Nalick v. Seagate Technology CA1/1

Filed 4/4/24 Nalick v. Seagate Technology CA¹/1 NOT TO BE PUBLISHED IN OFFICIAL REPORTS California Rules of Court, rule 8.1115(a), prohibits courts and parties from citing or relying on opinions not certified for publication or ordered published, except as specified by rule 8.1115(b). This opinion has not been certified for publication or ordered published for purposes of rule 8.1115.

IN THE COURT OF APPEAL OF THE STATE OF CALIFORNIA

FIRST APPELLATE DISTRICT

DIVISION ONE

SCOTT NALICK, Plaintiff and Appellant, A166356 v. (San Francisco City & County SEAGATE TECHNOLOGY LLC, Super. Ct. No. CGC-15-547787) Defendant and Respondent.

Plaintiff Scott Nalick, on behalf of himself and others similarly situated, appeals from an order decertifying a class alleging violations of the Consumer Legal Remedies Act (Civ. Code, § 1750 et seq.; CLRA) and California’s unfair competition law (Bus. & Prof. Code, § 17200 et seq.; UCL) under an omissions-based theory of liability. We affirm. I. INTRODUCTION In June 2017, Nalick filed the operative second amended complaint (SAC) on behalf of a putative class of purchasers of defendant Seagate Technology LLP’s (Seagate) three-terabyte hard disk drives bearing model No. ST3000DM001 (the Drives). In November 2017, the trial court granted in part Nalick’s motion for class certification—certifying only as to CLRA and UCL claims based on Seagate’s alleged omissions of the Drives’ failure rates. Regarding predominance, the court ruled that Seagate’s knowledge of the

Drives’ annual failure rates (AFR), reasonable consumers’ expectations of the Drives’ failure rates, and the Drives’ actual AFRs were all susceptible to common proof. Despite lacking evidence of the Drives’ actual AFRs, the court reasoned Nalick “may be able to demonstrate that whatever the failure rate was—and by definition, there was some failure rate, there always is—it was sufficiently high that consumers should have been alerted to it.” After the ruling, the parties stipulated the class and subclass definitions. In April 2019, the trial court denied a motion by Seagate for summary adjudication, but later granted Seagate’s motion for reconsideration and reversed its previous decision. Nalick appealed, and this court reversed the order granting summary adjudication in a nonpublished opinion. (Nalick v. Seagate Technology LLC (Mar. 25, 2021, A158237) (Seagate I).) Our opinion addressed Seagate’s duty to disclose under the CLRA and UCL, and concluded the evidence was sufficient to show the materiality of an AFR as low as 1 percent as well as Seagate’s knowledge of the allegedly high AFR. On remand, Seagate moved to decertify the class, challenging predominance, typicality, and ascertainability. Nalick opposed the motion, arguing Seagate did not meet its burden to establish changed circumstances necessary for class decertification and, alternatively, common issues predominated the case and the class was still ascertainable. The trial court granted the decertification motion. The court determined the motion was procedurally proper because it was “predicated on different arguments and evidence than [Seagate’s] opposition to class certification” and because a federal court in a similarly situated case denied certification of a nationwide class. In light of Nalick’s claims “rely[ing] on AFR to establish the predicate fact that the Drives had a high failure rate,” the court found Seagate’s newly submitted evidence demonstrated “the

measure of AFR itself [was] subject to individualized inquiries.” After also excluding an expert declaration submitted in support of Nalick’s opposition briefing, the court concluded Nalick failed to meet his burden to show that common issues predominate, the class action was manageable, and the class was sufficiently ascertainable. Nalick appealed. We find no abuse of discretion and affirm. II. FACTUAL AND PROCEDURAL BACKGROUND Seagate manufactures the Drives, which are used in multiple different products and serve different needs. Seagate publicly represented that the Drives provide “ ‘trusted performance, reliability, simplicity and capacity’ ” to protect against data loss. Nalick purchased a product containing one of the Drives, which failed approximately one year later, causing Nalick to lose his data. This was, he claims, because the Drives “contain[ed] latent defects that cause the Drives to fail at an extraordinarily high rate.” In 2017, after overcoming Seagate’s demurrer, Nalick filed the operative SAC asserting Seagate’s alleged misrepresentations and omissions about the Drives’ failure rates violated the CLRA, UCL, California’s false advertising law (Bus. & Prof. Code, § 17500 et seq.), and the Song-Beverly Consumer Warranty Act (Civ. Code, § 1790 et seq.) He also filed a motion for class certification for the alleged violations of the Song-Beverly Act, UCL, and CLRA. A. Class Certification In his motion for class certification, Nalick characterized the “core issues of liability” as “whether the Drives were suitable for their ordinary purpose and whether Seagate’s conduct was deceptive.” To prove the Drives were unfit for their ordinary purpose, Nalick presented evidence of (1) a

“study of over 50,000 consumer-grade hard drives that showed that the Drives failed at a 78% rate,” and (2) a recall of “OEM” products that used the Drives due to a three-year cumulative return rate of around 5 percent.1 Nalick did not allege the exact failure rate of the Drives, but his expert, Thomas Coughlin, testified by declaration that Seagate “quantifies failure rates using a metric known as ‘Annual Failure Rate’ or ‘AFR.’ ” As both Coughlin and Seagate explained, AFR is an estimated percentage a product will fail due to a manufacturer issue within one year. AFR measures the statistical reliability of the Drives as manufactured, before Seagate distributes or sells them. Seagate’s opposition attacked Nalick’s alleged common proof—i.e., the study purporting a 78 percent failure rate and a recall of OEM products due to postsale returns. Specifically, Seagate argued there was no evidence of any alleged common “defect” in the Drives and “no [admissible] evidence that the proposed class experienced a failure rate similar to the one” claimed by the study or OEM recall. (Italics omitted.) In November 2017, the trial court partially granted Nalick’s motion for class certification to certain CLRA and UCL claims based on alleged omissions of the Drives’ failure rates. Its finding that common issues predominated was premised on the fact “there was some failure rate.” The court explained Nalick would need “to demonstrate that whatever the failure rate was and, by definition, there was some failure rate, there always is — it was sufficiently high that consumers should have been alerted to it.” The court denied the motion as to the other claims, in part because Nalick failed

to present substantial evidence of a defect or affirmative misrepresentations in Seagate’s advertising. B. This Court Reverses Summary Adjudication² Following the grant of class certification, Seagate moved for summary adjudication, which was denied in April 2019. But about two months after initially denying Seagate’s motion, the trial court reconsidered and reversed its prior opinion. The court held on reconsideration that Nalick’s evidence was not sufficient to show a dispute of material fact regarding what AFR would be material to a reasonable consumer. And the court found Seagate’s evidence failed to demonstrate knowledge of an AFR above 3 percent. On appeal, we reversed. (Seagate I, supra, A158237.) First, we held the trial court improperly interpreted Daugherty v. American Honda Motor Co., Inc. (2006) 144 Cal.App.4th 824 as a bar to Seagate’s liability for omissions after the expiration of the warranty period. We explained California law “sanctioned an ‘omission claim when: the plaintiff alleges that the omission was material; second, the plaintiff [alleges] the defect was central to the product’s function; and third, the plaintiff [alleges] one of the four LiMandri[³] factors.’ ” (Seagate I, supra, A158237.) Because the Drives’ alleged defect—i.e., their high rate of failure—related to the Drives’ central functionality, we determined Nalick’s claims could proceed “provided he can meet the LiMandri test and demonstrate materiality and exclusive knowledge by Seagate.” (Ibid.)

Regarding the LiMandri test, we held Nalick may demonstrate a duty to disclose arising from Seagate’s alleged partial representations about the Drives’ reliability. In short, the fact that the certified class claims were not based on a latent defect or affirmative misrepresentation did not “limit[] how Nalick may prove his omissions-based claims.” (Seagate I, supra, A158237.) Turning to materiality and knowledge, we concluded the record reflected a genuine issue of material fact regarding at what level AFRs become material and whether Seagate had exclusive knowledge of such AFRs. Central to that finding was our disagreement with the trial court’s interpretation of the declaration of Stefan Boedeker, which was submitted in support of Nalick’s opposition to Seagate’s motion for reconsideration of the ruling denying summary judgment. Boedeker offered an opinion quantifying the economic loss for an AFR as low as 3 percent. The trial court interpreted Boedeker’s conclusion to be only AFRs at or above 3 percent are material. But Boedeker also analyzed AFRs “ ‘between 1% and 10% through interpolation.’ ” (Seagate I, supra, A158237.) Indeed, “Boedeker expressly stated ‘the model can be further refined to enable the quantification of economic losses for different interim AFRs,’ including ‘the specification of economic losses for all incremental AFRs between 1% and 50%.’ ” (Ibid.) Thus, the “court exceeded its role by adopting its own interpretation of Boedeker’s conclusions.” (Ibid.) We held that “Boedeker’s declaration provides sufficient evidence of materiality for AFR’s above 1 percent to raise a genuine issue of material fact.” (Ibid.) Finally, we determined the trial court’s “initial order” had correctly determined Seagate’s own evidence in support of its motion for summary adjudication raised a factual dispute regarding whether Seagate had knowledge of Drives with AFRs over 1 percent. (Seagate I, supra, A158237.)

That evidence consisted of the declarations of David and Sek Nam “Allen” Ng, which discussed return rate information—not AFRs. Specifically, they “ ‘identif[ied] return rates [of Drives] that were in excess of 1% per year over the first two years, even, in Trane’s case, when “No Trouble Found” and “Could Not Duplicate” returns are excluded.’ ” (Seagate I, supra, A158237.) We held “ ‘a trier of fact presented with Seagate’s evidence [regarding return rates in excess of 1%] could conclude that at least some of the drive families at issue here had AFR’s in excess of 1% and that Seagate knew of those AFR’s as a result of return data.’ ” (Ibid.) Accordingly, the order granting summary adjudication was reversed. C. Class Decertification After the reversal of summary adjudication and remand, Seagate moved to decertify the class. Seagate raised several arguments in its decertification motion, three of which are relevant here: (1) “changed circumstances, theories of liability, and evidence warrant revisiting class certification”; (2) common questions do not predominate because Nalick could not provide common proof of the Drives’ AFRs; and (3) the certified class was not ascertainable. As a threshold matter, Seagate asserted changed circumstances justified its request for decertification. First, Seagate argued Nalick’s theory of liability shifted from claiming a latent defect to an obligation to disclose an AFR above 1 percent. Second, in support of this “new theory of liability,” Seagate contended Nalick changed his evidence regarding AFRs. Third, Seagate alerted the court to a parallel action in the Northern District of California, where the federal court denied class certification based on the same products and theories of liability. Finally, Seagate claimed new case

law required omissions-based claims to establish proof of class exposure to a partial representation to give rise to a duty to disclose. On the merits of decertification, Seagate asserted individual questions predominate because the AFRs at issue vary. Specifically, there are different AFRs for the different products that use the Drives. Because Seagate’s ongoing reliability testing of the Drives was done on a rolling basis and fixes were implemented throughout the manufacturing process, Seagate asserted the AFRs of a single product would vary across testing cycles. Relatedly, Seagate argued two AFRs may not be equivalent because changes in manufacturing output affect the testing sample sizes and thus distort the AFR results. As to the federal case, Seagate underscored the federal court’s conclusion that “evidence of AFR rates for the drives at issue here is not common to the class” because of “ ‘the undisputed variance of [the AFR] metric across the class period.’ ” (Boldface omitted.) Regarding ascertainability, Seagate argued the class was impermissibly overbroad. Specifically, Seagate noted the class definition “covers all consumers who purchased the [Drives] in California, regardless of the product, or the version of the product, over a period of over ten years, and regardless of whether the consumer actually saw or relied on any partial representation that gave rise to a duty to disclose a certain AFR rate” and regardless of whether they “experienced any drive failure or suffered any harm.” After separating out potential class members based on these differences, Seagate asserted there would be 72 subclasses. In opposition, Nalick countered that Seagate’s motion was untimely and not justified by “significant” changes in circumstances. Nalick stated his theory of liability had not changed, as evidenced by the fact the court certified a class based on omissions, not a latent defect theory. Further, the federal

case Seagate relied on was decided nearly four years earlier—in July 2018— and was not “new,” according to Nalick, because it merely applied settled law. Nalick also claimed Seagate “present[ed] no new evidence” in support of decertification. Regarding predominance, Nalick argued Seagate “speculates that differences among the Drives’ versions might create individual issues” without submitting newly discovered evidence demonstrating that the Drives’ AFRs, or Seagate’s knowledge of them, actually varied over time. And, even assuming there was actual AFR variance, Nalick quoted the trial court’s original certification order stating the “reasonable consumer” test presents common questions despite Seagate’s evidence that “global failure rate for the Drives” may vary based on “their use.” To the extent return rates are not an accurate measure of the Drives’ AFRs, Nalick argued “that too is a common question.” In support of his opposition briefing, Nalick submitted the declaration of John Levy, Ph.D. Relying on the return data presented during summary adjudication proceedings, Levy declared “ ‘failure rates for each of the three Drive versions fell within a narrow range,’ ” which he concluded presents common rather than individual issues. Citing Levy’s declaration, Nalick asserted, “Because the Drive only had three versions (Classic, BP, or BP2) and two uses (internal or external), a trial would involve, at most, the presentation of six failure rates.” Regarding ascertainability, Nalick argued the court should not revisit its prior decision because the parties stipulated the class definitions after class certification. He further contended the differences Seagate cited to defeat ascertainability were irrelevant. Specifically, he argued that Seagate’s representations were similar regardless of the product, that Levy concluded

the AFRs “ ‘fell within a narrow range’ ” regardless of which Drive version was used in each product, and that absent class members do not need to demonstrate reliance. Following a hearing, the trial court decertified the class. The court concluded Seagate’s motion was procedurally proper because it was “predicated on different arguments and evidence than its opposition to class certification, which did not focus on AFRs.”⁴ The court also found the parallel federal case to be “new” because it was rendered postcertification. Turning to the merits, the court held individual questions predominate. The court observed that Nalick’s claims “rely on AFR to establish the predicate fact that the Drives had a high failure rate.” In accord with Seagate I, the court stated the predicate fact—i.e., the alleged AFRs— “supports the element[] of materiality.” But Seagate’s evidence concerning the AFRs demonstrated “different versions of the [Drives], rolling reliability testing, and fixes during the manufacturing process render the measure of AFR itself subject to individualized inquiries.” Further, whether AFRs were affected by mishandling “raise[d] individual inquiries into specific distribution channels, product placement at retailers, and whether a class member mishandled a Drive.” And, because “AFR variances are not necessarily applicable only to one set of class members,” the court “ ‘decline[d] to find [differences in the Drives’ AFRs were] a merits issue rather than a class certification issue.’ ”

In reaching the conclusion that individual questions predominate—and relatedly that AFR variations present significant manageability concerns for trial—the court expressly cited and discussed the declarations of Greg Almgren, Karl Schweiss, and Harrie Netel.⁵ Almgren’s declaration described the reliability testing of the Drives and explained AFR may refer to different estimates during reliability testing. Further, Almgren stated the specifications for a Drive used in one product differed from the specifications used in other products. Schweiss’s declaration declared Seagate never specified AFRs for certain products and AFRs for different products were conditioned on different “power-on hours” per year depending on their anticipated uses. Netel’s declaration explained how measures of AFRs during production actually varied by day and week due to the cycle of reliability testing, and he stated “ ‘the individual AFR for the 3TB drive, [] was not always considered to be as accurate because variances in manufacturing output resulted in variances of sample sizes each week.’ ” Netel further stated AFR testing measured the Drives’ reliability “ ‘before the products reach[ed] the hands of retailers and eventually, consumers.’ ” The trial court granted Seagate’s motion to strike Levy’s declaration on the grounds that its conclusions were speculative and not supported by evidence. The court observed Levy “did not offer any direct evidence of the Drives’ AFRs” and instead relied on return data. Thus, as the court explained, a critical assumption underlying Levy’s conclusion was that the Drives’ AFRs were “at least equal to” products’ return rates.

Upon examining the return data on which Levy relied, the court concluded “that assumption [wa]s not supported.” The applicable return data came from the Trane declaration filed in support of Seagate’s motion for summary judgment. After adjusting for “ ‘no trouble found’ ” and “ ‘could not duplicate,’ ” Trane calculated a range of adjusted return rates for different versions of the Drives. Trane explained these adjusted return rates included returns due to mishandling. Because the data measured returns a year after sale—in contrast to AFR’s measure of estimated failure rates before distribution and sale—the court faulted Levy for failing to cite any studies or evidence establishing return rates correspond to “minimum” AFRs. Accordingly, the trial court determined Levy’s conclusion was unsound. The court therefore held that Nalick failed to meet his burden to show common proof of the AFRs. Independent of Nalick’s failure to establish common issues predominate, the court found Nalick did not meet his burden to show the existence of an ascertainable class. The court explained Nalick did not show how the class definition “could be narrowed to include only those class members who have viable claims.” In particular, the court highlighted that the class definition included consumers “whose drives had AFR rates of less than 1%, in addition to consumers whose mishandling or overuse of their drives caused their failure.” Thus, the trial court granted Seagate’s motion to decertify the class. Nalick appealed.

III. DISCUSSION A. Legal Standards A trial court that discovers a class action is no longer proper retains flexibility to decertify the class. (Weinstat v. Dentsply lnternat., Inc. (2010) 180 Cal.App.4th 1213, 1226 (Weinstat).) But decertification is appropriate “ ‘only where it is clear there exist changed circumstances making continued class action treatment improper.’ ” (Green v. Obledo (1981) 29 Cal.3d 126, 148.) “A motion for decertification is not an opportunity for a disgruntled class defendant to seek a do-over of its previously unsuccessful opposition to certification.” (Williams v. Superior Court (2013) 221 Cal.App.4th 1353, 1360 (Williams).) We review a decertification order for an abuse of discretion. (Moen v. Regents of the University of California (2018) 25 Cal.App.5th 845, 853.) Unlike ordinary appellate review, however, “ ‘when denying class certification, the trial court must state its reasons, and we must review those reasons for correctness.’ ” (Peviani v. Arbors at California Oaks Property Owner, LLC (2021) 62 Cal.App.5th 874, 887.) Accordingly, decertification resting on an invalid reason, such as improper legal criteria or an incorrect assumption, is an abuse of discretion. (Williams, supra, 221 Cal.App.4th at p. 1361.) Nevertheless, appellate review is “narrowly circumscribed,” reversing only for manifest abuse of discretion “ ‘ “[b]ecause trial courts are ideally situated to evaluate the efficiencies and practicalities of permitting group action.” ’ ” (Brinker Restaurant Corp. v. Superior Court (2012) 53 Cal.4th 1004, 1022 (Brinker).) The trial court’s predominance finding is reviewed for substantial evidence. (Lockheed Martin Corp. v. Superior Court (2003) 29 Cal.4th 1096,

1106 (Lockheed).) The issue of predominance turns on “ ‘whether the theory of recovery advanced by the proponents of certification is, as an analytical matter, likely to prove amenable to class treatment.’ ” (Brinker, supra, 53 Cal.4th at p. 1021.) “ ‘Reviewing courts consistently look to the allegations of the complaint and the declarations of attorneys representing the plaintiff class to resolve this question.’ ” (Sav-On Drug Stores, Inc. v. Superior Court (2004) 34 Cal.4th 319, 327 (Sav-On).) On review for substantial evidence, appellate courts view the record in the light most favorable to the respondent and resolve all evidentiary conflicts and indulge all reasonable inferences in favor of the judgment. (Williams-Sonoma Song-Beverly Act Cases (2019) 40 Cal.App.5th 647, 650.) Evidence Code section 802 permits the court in its discretion to exclude expert testimony on the grounds that there is no reasonable basis for the opinion. (See Sargon Enterprises, Inc. v. University of Southern California (2012) 55 Cal.4th 747, 771 (Sargon).) We review the exclusion of expert testimony for abuse of discretion. (Id. at p. 773.) B. New Evidence Justified Revisiting Certification While a class “can be decertified at any time, even during trial, should it later appear individual issues dominate the case” (Macmanus v. A. E. Realty Partners (1987) 195 Cal.App.3d 1106, 1117), decertification “ ‘typically occur[s] in response to a significant change in circumstances’ ” (Williams, supra, 221 Cal.App.4th at p. 1360). And the party seeking decertification “generally has the burden to show that certification is no longer warranted.” (Kight v. CashCall, Inc. (2014) 231 Cal.App.4th 112, 126.) Thus, to allow courts the flexibility to decertify in the face of changed circumstances while curtailing abuse by class defendants, a class defendant is required to show new evidence or new law demonstrating such changed circumstances. (Weinstat, supra, 180 Cal.App.4th at p. 1226.)

The trial court articulated Seagate’s motion was proper because it was “predicated on different arguments and evidence than its opposition to class certification, which did not focus on AFRs.” Specifically, Seagate proffered the declarations of Almgren, Schweiss, and Netel, which all address how AFRs are calculated. Nalick counters that none of this “new” evidence was “newly discovered” after the initial class certification order. (Cf. Weinstat, supra, 180 Cal.App.4th at p. 1225 [decertification improper where evidence is “newly packaged, but not newly discovered”].) Nalick’s assertion is misleading and ignores the new evidence that he produced after class certification in opposition to Seagate’s motion for summary judgment. As the decertification order explained, the initial class certification proceedings did not focus on AFRs.⁶ Instead, Nalick’s alleged common proof of the Drives’ “abnormally high failure rate” was a study purporting the Drives’ failure rate was 78% and a recall of OEM products due to postsale returns. Seagate’s opposition was responsive—and naturally limited in its response—to the asserted common proof.⁷ Thus, neither party’s initial class certification briefing addressed how AFRs are measured or whether calculating AFRs for products across the class period would require individualized inquiries. Despite noting the absence of evidence regarding the failure rate, the trial court determined the “failure rate of a consumer hard drive” existed “by definition” and thus was susceptible to common proof. It granted class

certification on certain CLRA and UCL claims based on alleged omissions because Nalick “may be able to demonstrate that whatever the failure rate was—and, by definition, there was some failure rate, there always is—it was sufficiently high that consumers should have been alerted to it.” In other words, to succeed at trial, Nalick would need to present evidence of the actual AFRs. Accordingly, in opposition to Seagate’s motion for summary adjudication, Nalick submitted new evidence regarding the Drives’ AFRs. At this stage, such a showing was necessary because his claims “rely on AFR to establish the predicate fact that the Drives had a high failure rate” across the class. And while Nalick’s theory of liability did not change, his evidence and arguments in support of that theory did. Indeed, it was not until after class certification was granted, during summary adjudication proceedings, when Nalick identified evidence that a particular failure rate—1 percent—was material to a reasonable consumer. Tellingly, in its initial order denying Seagate’s motion for summary adjudication, the trial court noted decertification may be appropriate if evidence regarding the “exact failure rate” showed AFRs varied. This was prescient. After this court reversed the trial court’s subsequent order granting summary adjudication, Seagate moved to decertify the class and submitted additional evidence in support of its motion. Seagate’s newly submitted evidence showed—for the first time—the variability of Drives’ AFRs across both products and the class period. This was “new” evidence regarding common proof of the AFRs, not previously considered by the court, and arose in response to Nalick’s postcertification evidence and arguments regarding the materiality of an AFR at or above 1 percent.

Nalick’s cited authorities do not stand for the proposition that changed circumstances may only be demonstrated by newly arisen facts. In Williams, the decertification order relied on evidence from a declaration “filed in support of petitioner’s motion for certification.” (Williams, supra, 221 Cal.App.4th at p. 1368.) Because the same evidence had been presented in support of the original motion for class certification, it was demonstrably not new. Weinstat is similarly inapposite. There, the trial court decertified two claims based on allegedly new law, not new evidence. (Weinstat, supra, 180 Cal.App.4th at pp. 1217–1218.)8 The court decertified a UCL claim based on an incorrect interpretation of new case law, then decertified a breach of warranty claim on the same basis despite the changed “circumstance only pertain[ing] to the UCL cause of action.” (Weinstat, at pp. 1224, 1226.) The appellate court criticized the rulings as “in effect reassess[ing] the matter under existing law, coupled with newly packaged, but not newly discovered, evidence.” (Id. at p. 1225.) Weinstat did not establish a rule restricting new evidence to newly arisen facts. Nalick also cites People v. Anderson (2008) 169 Cal.App.4th 321, which accords with our analysis. There, the court denied a motion to suppress and that ruling was upheld in the first appeal, which reversed the defendant’s conviction on other grounds. (Id. at p. 325.) On remand, the defendant renewed the motion, and the court conducted a new hearing and considered new evidence without reconsidering evidence submitted at the first hearing. (Id. at p. 329.) On appeal the second time, the appellate court condoned this approach and articulated a framework for determining when a court must

8 Indeed, the defendant in Weinstat did not contend there was new

evidence, instead expressly arguing “there is no requirement of changed circumstances or new evidence” to revisit certification. (Weinstat, supra, 180 Cal.App.4th at p. 1226.)

redetermine evidentiary matters on remand where the underlying issue was “actually addressed by the [first] appellate court.” (Id. at pp. 332–333.) First, the court must “determine whether the renewed motion involves ‘new or different’ evidence. If not, the trial court must follow the ruling of the appellate court; if so, the trial court must redetermine the suppression motion.” (Id. at p. 333.) The Anderson court explained “new or different” evidence in such a context included evidence presented on remand that “raise[s] material factual or legal issues not considered by the original trial judge.” (Id. at pp. 333–334, italics omitted.) Under this standard, Nalick and Seagate presented new evidence—i.e., evidence not considered by the court at the initial class certification proceedings. Further, the reversal of the trial court’s grant of summary adjudication in Seagate I did not preclude decertification because the issue of common proof was not “actually addressed.” Under the law of the case doctrine, “[t]he decision of an appellate court, stating a rule of law necessary to the decision of the case, conclusively establishes that rule and makes it determinative of the rights of the same parties in any subsequent retrial or appeal in the same case.” (Morohoshi v. Pacific Home (2004) 34 Cal.4th 482, 491.) In Seagate I, we found the Boedeker declaration was “sufficient evidence of materiality for AFR’s above 1 percent to raise a genuine issue of material fact.” (Seagate I, supra, A158237.) We also found—based on the Trane and Ng declarations submitted in support of Seagate’s motion for summary adjudication—a trier of fact “could conclude that at least some of the drive families at issue here had AFR’s in excess of 1% and that Seagate knew of those AFR’s as a result of return data.” (Ibid., italics added.) We did not, however, hold that the Boedeker declaration or return rates to which Trane and Ng testified were common proof of a 1 percent AFR across the different Drive versions or across

the entire class. We did not expressly or implicitly decide whether common issues predominate or whether the factual inquiries into AFRs were manageable. (See Searle v. Allstate Life Ins. Co. (1985) 38 Cal.3d 425, 435 [explaining justification “for the [law of the case doctrine] is inoperative when the court hearing the subsequent appeal determines that there should be a reversal on a ground that was not considered on the prior appeal”].) Nalick’s authorities are unavailing for the same reason. For example, the cross-defendant in Nissan Motor Acceptance Cases (2021) 63 Cal.App.5th 793 “contend[ed] no evidence showed” justifiable reliance, but the appellate court rejected the same argument in an earlier appeal. (Id. at pp. 823, 824.) Because the cross-complainant presented the same evidence regarding the same issue—i.e., sufficiency of the evidence of justifiable reliance—the law of the case doctrine precluded a different result. (Id. at p. 824.) Moreover, the court expressly stated the law of the case doctrine was inapplicable to a separate challenge by the cross-defendant. (See id. at pp. 824–825 [reviewing whether there was substantial evidence of negligent representation because the first appeal “did not expressly or implicitly decide” the issue].)⁹ The same logic holds here. Because common proof was not expressly or implicitly decided in Seagate I, the court’s subsequent ruling that individualized issues predominate the inquiry into AFRs did not contravene the law of the case. In sum, the evidence submitted postcertification regarding the Drives’ actual AFRs demonstrated changed circumstances. Namely, that the predicate fact of Nalick’s claims was susceptible to individualized inquiries

and manageability concerns. The court therefore did not abuse its discretion by considering Seagate’s request for decertification.¹⁰ C. The Court Did Not Impermissibly Resolve Merits Questions A trial court’s inquiry into whether certification is proper is essentially procedural, and the “resolution of disputes over the merits of a case generally must be postponed until after class certification has been decided.” (Brinker, supra, 53 Cal.4th at p. 1023.) So the court “ordinarily must assume the claims have merit.” (Williams, supra, 221 Cal.App.4th at p. 1360.) But “no such assumption applies to the issues in dispute on [a certification] motion,” including predominance. (Apple Inc. v. Superior Court (2018) 19 Cal.App.5th 1101, 1119, fn. 3.) Issues related to predominance often overlap issues affecting the merits of a case. (Brinker, supra, 53 Cal.4th at pp. 1023–1024.) To decide whether individual or common issues predominate, the “court must examine the plaintiff’s theory of recovery, assess the nature of the legal and factual disputes likely to be presented,” and, if necessary, resolve disputed threshold legal or factual questions. (Id. at p. 1025; see also Sav-On, supra, 34 Cal.4th at p. 334 [“Predominance is a comparative concept”].) “ ‘Critically, if the parties’ evidence is conflicting on the issue of whether common or individual questions predominate (as it often is . . . ), the trial court is permitted to credit one party’s evidence over the other’s in determining whether the

requirements for class certification have been met.’ ” (Mies v. Sephora U.S.A., Inc. (2015) 234 Cal.App.4th 967, 981 (Mies).) Nalick argues the court impermissibly resolved merits questions by “reject[ing] a fundamental leg of Nalick’s theory of liability—that Seagate’s own internal return data provided critical, classwide evidence of the Drives’ common failure rates.” Not so. The predicate factual issue raised by Nalick’s theory of recovery was that the Drives’ AFR exceeds 1 percent. If Drives’ AFRs vary—including between different Drive versions and products and due to rolling fixes and testing cycles—then liability is susceptible to individualized inquiries. Therefore, the variability of AFRs is an issue of predominance that may be considered. As discussed below, Seagate presented substantial evidence showing the AFRs vary due to several factors, which presents “serious manageability concerns for trial.” The court then assessed whether Nalick met his burden of showing common issues predominate or that it “ ‘will be feasible to try the case as a class action.’ ” It found he did not.¹¹ Undertaking this balancing analysis was appropriate and did not improperly resolve merits questions. (See Mies, supra, 234 Cal.App.4th at p. 983.) Accordingly, we review the court’s finding that individualized issues predominate for substantial evidence. (Lockheed, supra, 29 Cal.4th at p. 1106.) D. Substantial Evidence Supports Predominance Finding 1. Seagate Evidence

Predominance turns on “whether the theory of recovery advanced by the proponents of certification is, as an analytical matter, likely to prove amenable to class treatment.” (Sav-On, supra, 34 Cal.4th at p. 327.) Accordingly, the trial court correctly observed Nalick’s claims “rely on AFR to establish the predicate fact that the Drives had a high failure rate.” The court then evaluated the evidence submitted in support of Seagate’s decertification motion—primarily, the declarations of Almgren and Netel. The court concluded “the different versions of the Grenada drives, rolling reliability testing, and fixes during the manufacturing process render the measure of AFR itself subject to individual inquiries.” Viewing the record in the light most favorable to Seagate and indulging all reasonable inferences in favor of the judgment, we affirm. (See Williams-Sonoma Song-Beverly Act Cases, supra, 40 Cal.App.5th at p. 650.) This conclusion was grounded in substantial evidence. The court’s decertification order relied on Almgren’s declaration describing “significant changes” between different versions of the Drives, which required different reliability demonstration testing, which in turn affected AFRs. Reliability demonstration testing “included calculating the mean time between/before failure (‘MTBF’) as well as AFRs.” But “AFR” may refer to different estimates—“ ‘raw’ AFR” reflected estimates before fixes were implemented, while “ ‘demonstrated reduced AFR’ ” reflected AFR after fixes were implemented. Moreover, the AFRs of different versions of the Drives were not necessarily equivalent because the projected AFR of any given version was impacted by the estimated power-on hours of that version’s intended use. For example, an internal hard drive was expected to be used approximately 40 hours a week, while an external hard drive was expected to be used less than half as much. Indeed, some versions of the Drives did not have an AFR

target at all. Such Drives instead had to meet an MTBF specification of 100,000 hours, which corresponds to an AFR around 1 percent. However, that equivalence depends on an assumption that the product will be used for less than 100 power-on hours. Further evidence the court cited was Netel’s declaration regarding Seagate’s ongoing reliability testing, which occurred “post launch of a product” and assessed the Drives “on a rolling basis with a new sample pulled for testing each week.” “Typical testing [of hard drives sold as internal desktop drives] takes six weeks, so daily or weekly results are assessed on the entire group of drives, which is a rolling group of drives that has been in testing for various lengths of time.” Ongoing reliability testing produced two different calculations of “AFR”: (1) “the average AFR” across hard drives; and, (2) “the individual AFR for the 3 TB drive, which was not always considered to be as accurate because variances in manufacturing output resulted in variances of sample sizes each week.” Seagate also monitored a third metric—“ ‘rolling AFR’ ”—an average of AFRs for a six-week period. The court also credited Netel’s declaration that Seagate implemented engineering change requests to improve the Drives throughout manufacturing: “ ‘Although often minor, it cannot be stated that a drive from the first week of manufacture is the exact same as a drive from the final week of manufacture.’ ” Moreover, if an ongoing reliability testing result—i.e., an MTBF or AFR calculation—exceeded a certain threshold, then Seagate would issue a “ship hold” and implement a fix. Thus, “ ‘the reported AFR in documents could be higher than the actual AFR of drives shipped to consumers for two reasons: (1) the reported AFR is the AFR that triggered the ship hold . . . ; (2) even after a fix was implemented, it could take time for

the fixed drives to make their way into testing and bring down the “raw” AFR to match the AFR of drives being shipped to consumers.’ ” This evidence adequately substantiated Seagate’s contention that individual issues predominate and present manageability issues.¹² The court then assessed Nalick’s countervailing evidence to determine if he met his burden of showing that common issues predominate or that it “will be feasible to try the case as a class action.” (Duran v. U.S. Bank National Assn. (2014) 59 Cal.4th 1, 27.) After evaluating and excluding Nalick’s expert testimony, it determined he did not. (See Sav-On, supra, 34 Cal.4th at p. 334 [“Predominance is a comparative concept”].) 2. Court Properly Excluded Nalick’s Predominance Evidence Trial courts have discretion to exclude expert testimony on the grounds that there is no reasonable basis for the opinion. (See Sargon, supra, 55 Cal.4th at p. 771.) A court’s inquiry into “whether expert opinion is founded on sound logic is not a decision on its persuasiveness.” (Id. at p. 772.) The court merely determines “whether the matter relied on can provide a reasonable basis for the opinion or whether that opinion is based on a leap of logic or conjecture.” (Ibid.) Thus, the issue before us is whether the court abused its discretion in excluding Levy’s testimony, not whether there is common proof of the AFRs. (Id. at p. 773.) But how AFRs are calculated is

relevant to help define the type of matter on which an expert may reasonably rely in reaching such a conclusion. (See id. at pp. 775–776.) Nalick’s asserted common proof of the Drives’ AFRs was limited to the Levy declaration, which relied on the Trane and Ng declarations regarding return rates. As the court’s decertification order describes, Levy opined the Drives’ return rates “ ‘establish a lower bound for the actual failure rates’ ”— i.e., a minimum AFR—and “ ‘the failure rates for each of the three Drive versions fell within a narrow range.’ ” The court was not persuaded. It concluded Levy’s opinion was speculative and unreliable because he “conflate[d] the two measures,” return rates and AFRs, “without any showing as to the relationship between them.” This conclusion does not contravene the law of the case. As discussed above, in Seagate I we determined a trier of fact presented with the same return rate data could conclude AFRs of “at least some of the drive families” exceeded 1 percent. (Seagate I, supra, A158237, italics added.) We did not rule such return data was common proof of the AFRs of all versions of the Drives across the class period. Nor did we determine such return data was a reasonable basis for establishing the Drives’ “minimum AFRs.” As the trial court explained, “[t]hat Seagate may have been put on notice by high return rates of the possibility of higher than usual AFRs does not establish that the two are interchangeable.” In other words, while sufficient to create a triable issue of material fact as to the Drives’ AFRs and Seagate’s knowledge of high failure rates, it was proper for the court to analyze whether the return data was a reasonable basis for Levy to opine “ ‘the failure rates for each of the three Drive versions fell within a narrow range’ ” and all correlate with the “minimum AFR.”

Levy asserted the return rate data could be used to establish the “minimum AFRs” because “ ‘[c]ommon sense suggests that many consumers will not return a failing device.’ ” But the court observed several flaws in using the return data as common proof of AFRs across the class period. To start, Nalick argued at summary judgment that Trane’s declaration “ ‘does not dispositively resolve the question of what the Drives’ AFRs actually were.’ ” And the court observed Ng’s declaration “relate[d] to OEM products that are outside the class definition in this case,” which Nalick acknowledges. Further, while Levy admitted the return rates were measured 12 and 24 months after sale and AFR measures the projected failure rate before sale, Levy did “not cite any studies or other evidence showing a correlation between the two or supporting his critical assumption that there is such a correlation.” Further undermining any correlation between return rates and AFRs— and showcasing the speculative nature of his conclusions—was Levy’s suggestion that Seagate’s AFR metrics may be artificially low. Levy surmised “ ‘the fact that the manufacturer . . . could not confirm that a hard drive failed or did not reproduce the failure in its testing does not necessarily mean that the Drive did not fail.’ ” The court was unimpressed that “without evidence” Levy “assume[d]” the “Drives’ failure rate must have been higher than [Seagate’s] own testing established.” The court also criticized Levy’s declaration for inadequately addressing how return data was common proof of different Drive versions’ AFRs. For example, Trane declared the return data encompassed mishandled products—an issue susceptible to individualized inquiries. Also, Levy opined the “minimum AFR”—i.e., the adjusted return rates—for three of the Drive versions ranged from 0.92 percent to 1.99 percent. In other words, at least

one version did not have an adjusted return rate above 1 percent. But Levy stated—without any supporting evidence—that this version “likely” had an AFR exceeding 1 percent because “ ‘some consumers may not return failing drives.’ ” Moreover, this range of adjusted return rates only reflected “return data for [Seagate’s] internal Drives.” And despite admitting he “did not break out the return rates for [Seagate’s] external products by Drive version,” Levy stated “without explanation” that “those drives, too, had a minimum AFR of 2.22%.” The trial court did not abuse its discretion by excluding Levy’s testimony. It properly found his conclusions were too speculative to show common proof of AFRs. Accordingly, the court held Nalick failed to meet his burden to show that common issues predominate over individualized issues, or that it “will be feasible to try the case as a class action.” (Duran v. U.S. Bank National Assn., supra, 59 Cal.4th at p. 27.) Finding substantial evidence in support of this holding, we affirm. 3. Ascertainability Nalick also argues the trial court erred in holding the class was not ascertainable by applying improper criteria and assumptions. We need not reach this issue, because we have determined the court did not abuse its discretion in finding common issues no longer predominated, warranting decertification of the class. (See Kight v. CashCall, Inc., supra, 231 Cal.App.4th at p. 128 [predominance sufficient ground for decertification].)

IV. DISPOSITION The order is affirmed. Seagate is entitled to costs on appeal. (Cal. Rules of Court, rule 8.278(a)(2).)

CASTRO, J.* WE CONCUR:

HUMES, P. J.

LANGHORNE WILSON, J.

A166356 Nalick v. Seagate Technology LLC

• • Judge of the Alameda County Superior Court, assigned by the Chief Justice pursuant to article VI, section 6 of the California Constitution.

Footnotes

[1] OEM products are ones in which the Drives are only a single component, such as a hard drive installed inside a desktop computer. OEM products were not included in the class definition.

[2] We incorporate by reference the procedural and factual background detailed in our previous nonpublished opinion, Seagate I, supra, A158237. But because the parties dispute its import, we summarize our prior opinion and then describe the subsequent procedural history.

[3] LiMandri v. Judkins (1997) 52 Cal.App.4th 326 (LiMandri).

[4] Later in its order, however, the court rejected the notion that Nalick’s theory of liability had changed. The court observed Nalick’s theory of liability was based on omissions of the Drives’ alleged high failure rates, and the fact that Nalick now identified AFRs above 1 percent as the failure rate a reasonable consumer would find material did not change his theory of liability.

[5] These declarations were submitted in support of Seagate’s March 2022 “Motion to Decertify the Class, or in the Alternative, to Modify the Class Definition.” Almgren had also submitted a separate declaration in June 2017 in support of Seagate’s opposition to Nalick’s motion for class certification.

[6] The only reference to the Drives’ AFRs in Nalick’s motion was an assertion that Seagate “published and disseminated detailed reliability and data specifications for the Drives, claiming that the ST3000DM001’s annualized failure rate (‘AFR’) was less than 1%.”

[7] Seagate did argue that evidence of misuse by certain class members raised individual issues for its affirmative defense.

[9] McCoy v. Hearst Corp. (1991) 227 Cal.App.3d 1657, 1661–1662, and Gibson v. State (1962) 208 Cal.App.2d 458, 461–462 are likewise inapposite. Both involved retrials after appellate rulings on the sufficiency of the evidence introduced at trial. Here, the issue is not the sufficiency of the evidence; it is whether there is common proof of the AFRs.

[10] Nalick also argues the court was wrong in finding a similar federal action was “new law” merely because it was rendered post-certification. We assume without deciding that Nalick is correct. (See Ovitz v. Schulman (2005) 133 Cal.App.4th 830, 848.) Decertification resting on an invalid reason is an abuse of discretion. (Williams, supra, 221 Cal.App.4th at p. 1361.) But, because the court expressly stated new evidence was an independent—and the primary—premise for considering decertification, this error is insufficient to reverse.

[11] As explained above, while the law of the case holds the return data is sufficient for a trier of fact to conclude “at least some of the drive families” had AFRs greater than 1 percent, it does not hold the return data is common proof of the AFRs class-wide. (See Seagate I, supra, A158237.)

[12] As the trial court recognized, the federal court reached the same conclusion. It found the differences among the Drives, and the different reliability demonstration testing standards applied to the internal and external drives, weighed against commonality. (See In re Seagate Technology LLC Litigation (2018) 326 F.R.D. 223, 244.) This reinforces our view that substantial evidence supports the court’s determination. (See also Brinker, supra, 53 Cal.4th at p. 1022 [trial courts are ideally situated to evaluate the efficiencies and practicalities of permitting group action].)