Skip to main content


Reset A A Font size: Print

United States Court of Appeals, Ninth Circuit.

IN RE: APPLE INC. DEVICE PERFORMANCE LITIGATION, Named Plaintiffs and Settlement Class Members, Plaintiff-Appellee, v.

Sarah Feldman; Hondo Jan, Objectors-Appellants, v. Apple Inc., Defendant-Appellee.

IN RE: Apple Inc. Device Performance Litigation, Named Plaintiffs and Settlement Class Members, Plaintiff-Appellee, v.

Best Companies, Inc., Objector-Appellant, v. Apple Inc., Defendant-Appellee.

IN RE: Apple Inc. Device Performance Litigation, Named Plaintiffs and Settlement Class Members, Plaintiff-Appellee, v.

Deborah Pantoni, Objector-Appellant, v. Apple Inc., Defendant-Appellee.

IN RE: Apple Inc. Device Performance Litigation, Named Plaintiffs and Settlement Class Members, Plaintiff-Appellee, v.

Anna St. John, Objector-Appellant, v. Apple Inc., Defendant-Appellee.

No. 21-15758, No. 21-15761, No. 21-15762, No. 21-15763

Decided: September 28, 2022

Before: Jacqueline H. Nguyen, John B. Owens, and Ryan D. Nelson, Circuit Judges. Kendrick Jan (argued), Kendrick Jan APC, San Diego, California, for Objector-Appellants Sarah Feldman and Hondo Jan. Scott A. Kamber (argued), Kamber Law LLC, Denver, Colorado, for Objector-Appellant Best Companies, Inc. John J. Pentz (argued), Law Offices of John J. Pentz, Sudbury, Massachusetts; Jan L. Westfall, San Diego, California; for Objector-Appellant Deborah Pantoni. Theodore H. Frank (argued) and Anna St. John, Hamilton Lincoln Law Institute, Center for Class Action Fairness, Washington, D.C., for Objector-Appellant Anna St. John. Mark C. Molumphy (argued), Joseph W. Cotchett, and Elle D. Lewis, Cotchett Pitre & McCarthy LLP, Burlingame, California; Laurence D. King, Kathleen A. Herkenhoff, and Matthew B. George, Kaplan Fox & Kilsheimer LLP, Oakland, California; Frederic S. Fox, Donald R. Hall, and Melinda C. Campbell, Kaplan Fox & Kilsheimer LLP, New York, New York; for Plaintiff-Appellee. Christopher Chorba (argued), Theodore J. Boutrous Jr., and Wesley Sze, Gibson Dunn & Crutcher LLP, Los Angeles, California; Kory Hines and Catherine McCaffrey, Gibson Dunn & Crutcher LLP, New York, New York; for Defendant-Appellee. Brett R. Nolan, Office of the Attorney General, Frankfort, Kentucky; Steve Marshall, Attorney General, Office of the Attorney General, Montgomery, Alabama; Mark Brnovich, Attorney General; Office of the Attorney General, Phoenix, Arizona; Leslie Rutledge, Attorney General, Office of the Attorney General, Little Rock, Arkansas; Ashley Moody, Attorney General, Office of the Attorney General, Tallahassee, Florida; Jeff Landry, Attorney General, Office of the Attorney General, Baton Rouge, Louisiana; Keith Ellison, Attorney General Office of the Attorney General, St. Paul, Minnesota; Douglas J. Peterson, Attorney General, Office of the Attorney General, Lincoln, Nebraska; Aaron D. Ford, Attorney General, Office of the Attorney General, Carson City, Nevada; Wayne Stenehjem, Attorney General, Office of the Attorney General, Bismarck, North Dakota; Dave Yost, Attorney General, Office of the Attorney General, Columbus, Ohio; Ken Paxton, Attorney General, Office of the Attorney General, Austin, Texas; Sean D. Reyes, Attorney General, Office of the Attorney General, Salt Lake City, Utah; for Amici Curiae Kentucky, Alabama, Arizona, Arkansas, Florida, Louisiana, Minnesota, Nebraska, Nevada, North Dakota, Ohio, Texas, and Utah. Shiyang Huang, Topeka, Kansas, as pro se Amicus Curiae.


In this multidistrict litigation against Apple Inc., the district court approved a $310 million class action settlement resolving allegations that Apple secretly throttled the system performance of certain model iPhones to mask battery defects. The court approved $80.6 million, representing 26% of the recovery, in fees to class counsel and also approved service awards to the named plaintiffs. Several class members who objected to these decisions now appeal.

The district court actively managed this difficult litigation, which involved the consolidation of dozens of federal lawsuits. Once the settlement was achieved following motions practice, discovery, and months of negotiations with the assistance of a mediator, the settlement administrator sent over 90 million class notices via email and over 5 million notices by postcard. About 99% of persons associated with potentially eligible devices received notice of the settlement. The settlement also received substantial press and social media coverage. We find that class members—including nonnatural persons—received adequate notice of the settlement. Any suggestion to the contrary is unsupported by the record.

The district court properly resolved most of the objections at issue on appeal. However, in finding the settlement fair, reasonable, and adequate, the district court committed legal error. As we have repeatedly admonished, settlement prior to class certification requires extra scrutiny. While we commend the district court's thoughtful and thorough analysis, which suggests that the court took great care in considering the terms of the settlement, its written order explicitly states that the court applied a presumption that the settlement was fair and reasonable. Because the district court cited the wrong legal standard, we vacate and remand for it to reconsider settlement approval under the correct standard. See Roes, 1–2 v. SFBSC Mgmt., LLC, 944 F.3d 1035, 1049 (9th Cir. 2019) (holding that the district court committed reversible error in approving a settlement negotiated prior to class certification by starting its analysis with a presumption of fairness and reasonableness).

I. Background

In January and December 2017, Apple released updates for its iPhones' system software (“iOS”) that under certain conditions slowed the performance of certain phones. At the time of the iOS releases, Apple acknowledged only that the updates provided “improvements” and fixed “bugs.” On December 20, 2017, after independent researchers published findings that the iOS updates degraded system performance, Apple publicly acknowledged as much.

Apple explained that as iPhone batteries age, they become less capable of supplying a phone's peak demands for electric current, 1 and that the iOS updates smoothed out a phone's electrical demands when necessary to avoid unexpected shutdowns. The following week, Apple disclosed that the smoothing feature affected the “the maximum performance of some system components.” However, Apple insisted that “[t]he level of perceived change depends on how much power management is required for a particular device” and that “[i]n some cases, a user may not notice any differences in daily device performance.”

Following Apple's disclosures, consumers around the country filed class action lawsuits concerning the unexpected shutdowns and iOS updates. The Judicial Panel on Multidistrict Litigation consolidated the 67 federal actions in the Northern District of California. The Judicial Council of California coordinated the four state court proceedings into a single action in the San Francisco Superior Court (the “JCCP action”), which proceeded parallel to the federal litigation.

In May 2018, the district court consolidated the individual federal cases, see Fed. R. Civ. P. 42(a)(2), and selected one of three competing proposals for the litigation's leadership structure. The court established a protocol for attorney work and expenses that required any compensable activity to be reasonable, non-duplicative, beneficial to the prosecution of the multidistrict litigation, and authorized by one of three attorneys managing the litigation. The court required plaintiffs' counsel to maintain contemporaneous records and provide quarterly reports for in camera review. And the court appointed a special master to oversee the discovery process who resolved several issues during the next two years.

In the consolidated amended complaint, a total of 122 named plaintiffs from every state and several territories and foreign countries alleged 76 claims against Apple. Plaintiffs asserted various fraud-based theories, breach of contract, trespass to chattels, and violation of the federal Computer Fraud and Abuse Act, 18 U.S.C. § 1030, California's Data Access and Fraud Act, Cal. Penal Code § 502, and several states' unfair competition and consumer protection statutes.

The district court eliminated many of the claims at issue in its rulings on Apple's two motions to dismiss. The court dismissed plaintiffs' claims regarding alleged battery defects but concluded that plaintiffs had viable claims relating to the iOS updates. In particular, the court concluded that plaintiffs could proceed on a “computer intrusion” theory for trespass to chattels and under the California and federal computer fraud statutes.

In February 2020, the parties reached a settlement. The settlement agreement resolved the claims of “all former or current U.S. owners” of certain iPhone models that ran specified versions of iOS by the time Apple first publicly disclosed that the iOS updates slowed phone performance under certain conditions. 2 The agreement resolved not only the federal multidistrict litigation, but also the JCCP action in California.

Apple agreed to pay $25 per eligible iPhone to settlement class members with approved claims, subject to two limitations. First, Apple agreed to pay the settlement class a minimum of $310 million and a maximum of $500 million. This meant that if the number of eligible iPhones with approved claims was less than about 12 million or more than 20 million, then the payment per device would be proportionately more or less than $25. Second, Apple's payment to settlement class members was subject to a deduction for any court-ordered attorney's fees, expenses, and service awards to the named plaintiffs.

To receive a cash payment, settlement class members had to certify under penalty of perjury that “they experienced diminished performance on [an] eligible device when running [the applicable iOS] before December 21, 2017.” Class members could also opt out of the settlement.

The district court granted preliminary approval to the settlement, provisionally certified the nationwide settlement class, and directed the parties to notify class members. Apple provided the settlement administrator with the names, contact information, and serial numbers associated with each potentially eligible device—i.e., each device covered by the settlement that had downloaded the applicable iOS during the relevant time period. 3 The settlement administrator sent 90,119,272 class notices via email and an additional 5,617,563 notices by postcard. In all, approximately 99% of the persons associated with potentially eligible devices received notice of the settlement.

The settlement administrator received 3,284,985 claims by the submission deadline and, as of January 2021, had approved 69% of them. 4 These response and claim approval rates meant that claimants could receive, on average, at least $128 less any court-ordered deduction for attorney's fees, costs, and incentive awards. 5 In addition, 622 persons timely opted out of the settlement.

Class counsel moved for final approval of the settlement and, separately, for attorney's fees, expenses, and service awards for the named plaintiffs. At issue here, counsel requested attorney's fees of $87.73 million and service awards of either $3,500 or $1,500 for each of the named plaintiffs, the larger amount for the nine named plaintiffs who were deposed.

The district court received various objections to the two motions from 144 class members. In addition, Apple opposed the requested amount of attorney's fees as excessive, and the United States as well as several states' attorneys general submitted briefs raising similar concerns. At two hearings lasting a total of eight hours, the district court heard from the parties and several objectors about the fairness of the proposed settlement and the requested compensation for class counsel and the named plaintiffs.

Ultimately, the district court granted final approval to the settlement and granted in part class counsel's fee motion. The court approved the requested expenses and service awards, finding them to be reasonable. However, the court found that the request for attorney's fees of $87.73 million was too high and instead awarded $80.6 million.

Five of the objectors challenge the district court's rulings in these four consolidated appeals. In case nos. 21-15758 and 21-15762, Sarah Feldman, Hondo Jan, and Deborah Pantoni (collectively, the “Feldman objectors”) challenge the settlement approval, the amount of attorney's fees, and the decision to grant service awards. In case no. 21-15761, Best Companies, Inc. (“BCI”) challenges the settlement approval as it relates to nonnatural persons. In case no. 21-15763, Anna St. John challenges the amount of attorney's fees.

II. Jurisdiction and Standard of Review

We have jurisdiction pursuant to 28 U.S.C. § 1291. We review a district court's decision to approve a class action settlement for clear abuse of discretion. Saucillo v. Peck, 25 F.4th 1118, 1129 (9th Cir. 2022). Under this “extremely limited” review, we will affirm if the district judge applies the proper legal standard and makes findings of fact that are not clearly erroneous. Id. (quoting In re Bluetooth Headset Prods. Liab. Litig., 654 F.3d 935, 940 (9th Cir. 2011)). We review the district court's decision to award attorney's fees and costs to class counsel, as well as the method of calculation, for abuse of discretion. In re Google Inc. St. View Elec. Commc'ns Litig., 21 F.4th 1102, 1110 (9th Cir. 2021).

We review legal questions de novo. See In re Volkswagen “Clean Diesel” Mktg., Sales Practices, & Prods. Liab. Litig., 914 F.3d 623, 640 (9th Cir. 2019). Legal questions include the interpretation of the settlement agreement, see Parsons v. Ryan, 949 F.3d 443, 453, 460 (9th Cir. 2020), “whether notice of a proposed settlement satisfies due process,” Torrisi v. Tucson Elec. Power Co., 8 F.3d 1370, 1374 (9th Cir. 1993), and whether Supreme Court precedent bars incentive awards altogether, see Chambers v. Whirlpool Corp., 980 F.3d 645, 656 (9th Cir. 2020) (reviewing a fee award's legal basis de novo).

III. Discussion

A. Nonnatural Persons Received Sufficient Notice of the Settlement

BCI contends that the district court provided inadequate notice of the settlement to nonnatural persons. Rule 23 requires that the district court “direct to class members the best notice that is practicable under the circumstances, including individual notice to all members who can be identified through reasonable effort.” Fed. R. Civ. P. 23(c)(2)(B). Procedural due process requires that the notice be “reasonably calculated, under all the circumstances, to apprise interested parties of the pendency of the action and afford them an opportunity to present their objections.” Roes, 944 F.3d at 1045 (quoting Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 174, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974)). But “neither Rule 23 nor the Due Process Clause requires actual notice to each individual class member.” Briseno v. ConAgra Foods, Inc., 844 F.3d 1121, 1128 (9th Cir. 2017).

The notice here satisfied both Rule 23 and due process. The settlement administrator contacted 99% of the persons associated with potentially eligible devices via the email and postal addresses in Apple's records and later sent follow-up notices to persons who had not yet responded. Additional class members received notice through the settlement's substantial coverage in the press and on social media.

BCI complains that “tying the notice solely to the Apple ID on an Affected Device ․ focused on notifying users of Affected Devices, but not necessarily the owners.” But the focus on users was reasonable given the requirement that claimants attest they experienced diminished performance on their devices. As BCI acknowledges, nonnatural persons “generally could not be users” of eligible devices. Moreover, BCI does not satisfactorily explain how the settlement administrator could have provided better notice to nonnatural persons.

While BCI speculates that Apple could have furnished corporate purchasers' contact information based on internal sales records, Apple informed the district court that it lacks that capability. Due process does not require Apple to perform impossible feats. Cf. Roes, 944 F.3d at 1047 n.9 (holding that “email notice would in no way have been ‘practicable under the circumstances’ ” where the defendants “did not have e-mail addresses for the class members”).

Similarly, BCI suggests that the parties could have identified corporate purchasers by “subpoena[ing] sales records from the major U.S. cellular carriers,” but there is no reason to assume that cellular carriers supply a significant share of corporate devices. Apple does not track corporate ownership of its devices in part because nonnatural persons purchase devices “in a variety of ways.” Even if the various third parties selling iPhones to corporations were not so numerous as to make subpoenaing them impracticable, such an approach would have risked delays and increased costs from litigation over the subpoenas. See, e.g., Ostrowski v., Inc., No. C16-1378-JCC, 2016 WL 4992051, at *1 (W.D. Wash. Sept. 16, 2016) (addressing third party retailer's objections to subpoena for customer information). Rule 23 and due process require only a “reasonable effort” to notify individual class members. Fed. R. Civ. P. 23(c)(2)(B); cf. In re Hyundai & Kia Fuel Econ. Litig., 926 F.3d 539, 568 (9th Cir. 2019) (en banc) (rejecting argument that defendant automakers “should have automatically made lump sum payments to class members” where the automakers lacked “complete records of resales of the class vehicles” and “could [not] have identified subsequent purchasers who were also part of the class”).

Lastly, BCI asserts that the parties could have given nonnatural persons constructive notice of the settlement through publication. But the free media coverage and individual notice to device users was more than adequate to reach nonnatural persons. The media covered the settlement in 2,670 pieces with a combined readership of approximately 7.31 million people. This case is thus readily distinguishable from Roes, where we found it “particularly problematic” that, despite concerns that certain class members “might be difficult to reach by mail, the settlement provided no other means of reaching [them].” Roes, 944 F.3d at 1046. Indeed, the large share of corporate claimants belies a lack of notice. More than one million nonnatural persons submitted claims, comprising nearly a third of the total. Of the 67 corporate class members who objected to the settlement, only BCI challenged the notice. 6

The notice to nonnatural persons was reasonable, and the district court did not abuse its discretion by authorizing it.

B. The Attestation Requirement Did Not Render the Settlement Unfair

For a settlement to be “fair, reasonable, and adequate,” Rule 23(e) requires that “the class representatives and class counsel have adequately represented the class,” “the relief provided for the class is adequate,” and the settlement “treats class members equitably relative to each other.” Fed. R. Civ. P. 23(e)(2)(A), (C), (D). The Feldman objectors contest each of these criteria. 7 However, most of their arguments boil down to the same core complaint: the settlement extinguishes the claims of “all former or current U.S. owners” of certain devices who downloaded iOS software before Apple disclosed potential defects, yet the settlement limits recovery to the subset of owners who can attest that “they experienced” the alleged defects.

The fundamental problem with the Feldman objectors' arguments is their assumption that “all Class members suffered the same impairment of iPhone performance and uniform damages.” Throughout this litigation, Apple has disputed “that all devices were used in a way that would have activated the performance management feature” that slowed system performance. Apple insists that “even when [the feature] was activated, users may not have ․ noticed any differences.” Although plaintiffs alleged that “the iOS updates affected all Plaintiffs alike,” the parties agreed to the attestation requirement as a compromise.

That compromise was reasonable. It reflected “the bargaining and compromise inherent in settling disputes.” California v. IntelliGender, LLC, 771 F.3d 1169, 1179 (9th Cir. 2014). At the time of the settlement, the only claims that remained—concerning Apple's alleged failure to disclose the nature of the iOS updates—all required a showing of damages for a plaintiff to recover. Neither the federal Computer Fraud and Abuse Act nor California's Data Access and Fraud Act provides for statutory damages, see 18 U.S.C. § 1030(g); Cal. Penal Code § 502(e)(1), and trespass to chattels is not actionable without damage to or interference with the phone's operation. See hiQ Labs, Inc. v. LinkedIn Corp., 31 F.4th 1180, 1201 n.21 (9th Cir. 2022). If a class member did not perceive and could not otherwise detect the throttling, then the class member's inability to make the attestation relinquished a valueless claim. Apple, on the other hand, risked increased liability by proceeding to trial because plaintiffs might have proven that the system slowdowns affected every device. The settlement allowed Apple to limit its exposure while ensuring that compensation was available to every class member who suffered a compensable injury.

That not every class member had an actionable claim is not tantamount to two adverse groups requiring separate representation, as in Amchem Products, Inc. v. Windsor, 521 U.S. 591, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997). In Amchem, every class member suffered an injury—exposure to asbestos—but the injuries manifested over different time horizons “because of a latency period that may last as long as 40 years.” Id. at 598, 117 S.Ct. 2231. Class members whose injuries had already manifested sought “generous immediate payments,” bringing them into conflict with the other class members who sought an “ample, inflation-protected fund for the future.” Id. at 626, 117 S.Ct. 2231. Thus, “the interests of those within the single class [were] not aligned” in “significant respects.” Id. No such conflict exists here. All class members who were injured by Apple's failure to disclose the nature of the iOS updates experienced injury during the same time frame and in the same manner.

Nor does the possibility that some class members suffered no damages mean that they lack standing and must be dismissed. 8 While class members “must maintain their personal interest in the dispute at all stages of litigation” to have a judgment bind them, TransUnion LLC v. Ramirez, ––– U.S. ––––, 141 S. Ct. 2190, 2208, 210 L.Ed.2d 568 (2021), “[a]t the pleading stage, general factual allegations of injury resulting from the defendant's conduct may suffice,” Lujan v. Defs. of Wildlife, 504 U.S. 555, 561, 112 S.Ct. 2130, 119 L.Ed.2d 351 (1992). At the time the parties settled, prior to class certification or summary judgment, plaintiffs alleged that all putative class members experienced throttling from Apple's allegedly unlawful intrusion into their phones. That sufficed to establish standing. Had the parties brought the case to trial, as in TransUnion, 141 S. Ct. at 2202, plaintiffs' allegation of classwide injury would have been either proven or disproven. The risk that this uncertain outcome posed to the parties was one of the factors that induced them to settle.

C. The District Court Cited the Wrong Legal Standard in Examining the Settlement's Fairness

The Feldman objectors advance other reasons why the settlement was unfair, such as the size of the recovery and language on the claim form that, they argue, prevented former device owners from recovering. We need not reach these issues. More fundamentally, the Feldman objectors argue that the district court “improperly ․ applied a presumption of reasonableness” to the settlement rather than “applying heightened scrutiny.” 9

Although appellate review of a settlement's substantive fairness is limited, “we hold district courts to a higher procedural standard when making that determination of substantive fairness: ‘To survive appellate review, the district court must show it has explored comprehensively all [Rule 23(e)(2)] factors, and must give a reasoned response to all non-frivolous objections.’ ” McKinney-Drobnis v. Oreshack, 16 F.4th 594, 606 (9th Cir. 2021) (quoting Allen v. Bedolla, 787 F.3d 1218, 1223–24 (9th Cir. 2015)). And “when a settlement precedes class certification ․ , the district court must apply ‘an even higher level of scrutiny.’ ” Id. at 607 (quoting Roes, 944 F.3d at 1049). This additional scrutiny requires the court to look for and scrutinize “any subtle signs that class counsel have allowed pursuit of their own self-interests to infect the negotiations.” Id. (quoting Roes, 944 F.3d at 1043).

When granting final settlement approval, the district court carefully considered and thoughtfully responded to “the more serious, common, and representative objections.” But notwithstanding its careful consideration of the settlement's fairness, the court's written order explicitly presumed that the settlement was fair and reasonable.

At the outset, the district court stated that it “should give ‘proper deference to the private consensual decision of the parties’ ” (quoting Hanlon, 150 F.3d at 1027), 10 and that the “recommendations of plaintiffs' counsel should be given a presumption of reasonableness” (quoting Knight v. Red Door Salons, Inc., No. 08-01520 SC, 2009 WL 248367, at *4 (N.D. Cal. Feb. 2, 2009)). The district court also cited “a presumption that the agreement is fair” given “[t]he involvement of experienced class action counsel and the fact that the settlement agreement was reached in arm's length negotiations, after relevant discovery had taken place” (quoting Linney v. Cellular Alaska P'ship, No. C-96-3008 DLJ, 1997 WL 450064, at *5 (N.D. Cal. July 18, 1997)). And the district court suggested that its scrutiny was limited to “obvious deficiencies” (quoting In re Zynga Inc. Sec. Litig., No. 12-CV-04007-JSC, 2015 WL 6471171, at *8 (N.D. Cal. Oct. 27, 2015)), rather than any “subtle signs” of collusion, McKinney-Drobnis, 16 F.4th at 607.

In Roes, we reversed the district court for “begin[ning] its analysis with a presumption that the settlement [was] fair and reasonable.” Roes, 944 F.3d at 1049 (emphasis omitted). We observed that this presumption once “was commonly applied by district courts,” id. at 1049 n.12, but is nonetheless “erroneous” as to “settlements negotiated prior to class certification,” id. at 1049.

Here, as in Roes, the district court cited the wrong legal standard and failed to “acknowledge ․ longstanding [circuit] precedent requiring a heightened fairness inquiry prior to class certification.” Id. In Roes, the problems ran deeper than an incorrect statement of the law; the district court did not “investigate or adequately address” the “numerous problematic aspects of the settlement and subtle signs of implicit collusion.” Id. Here, while the district court's probing analysis suggests that it may have applied heightened scrutiny, its written order relied on a flawed legal standard.

“[A]pplication of an incorrect legal standard alone constitutes an abuse of discretion,” Saucillo, 25 F.4th at 1131 n.9, and this type of error is not subject to review for harmlessness, see id. at 1133. Therefore, the district court abused its discretion by stating that it applied a presumption of reasonableness and fairness to the settlement. We vacate the order granting final settlement approval so that on remand the district court can evaluate the settlement under the correct standard.

In light of our vacatur of the settlement approval, we also vacate the district court's order awarding attorney's fees, expenses, and incentive payments. However, we address two issues that appellants raise regarding the latter order to provide guidance in the event the district court again approves the settlement.

D. The District Court Should Reconsider the Inclusion of JCCP Work in the Lodestar Calculation

The Feldman objectors contend that the district court abused its discretion by awarding attorney's fees based in part on hours that JCCP counsel spent on the parallel state litigation. While we express no opinion as to whether the inclusion of JCCP-related work was reasonable, we agree with objectors that the district court's explanation for considering JCCP-related work conflicted with the court's overall rationale for its fee award.

When calculating an attorney's fee award, a district court can employ one of two methods—the lodestar or a percentage of the recovery. See Kim v. Allison, 8 F.4th 1170, 1180 (9th Cir. 2021). Under the lodestar method, the district court “multiplies the number of hours the prevailing party reasonably spent on litigation by a reasonable hourly rate to determine a presumptively reasonable fee award”—the lodestar. Id. The lodestar amount can then be adjusted “by an appropriate positive or negative multiplier” to account for factors such as “the quality of representation, the benefit obtained for the class, the complexity and novelty of the issues presented, and the risk of nonpayment.” Id. at 1180–81 (quoting Bluetooth Headset, 654 F.3d at 941–42).

The percentage-of-recovery method expresses fees as a percentage of a recovered common fund, calculated to be “sufficient to provide class counsel with a reasonable fee.” Id. at 1181 (quoting Hanlon, 150 F.3d at 1029). The benchmark percentage is 25%, but, similar to the lodestar, the benchmark percentage “can be adjusted upward or downward, depending on the circumstances.” Hyundai & Kia Fuel Econ., 926 F.3d at 570.

District courts “have discretion to employ either the lodestar method or the percentage-of-recovery method.” Id. at 570 (quoting Bluetooth Headset, 654 F.3d at 942). Whichever method is chosen, courts often employ the other method as a cross-check that the award is reasonable. See In re Online DVD-Rental Antitrust Litig., 779 F.3d 934, 949 (9th Cir. 2015). A cross-check is discretionary, but we encourage one when utilizing the percentage-of-recovery method. See In re Optical Disk Drive Prods. Antitrust Litig., 959 F.3d 922, 930 (9th Cir. 2020).

Here, the district court found that 26% of the $310 million fund—more than the benchmark but less than the 28.3% requested by class counsel—was a reasonable percentage. 11 The court then performed a lodestar analysis as a cross-check. Because 26% of the fund exceeded the lodestar amount by about $44.5 million, the district court's award reflected a 2.232 positive multiplier. The district court found that this multiplier was “well justified” but “an award exceeding a 2.232 multiplier would result in ‘windfall profits for class counsel in light of the hours spent on the case’ ” (quoting Bluetooth Headset, 654 F.3d at 942).

The lodestar amount claimed by class counsel and accepted by the district court included nearly $4 million in attorney's fees generated by JCCP counsel in the state proceedings. The Feldman objectors argue these hours “increased inefficiency and duplication” and JCCP counsel “does not represent parties in this litigation.”

The district court agreed that this argument (made below by Apple) “has some merit.” However, the court declined to consider it. In the court's view, “the inclusion of JCCP Counsel's hours does not have any bearing on the ․ percentage-of-the-fund calculation.” It reasoned that removing JCCP counsel's $4 million in fees from the lodestar “would have an insignificant impact on the multiplier,” since “[m]ultipliers of 1 to 4 are commonly found to be appropriate in common fund cases” (quoting Aboudi v. T-Mobile USA, Inc., No. 12CV2169 BTM NLS, 2015 WL 4923602, at *7 (S.D. Cal. Aug. 18, 2015)).

The district court's rationale directly contradicts its finding that “an award exceeding a 2.232 multiplier” was unwarranted because it would represent a “windfall profit[ ]” to class counsel. If the court had excluded JCCP counsel's fees from the lodestar, then the $80.6 million fee award would represent a multiplier of 2.51 over the lodestar of $32.1 million. That's more than even the 2.43 multiplier the court described as “high” when rejecting class counsel's full fee request. Thus, the impact of the JCCP fees on the multiplier was not at all “insignificant,” and the district court's failure to consider whether the JCCP fees should be included in the lodestar was an abuse of discretion.

E. Supreme Court Precedent Does Not Foreclose Incentive Payments to Class Representatives

The Feldman objectors contend that district courts lack discretion to award any service fees or incentive payments to class representatives. 12 We have repeatedly held that “reasonable incentive awards” to class representatives “are permitted,” Roes, 944 F.3d at 1057 (citing Rodriguez v. W. Publ'g Corp., 563 F.3d 948, 958–59 (9th Cir. 2009)), and the Supreme Court recently acknowledged that “[a] class representative might receive a share of class recovery above and beyond her individual claim” through an incentive award, China Agritech, Inc. v. Resh, ––– U.S. ––––, 138 S. Ct. 1800, 1811 n.7, 201 L.Ed.2d 123 (2018). Nonetheless, the Feldman objectors contend that our twenty-first century precedent allowing such awards conflicts with Supreme Court precedent from the nineteenth century—Trustees v. Greenough, 105 U.S. 527, 26 L.Ed. 1157 (1881), and Central Railroad & Banking Co. v. Pettus, 113 U.S. 116, 5 S.Ct. 387, 28 L.Ed. 915 (1885). 13 To the contrary, we have previously considered this nineteenth century caselaw in the context of incentive awards and found nothing discordant.

Greenough and Pettus established “the ‘common fund doctrine,’ a traditional equitable doctrine ‘rooted in concepts of quasi-contract and restitution.’ ” Rodriguez v. Disner, 688 F.3d 645, 653 (9th Cir. 2012) (quoting Vincent v. Hughes Air W., Inc., 557 F.2d 759, 770 (9th Cir. 1977)). Under this doctrine, “a private plaintiff, or his attorney, whose efforts create, discover, increase or preserve a fund to which others also have a claim is entitled to recover from the fund the costs of his litigation, including attorneys' fees.” Vincent, 557 F.2d at 769. The doctrine's “fundamental purpose ․ is to spread the burden of a party's litigation expenses among those who are benefited.” Paul, Johnson, Alston & Hunt v. Graulty, 886 F.2d 268, 271 (9th Cir. 1989).

While private plaintiffs who recover a common fund are entitled to “an extra reward,” they are limited to “that which is deemed ‘reasonable’ under the circumstances.” Id. Greenough, for example, prohibited recovery for the plaintiff's “personal services and private expenses” because the private plaintiff was a creditor who needed no inducement to bring suit. Greenough, 105 U.S. at 537. If private plaintiffs “could calculate upon the allowance of a salary for their time and of having all their private expenses paid,” the Court explained, it “would present too great a temptation.” Id. at 538. Private parties would be inclined “to intermeddle in the management of valuable property or funds in which they have only the interest of creditors, and that perhaps only to a small amount.” Id.

In class action litigation, the common fund doctrine supports reasonable awards to a litigant or lawyer. See Rodriguez, 688 F.3d at 653; see also Staton v. Boeing Co., 327 F.3d 938, 977 (9th Cir. 2003) (holding that “named plaintiffs ․ are eligible for reasonable incentive payments” in addition to reimbursement “for their substantiated litigation expenses, and identifiable services rendered to the class directly under the supervision of class counsel”). In Staton, we explained that the common fund doctrine allows reasonable incentive payments to class representatives but not “special rewards.” 327 F.3d at 976. An incentive payment cannot be so large that it amounts to “a preferred position in the settlement,” id. (quoting Officers for Justice, 688 F.2d at 632), or “a salary,” Greenough, 105 U.S. at 538. Thus, district courts must “evaluate the propriety of requested incentive payments” by considering, among other factors, “the actions the plaintiff has taken to protect the interests of the class, the degree to which the class has benefitted from those actions, the amount of time and effort the plaintiff expended in pursuing the litigation,” and any financial or reputational risks the plaintiff faced. Roes, 944 F.3d at 1057 (cleaned up) (quoting Staton, 327 F.3d at 977).

In Greenough, the Supreme Court disapproved a 10-year allowance for “personal services,” equivalent today to approximately $76,000 per year, as well as “personal expenditures” on “railroad fares and hotel bills” worth approximately $458,000 today. 14 Greenough, 105 U.S. at 530. That is a far cry from the $1,500 that the district court awarded most of the class representatives here.

There are, of course, cases where the incentive award is more aptly analogized to a salary—in such cases, the award exceeds the amount necessary to serve its purpose and creates the very conflict between class representatives and other class members that Greenough and Pettus sought to prevent. See, e.g., Roes, 944 F.3d at 1056 (criticizing $20,000 incentive payments that “appear to be completely divorced from any benefit or service to the class”); Staton, 327 F.3d at 977 (rejecting incentive payments averaging more than $30,000). The point is that incentive awards cannot categorically be rejected or approved. See Online DVD-Rental, 779 F.3d at 943 (“Incentive payments to class representatives do not, by themselves, create an impermissible conflict between class members and their representatives.”). So long as they are reasonable, they can be awarded.

IV. Conclusion

Because the district court applied the wrong legal standard when reviewing the settlement's fairness, we vacate the orders granting final settlement approval and awarding fees, expenses, and incentive awards, and we remand for application of the correct standard. The parties shall bear their own costs on appeal.



1.   Other factors, in particular cold conditions and low battery charge, also affect a battery's ability to supply peak electrical demands.

2.   The settlement applied to iPhone 6, 6 Plus, 6s, 6s Plus, and SE devices that ran iOS 10.2.1 or later and iPhone 7 and 7 Plus devices that ran iOS 11.2 or later.

3.   Apple's records showed which devices had downloaded a particular version of iOS and when the download occurred, but Apple apparently did not know when, if ever, a device owner installed the iOS after downloading it. Some iPhones had the applicable iOS pre-installed, and Apple included those devices in the information it provided to the settlement administrator.

4.   Of the 974,300 disapproved claims, 73% did not match an eligible device, 17% had inadequate attestations, and 10% were duplicative or had uncured deficiencies.

5.   Based on the number of claims, the settlement agreement required Apple to pay the minimum of $310 million, including settlement administration costs, leaving about $297.25 million available for distribution to potentially 2,310,439 claimants. The amount available per claim would exceed $128 if some of the 41,579 pending claims were disapproved. And to the extent the number of claimants exceeded the number of eligible devices (for example, if both corporate purchasers and their employees using the phones made claims), the recovery per device would exceed the average recovery per claimant.

6.   We easily reject BCI's argument that the settlement does not apply to nonnatural persons. The settlement references “individuals,” but that word is not “a legal term of art that applies only to natural persons,” and its ordinary meaning “does not necessarily exclude corporations.” United States v. Middleton, 231 F.3d 1207, 1210 (9th Cir. 2000). Regardless, while the settlement includes “individuals,” it applies more broadly to iPhone “owners”—a term that plainly encompasses nonnatural persons. And the settlement agreement is not entirely silent as to nonnatural persons; it protects them from fraud by allowing the settlement administrator to reject a claim form if “the person submitting [it] requests that payment be made to [an] ․ entity other than the Settlement Class Member for whom the Claim Form is submitted.”

7.   The Feldman objectors also argue that the district court should have made detailed findings that Rule 23(a) and (b)(3)'s prerequisites were satisfied. We have previously rejected a “claim [of] error in the brevity of [such] findings” where “the record provides more than adequate foundation” for review, Hanlon v. Chrysler Corp., 150 F.3d 1011, 1023 (9th Cir. 1998), and this case is no different. Moreover, the crux of the Feldman objectors' challenge to the commonality, typicality, adequacy, and predominance requirements, see Fed. R. Civ. P. 23(a)(2)–(4), (b)(3), concerns the class definition—namely, that it includes class members who aren't entitled to compensation because they couldn't attest to injury. We consider this issue, as to which the district court made detailed findings, in connection with Rule 23(e)(2).

8.   We say “possibility” because it is unknown why many class members submitted no claim. It is possible that some of these class members could not in good faith attest to experiencing a system slowdown, but it is also possible that they simply lacked interest in pursuing compensation.

9.   The Feldman objectors first challenged the district court's level of scrutiny in their reply brief. Normally, arguments not raised in an opening brief are forfeited. See, e.g., Iraheta-Martinez v. Garland, 12 F.4th 942, 959 (9th Cir. 2021). We can, however, exercise “discretion to consider a purely legal question when the record relevant to the matter is fully developed.” Saucillo, 25 F.4th at 1130 n.7 (quoting United States v. Northrop Corp., 59 F.3d 953, 957 n.2 (9th Cir. 1995)). We do so here because, as in Saucillo, the issue concerns the appropriate legal standard and “the district court ․ overlayed its entire discussion of the settlement agreement with [an] erroneous presumption.” Id. at 1132.

10.   The district court acknowledged that deference to the parties' consensual decision “must be limited” to ensure “that the agreement is not the product of fraud or overreaching by, or collusion between, the negotiating parties, and that the settlement, taken as a whole, is fair, reasonable and adequate to all concerned.” Hanlon, 150 F.3d at 1027 (quoting Officers for Just. v. Civ. Serv. Comm'n, 688 F.2d 615, 625 (9th Cir. 1982)). But the district court did not acknowledge Hanlon's holding that “settlement approval that takes place prior to formal class certification requires a higher standard of fairness.” Id. at 1026.

11.   It's not entirely clear how the district court arrived at this percentage. The court inaccurately characterized 26% as “the benchmark in the Ninth Circuit,” but it also recognized that 26% was a small upward adjustment when it found that “a larger upward adjustment” was unwarranted. The record suggests that the district court may have reached 26% by balancing various factors supporting an upward or downward adjustment to the benchmark. See Vizcaino v. Microsoft Corp., 290 F.3d 1043 (9th Cir. 2002). If, on remand, the district court again awards attorney's fees and uses the percentage-of-recovery method, it should more clearly explain why a deviation from the benchmark is or isn't appropriate before proceeding to a lodestar cross-check.

12.   In the alternative, the Feldman objectors argue that the awards here were an abuse of discretion, an issue we do not address.

13.   The Second Circuit rejected a similar argument. Melito v. Experian Mktg. Sols., Inc., 923 F.3d 85, 96 (2d Cir. 2019). The Eleventh Circuit also rejected it, Muransky v. Godiva Chocolatier, Inc., 922 F.3d 1175, 1196 (11th Cir. 2019), but later vacated the opinion and, on rehearing en banc, dismissed the case for lack of standing, 979 F.3d 917 (11th Cir. 2020). Meanwhile, the majority of another Eleventh Circuit panel reached the opposite conclusion—that Greenough and Pettus prohibit any incentive award to class representatives. Johnson v. NPAS Sols., LLC, 975 F.3d 1244, 1255 (11th Cir. 2020).

14.   See Fed. Rsrv. Bank of Minneapolis, Consumer Price Index, 1800-, (last visited Sept. 19, 2022) (showing that the cost of living has increased 30.5-fold between 1881, when the Supreme Court decided Greenough, and 2022).

NGUYEN, Circuit Judge:

Copied to clipboard