (“SLA”) for iOS 8, so the Court begins with the Nedlloyd test. The first step in the Nedlloyd test is to determine whether the claims at issue fall within the scope of the relevant choice-of-law clause. Nedlloyd, 3 Cal.4th at 468. The SLA's choice-of-law clause reads, “This License will be governed by and construed in accordance with the laws of the State of California, excluding its conflict of law principles.” Decl. of Robert K. Shelquist, Ex. 13 (“SLA”) § 12, ECF No. 194-14. The use of broad language like “governed by” means that this choice-of-law clause “appl[ies] to all causes of action arising from or related to” the SLA. Nedlloyd, 3 Cal.4th at 468. This holds true even if those causes of action are non-contractual, “including tortious breaches of duties emanating from the agreement or the legal relationships it creates.” Id. at 470.
The parties disagree on whether Plaintiffs' consumer protection claims arise from or relate to the SLA. There is undoubtedly some connection between Plaintiffs' claim and the SLA- Plaintiffs allege iOS 8 took up too much space and the SLA governs the use of iOS 8. But the parties dispute whether this nexus between Plaintiffs' claims and the SLA is close enough to justify applying the choice-of-law clause. The law is not entirely clear on this point. Nedlloyd states that choice-of-law clauses should apply broadly to “all disputes arising out of the transaction or relationship.” Id. at 469. However, Nedlloyd does not define what “arising out of or related to” means. And surveying other areas of California law shows that the definition of “arising out of or related to” can vary depending on context. For example, in the insurance context, “California courts have consistently given a broad interpretation to the terms ‘arising out of' or ‘arising from,'” holding that such language “does not import any particular standard of causation or theory of liability” and only requires “a minimal causal connection or incidental relationship.” Medill v. Westport Ins. Corp., 143 Cal.App.4th 819, 830 (2006) (quoting Acceptance Ins. Co. v. Syufy Enters., 69 Cal.App.4th 321, 328 (1999)). By contrast, in the workers' compensation context, an injury does not “arise out of the employment” unless it was proximately caused by the employment. Guerra v. Workers' Comp. Appeals Bd., 246 Cal.App.4th 1301, 1307 (2016).
Since there is no clear precedent explaining how closely claims must be tethered to a contract to arise out of or relate to that contract, the Court's task is to “predict how the highest state court would decide the issue using intermediate appellate court decisions, decisions from other jurisdictions, statutes, treaties and restatements for guidance.” Glendale Assocs., Ltd. v. N.L.R.B., 347 F.3d 1145, 1154 (9th Cir. 2003) (quoting N.L.R.B. v. Calkins, 187 F.3d 1080, 1089 (9th Cir. 1999)). In doing so, the Court finds most persuasive the line of California cases addressing when non-contract claims are sufficiently connected to a contract to be subject to an arbitration clause. Determining whether non-contract claims fall within the scope of an arbitration clause is closely analogous in nature to determining when non-contract claims fall within the scope a choice-of-law clause. California courts also similarly favor enforcing both arbitration and choice-of-law clauses, further underscoring the similarity between how California law treats the two types of clauses. Compare Ramos v. Superior Ct., 28 Cal.App. 5th 1042, 1051 (2018) (“Any doubts concerning the scope of arbitrable issues will be resolved in favor of arbitration.”), with Nedlloyd, 3 Cal.4th at 465 (“California courts shall apply . . . a strong policy favoring enforcement of such [choice-of-law] provisions.”).
California courts addressing arbitration clauses have consistently held that, when an arbitration clause applies to claims “arising from or related to” the relevant contract, all claims with “their roots in the relationship between the parties which was created by the contract” are subject to arbitration, regardless of whether those claims are contractual. Howard v. Goldbloom, 30 Cal.App. 5th 659, 664-65 (2018) (citations omitted); see also Ramos, 28 Cal.App. 5th at 1051-52; Rice v. Downs, 248 Cal.App.4th 175, 186 (2016); Bos Material Handling, Inc. v. Crown Controls Corp., 137 Cal.App.3d 99, 105 (1982). This relationship-oriented test is consistent with Nedlloyd, which likewise grounds its choice-of-law test on the importance of maintaining clear boundaries around the relationships created by contract. Nedlloyd, 3 Cal.4th at 469-470 (causes of action that “emanat[e] from the agreement or the legal relationship it creates” fall within the scope of the agreement's choice-of-law clause) (emphasis added). Therefore, the Court predicts that the California Supreme Court would use a “rooted in the relationship” test or something similar to determine whether claims arise from or relate to contracts for choice-of-law purposes.
The Court finds that none of Plaintiffs' claims are rooted in the relationship created by the SLA. For a consumer to agree to the SLA and enter into a contract with Apple, she needs to “US[E] [HER] iOS DEVICE OR DOWNLOAD[] A SOFTWARE UPDATE.” SLA at p. 1. Because the consumer cannot use her device or download an update until after she purchases her device, the SLA does not create a relationship between the consumer and Apple until after purchase. However, Plaintiffs' claims are that Apple misled them about the storage capacity of various Apple devices before purchase, and that those misrepresentations caused injury by influencing them to purchase those devices. Given this sequence of events, Plaintiffs' and absent class members' claims could not have been rooted in a relationship created by the SLA since no such relationship existed at the time Plaintiffs and absent class members were allegedly injured. This is especially so for the Upgrade Subclass. Many Upgrade Subclass members would have purchased their Apple devices before iOS 8 and the accompanying SLA even existed, so any claims that those Upgrade Subclass members were injured when purchasing their Apple devices are obviously not rooted in the SLA for iOS 8.
The Court's previous conclusion that the SLA's choice-of-law clause applies to the Upgrade Subclass no longer holds because Plaintiffs have changed their theory of liability for the Upgrade Subclass. The Court previously concluded that the choice-of-law clause applied to the Upgrade Subclass because Plaintiffs were claiming injury from misrepresentations made during the upgrade process, and the SLA explicitly controlled the upgrade process. SLA § 1(b) (“The terms of this License will govern any iOS Software Updates provided by Apple that replace and/or supplement the Original iOS Software product ....”). Now that Plaintiffs are focusing on misrepresentations at the time of purchase, this reasoning no longer fits with the Upgrade Subclass's claims. To the extent that Plaintiffs rely on similar choice-of-law clauses in the SLAs for earlier versions of iOS, those clauses are not relevant here because Plaintiffs' claims do not involve any of those earlier iOS versions.
The Ninth Circuit's decision applying California choice-of-law rules in In re Hyundai & Kia Fuel Economy Litigation, 926 F.3d 539 (9th Cir. 2019), further reinforces this conclusion. There, the en banc court explained that claims based on advertising misrepresentations about a product do not arise from any contract formed when a consumer purchases that product. Id. at 561 n.5. And as the Kia & Hyundai court further observed, when the relevant contract says nothing about the substance of allegedly misleading advertising, there is even less reason to think that claims about misleading advertising arose out of that contract. So too here. Plaintiffs claim that Apple falsely advertised available storage capacity, but the SLA says nothing about storage capacity or the amount of space that iOS 8 or other apps would take up.
Accordingly, Plaintiffs' claims fall outside the scope of the SLA's choice-of-law clause, and the Court does not need to assess whether the clause is enforceable. Either way, the SLA's choice-of-law clause does not apply here.
b. Governmental Interests
Since there is no applicable choice-of-law clause, the Court turns to the governmental interest test. California courts apply California law by default “unless a party litigant timely invokes the law of a foreign state.” In re Hyundai & Kia, 926 F.3d at 561 (quoting Wash. Mut. Bank, 24 Cal.4th at 919). In that case, it is the burden of the party invoking foreign law, even in the context of class certification, to show that foreign law should apply. Id. at 561-63 (finding predominance to be satisfied when objecting parties failed to meet their burden to show that different states' law applied). To meet its burden, the party seeking to apply foreign law must prove that (1) the foreign law materially differs from California law; (2) there is a “true conflict” in the sense that each state has an interest in applying its own law in the case; and (3) that the foreign state's interest would be more impaired than California's interest if California law applied. Id. at 562 (citations omitted).
Apple, as the proponent of foreign law in this case, bears the burden of showing that foreign law applies. It has met its burden as to the nationwide Upgrade and Preinstall Subclasses. First, as Apple demonstrates and Plaintiffs do not contest, consumer protection laws vary from state-to-state on material issues such as reliance, scienter, and statutes of limitations. Class Opp'n, App. A, ECF No. 201-1. Second, there is a true conflict because each state “has an interest in balancing the range of products and prices offered to consumers with the legal protections afforded to them.” Mazza, 666 F.3d at 592. Each state also has an interest in “being able to assure individuals and commercial entities operating within its territory that applicable limitations on liability set forth in the jurisdiction's law will be available.” Id. at 592-93 (quoting McCann v. Foster Wheeler LLC, 48 Cal.4th 68, 97-98 (2010)). Third, Apple has shown that other states' interests would be more impaired than California's because “the place of the wrong has the predominant interest.” Id. at 593 (quoting Hernandez v. Burger, 102 Cal.App.3d 795, 802 (1980)). For nationwide classes, the place of wrong will vary, therefore requiring the Court to consider several states' laws and causing choice-of-law issues to defeat predominance.
However, Apple has failed to show that choice-of-law issues create predominance problems for the California Preinstall Subclass. Although Apple has shown that foreign law differs from California law and that there is a true conflict, it has not shown that foreign states' interests would be impaired. The California Preinstall Subclass is defined to include only individuals who purchased their Apple devices in California. Class Mot. 11. For those individuals, California is the place of the wrong, so California has the greatest interest. Apple speculates that individuals who purchased their devices in California may have nonetheless received the alleged misrepresentations elsewhere. Class Opp'n 27 n.4. But that speculation is not enough to meet Apple's burden. And in any case, the place of the wrong is “the state where the last event necessary to make the actor liable occurred.” Mazza, 666 F.3d at 593-94. The last event necessary for liability is the purchasing of an Apple device, and that event occurred in California for the California Preinstall Subclass.
As a result, the Court finds that choice-of-law issues defeat predominance for the nationwide Upgrade and Preinstall Subclasses but not for the California Preinstall Subclass.
2. Cohesion
In addition to choice-of-law issues, a lack of “cohesion among [class] members [due to being] exposed to quite disparate information from various representatives of the defendant” can defeat predominance. Stearns v. Ticketmaster Corp., 655 F.3d 1013, 1020 (9th Cir. 2011), abrogated on other grounds by Comcast, 569 U.S. 27; see also Berger v. Home Depot USA, Inc., 741 F.3d 1061, 1069 (9th Cir. 2014) (cohesion was lacking when there were five different allegedly misleading contracts that each required separate analysis and when signs discussing the allegedly misleading subject varied between stores and over time), abrogated on other grounds by Microsoft Corp. v. Baker, 582 U.S. 23 (2017). Put differently, Plaintiffs cannot satisfy predominance if “many class members were never exposed to the allegedly misleading [statements].” Mazza, 666 F.3d at 595. This reflects common sense. In a case about misrepresentations, the core misconduct alleged is that the defendant made false and misleading statements. If the statements alleged to be misleading varied in material ways across the putative class, or if a significant portion of class members were not exposed to those statements, then the very core of the case would vary on an individualized basis.
However, cohesion does not require all alleged misrepresentations to be exactly identical. That is because courts take “the common sense approach that the class is united by a common interest in determining whether a defendant's course of conduct is in its broad outlines actionable.” DZ Rsrv., 96 F.4th at 1236 (quoting Blackie v. Barrack, 524 F.2d 891, 902 (9th Cir. 1975)). Slight differences in wording, format, or context do not displace common questions when the underlying misrepresentation is materially the same across all class members. Id. (collecting cases); see also Walker v. Life Ins. Co. of the Sw., 953 F.3d 624, 631 (9th Cir. 2020) (affirming the district court's reasoning that there was no predominance issue when all proposed class members received the same illustration alleged to be misleading even though some of the information surrounding those illustrations may have varied individually); Miller v. Travel Guard Grp., Inc., No. 21-cv-09751, 2023 WL 7106479, at *14 (N.D. Cal. Sept. 15, 2023) (no cohesion issue when all alleged misrepresentations communicated the same three misleading ideas). The key question is whether the nature of the alleged misrepresentations is the same for the entire class. Cf. Ellis, 657 F.3d at 984 (“Typicality refers to the nature of the claim or defense of the class representative, and not to the specific facts from which it arose or the relief sought.”).
Plaintiffs have shown that their proposed classes are sufficiently cohesive under this standard. Although Apple tries to frame this case as one about false labeling and casts doubt about whether any significant number of class members saw Apple's labels, Class Opp'n 15, that is too narrow a view of Plaintiffs' claims. Plaintiffs challenged not just Apple's labels, but Apple's entire marketing and advertising campaign identifying its devices as offering 16 gigabytes of storage. Second Consol. & Am. Class Action Compl. (“Compl.”) ¶¶ 1, 19, 38, ECF No. 189. Considering Apple's marketing and advertising as a whole, the Court finds that Apple's representations about the storage capacities of its devices were ubiquitous and consistent. Storage space is how Apple differentiated between different versions of devices within the same model family. See Dep. of Raja Bose 56:25-58:11. Because Apple's devices were identified in part by their advertised storage capacity, a consumer would necessarily need to know the advertised storage capacity to purchase a device. For instance, a consumer could not simply ask for an iPad Air-she would need to identify whether she wanted the 16 GB, 32 GB, 64 GB, or 128 GB version. Am. Groehn Rep. tbl.5. Given how Apple made storage capacity a central component of its marketing, “there [is] little doubt that almost every class member [has] been exposed to” Apple's alleged misrepresentations about storage capacity. Mazza, 666 F.3d at 596.
3. Materiality and Reliance
A plaintiff seeking to certify classes under the UCL, FAL, and CLRA must also grapple with the issues of materiality and reliance. The CLRA treats those two issues differently than the UCL and FAL, so the Court addresses the CLRA separately.
a. CLRA
To succeed on their CLRA claim, plaintiffs must prove that they and absent class members relied on Apple's alleged misrepresentations. Rojas-Lozano v. Google, Inc., 159 F.Supp.3d 1101, 1112 (N.D. Cal. 2016) (citations omitted). Reliance can often “raise so many individualized questions as to defeat predominance.” Walker, 953 F.3d at 631. But California law leaves putative classes with a way to avoid those individualized questions: When an alleged misrepresentation is material, a rebuttable presumption of reliance arises. Mass. Mut. Life Ins. Co. v. Superior Ct., 97 Cal.App.4th 1282, 1292 (2002) (quoting Vasquez v. Superior Ct., 4 Cal.3d 800, 814 (1971)). So, if materiality is a common question, then a finding of materiality will raise a classwide presumption of reliance that, unless Apple is able to rebut the presumption, elides any individualized issues that may arise.
Before beginning its analysis of the facts in this case, the Court must first determine what materiality and reliance mean in the CLRA context. Despite the pervasiveness of CLRA cases, the concepts of materiality and reliance are surprisingly muddled in the case law. Start with materiality. Some courts hold that a misrepresentation is material “if it induced the consumer to alter his position to his detriment.” Tucker v. Pac. Bell Mobile Servs., 208 Cal.App.4th 201, 222 (2012) (quoting In re Vioxx Class Cases, 180 Cal.App.4th 116, 129 (2009)); see also Algarin v. Maybelline, LLC, 300 F.R.D. 444, 453 (S.D. Cal. 2014); True v. Am. Honda Motor Co., 520 F.Supp.2d 1175, 1182 (C.D. Cal. 2007). Others take a more relaxed approach, holding that a misrepresentation is material if “a reasonable man would attach importance to its existence or nonexistence in determining his choice of action in the transaction in question.” Steroid Hormone Prod. Cases, 181 Cal.App.4th 145, 157 (2010) (quoting Engalla v. Permanente Med. Grp., Inc., 15 Cal.4th 951, 977 (1997)); see also Gutierrez v. Carmax Auto Superstores Cal., 19 Cal.App. 5th 1234, 1258 (2018) (a fact is material when “a reasonable consumer would deem it important in determining how to act in the transaction at issue”); Hinojos v. Kohl's Corp., 718 F.3d 1098, 1107 (9th Cir. 2013) (in the UCL context, a misrepresentation is material “if a reasonable consumer would attach importance to it”); Rasmussen v. Apple Inc., 27 F.Supp.3d 1027, 1033 (N.D. Cal. 2014).
Courts have been similarly inconsistent in their approaches to reliance. Some courts strictly require plaintiffs to show that they would have acted differently in the absence of the alleged misrepresentation in order to prove reliance. E.g., Gray v. Dignity Health, 70 Cal.App. 5th 225, 243 n.13 (2021); Buckland v. Threshold Enters., Ltd., 155 Cal.App.4th 798, 810 (2007), disapproved of on other grounds by Kwikset, 51 Cal.4th 310. Others adopt the less demanding test that, while reliance requires plaintiffs to show that a misrepresentation “was an immediate cause” of injury, the misrepresentation need not be the “sole or even decisive cause of the injury.” Kwikset, 51 Cal.4th at 327 (emphasis added); see also Lytle, 2024 WL 3915361, at *15 (quoting Moore v. Mars Petcare US, Inc., 966 F.3d 1007, 1020 (9th Cir. 2020)).
Faced with this chaos, the Court looks to the sources of these competing tests to determine which ones it should adopt. On that front, the path to resolving conflicts within the reliance case law is more apparent, so the Court begins there. The less demanding reliance test-requiring only that the misrepresentation be a cause of injury-clearly has the stronger pedigree. It is based on California Supreme Court precedent, which controls over California Court of Appeal cases. The Ninth Circuit has also adopted the less demanding test in multiple precedential opinions, and those opinions strictly bind this Court. Therefore, the Court applies the less demanding test for reliance.
The choice of materiality tests is less straightforward. The less demanding materiality test has some basis in California Supreme Court and Ninth Circuit precedent, unlike the stricter materiality test. But those precedents are not directly on point; rather, they are analogies from similar laws. See Engalla, 15 Cal.4th at 973-74 (common law fraud in the inducement); Hinojos, 718 F.3d at 1107 (UCL). When that observation is combined with the Court's adoption of the less demanding reliance test, though, the correct choice of materiality tests becomes more apparent. Under the Court's reliance test, a misrepresentation explicitly does not need to be the decisive cause of a consumer's injury for the consumer to have relied on that misrepresentation. Kwikset, 51 Cal.4th at 327. The strict materiality test, however, does require the misrepresentation to be the decisive cause because it requires the misrepresentation to alter the consumer's behavior. However, it is nonsensical to require the misrepresentation to be the decisive cause in order to establish a presumption of reliance while directly establishing reliance specifically does not require the misrepresentation to be the decisive cause. The strict materiality test is thus incompatible with the Court's reliance test, so the Court adopts the less demanding materiality test.
To summarize, then, plaintiffs must show that a reasonable consumer would attach importance to the alleged misrepresentations to demonstrate materiality. And to show reliance, plaintiffs must show that the alleged misrepresentations were a cause of their decision to purchase Apple devices though not necessarily the sole or decisive cause.
Applying these rules, the Court finds that materiality is a common question and that Apple's alleged misrepresentations about storage capacity are material across the entire class. The reasonable consumer standard is an objective test, so there is no need to look at individual circumstances. Bailey, 338 F.R.D. at 399 (citation omitted); cf. DZ Rsrv., 96 F.4th at 1035 (materiality for purposes of California common-law fraud uses an objective “reasonable consumer” that created common questions); Yokoyama v. Midland Nat. Life Ins. Co., 594 F.3d 1087, 1089, 1093 (9th Cir. 2010) (finding that Hawaii law applying a reasonable consumer standard posed common questions). Rather, the “reasonable consumer” is a hypothetical individual reflecting the aggregated preferences and motivations of “a significant portion of the general consuming public or of targeted consumers, acting reasonably in the circumstances.” Moore, 966 F.3d at 1017 (citation omitted). In this case, Dr. Groehn and Apple's expert have provided survey evidence that a significant portion of consumers found storage capacity to be important. Am. Groehn Rep. figs.12 & 13; Am. Expert Rep. of Sarah Butler (“Am. Butler Rep.”) ¶¶ 78-79, ECF No. 151-3. Even though storage capacity may not have been the single most important factor to consumers, as the Court previously held, it did not need to be. Prior Order 21. Therefore, Apple's alleged misrepresentations are material, and a classwide presumption of reliance arises.
Previously, the Court concluded that Apple had produced sufficient evidence to rebut the presumption of reliance. Prior Order 20-21. To reach that conclusion, the Court focused on evidence showing that a substantial portion of consumers would not have changed their decisions to buy 16 GB Apple devices even if told that iOS 8 limited available storage capacity. Id. (citing Am. Butler Rep. tbls 13 & 14). In doing so, the Court assumed without deciding that the more stringent test for reliance applied. The Court made this assumption because Apple argued that reliance required a change in consumer behavior, and Plaintiffs did not challenge that formulation of reliance. In this round of briefing, though, Plaintiffs set forth a competing view of reliance that required the Court to more closely analyze the standard for reliance. With the benefit of this further briefing and additional scrutiny, the Court finds that, contrary to its previous assumption, reliance depends only on whether the alleged misrepresentation was a cause of the decision to purchase a 16 GB device. Lytle, 2024 WL 3915361, at *15.
On that front, Apple argues that Apple's representations about storage capacity were not a cause at all for many putative class members because they could have been exposed to information from other sources. Apple points to the potential for class members to have encountered media articles or to have had conversations with salespeople, family, and friends. Class Opp'n 6-8.
Apple also suggests that class members' previous experience with other Apple products may have given them unique knowledge. Id. According to Apple, any of these could have revealed that iOS consumed storage space. Id. This argument does not take Apple far enough, though. As the Ninth Circuit emphasized in an earlier memorandum opinion in this case, consumers' knowledge that iOS 8 takes up some space is not the key issue; what matters is consumers' understanding of how much space will be taken up. Ninth Cir. Mem. Opn. 3, ECF No. 66. At best, Apple shows that a substantial portion of consumers understood that iOS 8 would take up some space. Am. Butler Rep. tbls.10 & 12. Apple offers no evidence showing how many consumers may have known the exact amount of space taken by iOS 8. That number could be significant, or it could be vanishingly small. If individualized issues involving consumers' external knowledge are “limited to a small number of class members,” they do not defeat predominance. Van v. LLR, Inc., 61 F.4th 1053, 1067 n.11 (9th Cir. 2023). But the Court cannot tell how many consumers knew the exact size of iOS 8 or received relevant external information. Since Apple leaves the Court to speculate about the number of consumers with outside knowledge, it fails to meet its burden to rebut the assumption of reliance. Id. at 1068 (“We do not permit a defendant to support its invocation of individualized issues with mere speculation.”).
As a result, materiality and reliance issues do not defeat predominance as to Plaintiffs' CLRA claims.
b. UCL and FAL
The UCL and FAL require plaintiffs to show only that “members of the public are likely to be deceived.” Pulaski, 802 F.3d at 985-86 (quoting In re Tobacco II, 46 Cal.4th at 312, 320). There is no requirement that plaintiffs establish proof of reliance as to absent class members, id., so reliance cannot raise individualized issues. Materiality, meanwhile, is only relevant to certain defenses that Apple can raise. Hinojos, 718 F.3d at 1107. And in any case, as the Court found above in its analysis of the CLRA, materiality can be resolved on a classwide basis.
Apple takes issue with the Court's statement of law that the UCL and FAL do not require proof of reliance for absent class members. Class Opp'n 19-20. Apple's arguments on this point are not easy to follow-on one hand Apple seems to acknowledge that Proposition 64's amendments to the UCL do not require proof of reliance for absent class members, and on the other Apple seems to suggest that not requiring such proof runs counter to Proposition 64. Id. But regardless of what Apple means to argue, the law is clear. The California Supreme Court explicitly interpreted Proposition 64 to mandate a reliance requirement “only [as] to the class representatives, and not [as to] all absent class members.” In re Tobacco II, 46 Cal.4th at 306 ; see also Mazza, 666 F.3d at 595-96 (“[T]he California Supreme court reconfirmed that class members do not need to demonstrate individualized reliance, and that Proposition 64 imposes its reliance requirements only on the named plaintiff, not unnamed class members.”); Walker, 953 F.3d at 630-31 (similar). Apple is wrong to suggest that California law requires proof of reliance for absent class members.
The California Supreme Court spoke in terms of statutory standing requirements rather than reliance but explained that standing is established by reliance.
Apple's remaining arguments fare no better. Apple asserts that, by certifying a class without requiring proof of reliance for absent class members, the Court impermissibly “arm[s] absent class members with additional substantive rights.” Class Opp'n 20. In support of this proposition, Apple cites to TransUnion LLC v. Ramirez, 594 U.S. 413 (2021), and Shady Grove Orthopedic Associates, P.A. v. Allstate Insurance Co., 559 U.S. 393 (2010). Id. Unfortunately, Apple does not elaborate on these citations. And the Court cannot intuit how these two cases fit together because they discuss very different issues. TransUnion concerns the concreteness of the injury necessary for Article III standing while Shady Grove concerns the extent to which Rule 23 preempts state law. Finding no apparent connection between the two cases, the Court addresses each of their issues as separate arguments.
Standing. To the extent Apple contends that lack of reliance equals lack of standing since reliance is needed for injury, the Ninth Circuit has already rejected this exact argument in Mazza. 666 F.3d at 594-95. There, the Ninth Circuit explained that, despite the lack of reliance, the plaintiffs established there was injury for Article III purposes based on the theory that class members paid more for a product than they otherwise would have paid or that they bought the product when they otherwise would not have. Id. This is a classic monetary injury satisfying Article III's requirements. TransUnion, 594 U.S. at 425. It is also the same theory of injury that Plaintiffs adopt in this case. Compl. ¶ 37 (“Had Plaintiffs known that the operating system and other pre-installed software consumes a substantial portion of the storage capacity of the Devices, they would have reconsidered their decisions to purchase Devices, or would have paid less.”). Thus, Article III standing is no obstacle to class certification.
Preemption. The Court also finds unavailing any suggestion that Rule 23 preempts the In re Tobacco II rule requiring reliance for class representatives but not absent class members. In deciding whether Rule 23 preempts state law, courts apply a two-step test. First, they ask whether Rule 23 conflicts with the relevant state law because they give different answers to the “same question,” meaning that one provides that a suit may be maintained as a class action while the other provides that the suit may not be maintained as a class action. Shady Grove, 559 U.S. at 398-99. Second, if there is a conflict, Rule 23 prevails “unless it exceeds statutory authorization or Congress's rulemaking power.” Id. at 398. The Rules Enabling Act, 28 U.S.C. § 2072, sets the parameters of this second step.
Any argument that Rule 23 preempts In re Tobacco II fails at the first step. Nothing in the text of Rule 23 mandates that the elements of a class representative's claims be identical to the elements of an absent class member's claims. The closest that Rule 23 comes to such a condition is through typicality, which requires that “the claims or defenses of the representative parties are typical of the claims or defenses of the class.” Fed.R.Civ.P. 23(a)(3). On its face, the typicality requirement says nothing about the elements of the class representatives' or absent class members' claims. Case law also clarifies that it is not the elements that matter, but rather the nature of the claims and defenses that matters. Ellis, 657 F.3d at 984. Further, typicality does not require complete identity of claims and defenses between class representatives and absent class members. Their claims and defenses need only be “reasonably co-extensive.” Castillo, 980 F.3d at 729. Rule 23 permits class certification even when class representatives are subject to defenses that absent class members are not, so long as those unique defenses do not “threaten to become the focus of the litigation.” Hanon, 976 F.2d at 508.
Thus, typicality can easily be reconciled with the In re Tobacco II rule that only class representatives, but not absent class members, need to establish reliance. The reliance requirement for class representatives is, in effect, a unique defense that applies only to class representatives but not to absent class members. If there are meritorious arguments against a class representative's reliance that would threaten to become the focus of litigation, then the representative is not typical, and a class may not be certified. In re Tobacco II does not require otherwise, so it does not conflict with that outcome. If, however, there are no meritorious reliance arguments against the class representative, reliance would not threaten to become the focus of litigation and the class could be certified. Again, In re Tobacco II does not conflict with that outcome by requiring otherwise. In short, Rule 23 is fully consistent with In re Tobacco II.
Even if there were a conflict, it would violate the Rules Enabling Act for Rule 23 to preempt the In re Tobacco II reliance rule by requiring plaintiffs to prove reliance for absent class members. The Rules Enabling Act empowers the U.S. Supreme Court to promulgate binding procedural rules, including, as relevant here, the Federal Rules of Civil Procedure. 28 U.S.C. § 2072(a). However, the authority to do so is limited: such a rule cannot “abridge, enlarge or modify any substantive right.” Id. § 2072(b). This limitation ensures that rules promulgated under § 2072 “really regulat[e] procedure.” Shady Grove, 559 U.S. at 407 (plurality opinion) (alteration in original) (quoting Sibbach v. Wilson & Co., 312 U.S. 1, 14 (1941)). A rule violates that limitation when it “alter[s] the rules of decision by which the court will adjudicate [substantive] rights.” Id. (cleaned up). If Apple were correct that Rule 23 requires plaintiffs to prove reliance for absent class members, that would “amount to abridging a substantive right,” since plaintiffs would have to prove reliance for absent class members in federal class actions but not in state class actions. DZ Rsrv., 96 F.4th at 1238. So “[c]ontrary to [Apple's] contention, the Rules Enabling Act requires application” of In re Tobacco II, it does not prohibit such application. Id.
The Court applied the Shady Grove plurality's test for determining whether Rule 23 would violate the Rules Enabling Act as applied. The plurality test is strictly based on prior Supreme Court precedent, so even if it is not technically binding, it is a persuasive statement of binding precedent. See Abbas v. Foreign Pol'y Grp., LLC, 783 F.3d 1328, 1336 (D.C. Cir. 2015). Further, the concurrence's proposed test is even broader than the plurality's, so applying the concurrence's test would lead to the same result. See Shady Grove, 559 U.S. at 422-23 (Stevens, J., concurring).
In summary, issues surrounding materiality and reliance do not defeat class certification of Plaintiffs' UCL and FAL claims.
4. Injury
Most of Apple's arguments regarding injury repackage those that the Court already addressed above in the Cohesion and Materiality and Reliance sections of this Order, and the Court does not repeat its analysis of those arguments here. Apple raises only two additional arguments as to injury, one of which is better addressed as an attack on Plaintiffs' theory of damages. Accordingly, the only argument that the Court discusses in this section is Apple's claim that consumers who knew iOS 8 took up storage capacity were not injured because they received what they expected. Class Opp'n 22. This “benefit of the bargain” defense is only viable when the alleged misrepresentations are not material. Hinojos, 718 F.3d at 1107 (citing Kwikset, 51 Cal.4th at 332-33). As the Court already concluded, the alleged misrepresentations here are material. Supra Section III.B.3. Accordingly, Apple cannot raise a “benefit of the bargain” defense, and Apple's argument about consumer knowledge poses no obstacle to predominance.
5. Damages
For Plaintiffs to demonstrate predominance, they “must proffer a damages model showing that ‘damages are susceptible of measurement across the entire class for purposes of Rule 23(b)(3).'” Vizcarra, 339 F.R.D. at 553 (quoting Comcast, 569 U.S. at 35). Plaintiffs have offered two potential damage models for this purpose-the Incremental Price Regression and Hedonic Regression Models. Am. Groehn Rep. ¶¶ 94-111. Apple raises a bevy of objections, but those are largely Daubert issues that the Court resolved in Section II above. Daubert issues aside, Apple makes just two arguments against Plaintiffs' damages theory. The Court finds that neither ultimately bars class certification.
First, Apple contends that Plaintiffs improperly applied a “fraud on the market” theory to measure damages. Class Opp'n 22; see also Expert Mot. 17-18. This is a somewhat curious argument because “fraud on the market” is not a theory of damages at all. Instead, it is a theory of reliance in federal securities fraud litigation. Amgen, 568 U.S. at 461. “Fraud on the market” flows from the premise that, in efficient markets, the market price of shares reflects and incorporates all publicly available information. Id. at 461-62. As such, when an investor purchases a share at market price, she implicitly relies on all publicly available information, including any alleged misrepresentations. Id. As a result, in efficient markets, there is a rebuttable presumption of reliance. Id. However, Plaintiffs here do not attempt to prove reliance through a “fraud on the market” theory by claiming that Apple's alleged misrepresentations were incorporated into the prices of Apple devices. Instead, the presumption of reliance in this case arises from the materiality of the alleged misrepresentations. Mass. Mut., 97 Cal.App.4th at 1292.
Rather than objecting to use of the “fraud on the market” theory, Apple seems to be criticizing Plaintiffs for adopting the theory's underlying premise. Apple appears to reason as follows: (a) the market for iPhones and iPads is not an efficient market; (b) because the market is not efficient, some information will not be reflected in price, and even material misrepresentations may not be reflected in price; (c) since not all information is necessarily incorporated into the price of iPhones and iPads, “damages resulting from the alleged misrepresentations will not possibly be uniform” because difference consumers might value storage differently. Expert Mot. 18 (citing In re POM Wonderful LLC, MDL No. 2199, 2014 WL 1225184, at *1-4 (C.D. Cal. Mar. 25, 2014)). From this, Apple concludes that “it is wrong to simply assume that all consumers would have paid less absent the misrepresentation,” and therefore that Plaintiffs' assumption of uniform damages invalidates their theory of damages. Id.
This line of reasoning might hold water if the measure of Plaintiffs' price premium damages were the price that the market would have set absent the alleged misrepresentations. But that is not the measure of damages. Instead, damages are measured by what a reasonable consumer would be willing to pay. Pulaski, 802 F.3d at 989. The reasonable consumer is a single, hypothetical person and would be the same across the entire class. See Moore, 966 F.3d at 1017. For this reason, damages would be uniform as Plaintiffs presumed.
Although not necessary to the Court's reasoning, the Court observes that few, if any, markets for consumer products will be efficient. If the lack of efficiency means that “damages resulting from the alleged misrepresentations will not possibly be uniform,” In re POM Wonderful, 2014 WL 1225184, at *4, that would lead to the seemingly categorical conclusion that cases regarding misrepresentations about consumer products can never be maintained as class actions. This, of course, would be a dramatic result that flies in the face of the many consumer class actions that courts have certified.
Apple's final argument is that it makes no sense for Plaintiffs to claim that members of the Upgrade Subclass were injured at the time of purchase before iOS 8 came into existence. Class Opp'n 26. If that is true-and the Court expresses no opinion on whether these claimed damages are viable-that would negate damages as to the entire Upgrade Subclass. So even if Apple's skepticism about Plaintiffs' theory is founded in this regard, it would not create individual issues that defeat predominance.
As such, there are no damages issues that prevent class certification.
IV. CONCLUSION
For the foregoing reasons, the Court GRANTS IN PART and DENIES IN PART Apple's motion to exclude and GRANTS IN PART and DENIES IN PART Plaintiffs' motion for class certification. The Court CERTIFIES the California Preinstall Subclass and declines to certify the nationwide Upgrade and Preinstall Subclasses. Plaintiffs ask the Court to certify the two nationwide classes as liability-only classes under Rule 23(c)(4) if the Court declines to certify them under Rule 23(b)(3), Class Mot. 33-35, but the Court declines that request as well because any such nationwide liability-only classes would still face intractable individualized choice-of-law issues.
Within twenty-one (21) days of this Order, the parties shall submit a stipulated schedule for this case up to and including the trial setting conference. Though the parties need not agree on dates for the pretrial conference or trial, they should indicate the time that they expect the case to be ready for trial. If the parties are unable to reach agreement, instead of a stipulation, they shall file a joint status report containing their competing scheduling proposals.
IT IS SO ORDERED.