OPINION

GOULD, Circuit Judge:

This case asks us to decide whether a settlement agreement reached before class certification between Plaintiffs and Defendants is fair, reasonable, and adequate. We agree with the district court that the settlement merits approval, and we affirm.

I

Coverall North America, Inc. ("Coverall") is a janitorial franchising company operating in California. Plaintiffs brought a class action suit against Coverall in 2009 alleging that (1) Coverall misclassified its California franchisees as independent contractors, thereby avoiding the protections afforded by California's labor laws to franchisees; and (2) Coverall breached its franchise agreements, and committed fraudulent and unfair practices, by removing customer accounts from franchisees without cause so that it could resell those accounts to other franchisees. In August 2011, after about two years of significant litigation, the parties agreed on a settlement. The sole objector, Amrit Singh, filed an objection to the proposed settlement on November 14, 2011, and although the objection was not timely, the district court accepted the filing "in the interest of determining the issues on the merits." After a fairness hearing on November 21, 2011, the district court approved the settlement agreement on February 23, 2012 pursuant to Federal Rule of Civil Procedure 23(e).

The settlement agreement is expansive, but the most contested provisions include the following: (1) Coverall pledges to assign customer accounts to current franchisees, with the assignments remaining conditional until franchisees have paid their franchise fees in full; (2) former franchise owners will receive $475 each and will receive a $750 purchase credit toward a new Coverall franchise; and (3) new franchisees will have a 30-day right to rescind their franchise agreements, and upon rescission will receive all the of money they have paid during that period under the franchise agreement except for the $75 background investigation fee. The settlement agreement also outlines other changes to the franchise agreements and Coverall's operating procedures. Beyond the agreement generally, Singh contests the award of $994,800 in attorneys' fees to Plaintiffs' attorneys. As of the fairness hearing on November 21, 2011, two class members had opted out of the agreement and Singh is the only objector.

II

We review the district court's approval of a proposed class action settlement agreement for abuse of discretion. Rodriguez v. W. Publ'g Corp., 563 F.3d 948, 963 (9th Cir.2009); see also United States v. Hinkson, 585 F.3d 1247, 1250 (9th Cir.2009) (en banc) (giving general abuse of discretion standard in contexts beyond class actions). Our review of a class action settlement is "extremely limited," In re Mego Fin. Corp. Sec. Litig., 213 F.3d 454, 458 (9th Cir.2000), and we will only reverse upon "a strong showing that the district court's decision was a clear abuse of discretion," Staton v. Boeing Co., 327 F.3d 938, 960 (9th Cir.2003) (quotation marks and citations omitted). We also review for abuse of discretion the calculation and award of attorneys' fees. In re Bluetooth Headset Prods. Liab. Litig., 654 F.3d 935, 940 (9th Cir.2011).

III

When a class action settlement agreement is submitted for approval, "[t]he initial decision to approve or reject a settlement proposal is committed to the sound discretion of the trial judge." Officers for Justice v. Civil Serv. Comm'n, 688 F.2d 615, 625 (9th Cir.1982). The district court judge must determine whether the settlement is "fundamentally fair, adequate and reasonable." Id.

We start with fees. Although the parties have agreed on the award of attorneys' fees, the district court has an "independent obligation to ensure that the award, like the settlement itself, is reasonable." Bluetooth, 654 F.3d at 941. However, we recognize that in the settlement context fees are a subject of compromise. Staton, 327 F.3d at 966. We have made clear that "since the proper amount of fees is often open to dispute and the parties are compromising precisely to avoid litigation, the [district] court need not inquire into the reasonableness of the fees at even the high end with precisely the same level of scrutiny as when the fee amount is litigated." Id.

We note at the outset that two different methods may be used "for calculating a reasonable attorneys' fee depending on the circumstances." Bluetooth, 654 F.3d at 941. The lodestar method is most appropriate where the relief sought is "primarily injunctive in nature," and a fee-shifting statute authorizes "the award of fees to ensure compensation for counsel undertaking socially beneficial litigation." Id.; see also Hanlon v. Chrysler Corp., 150 F.3d 1011, 1029 (9th Cir.1998). That is precisely the situation we face here, because the settlement provisions that Plaintiffs have sought and agreed upon are mostly injunctive in nature and a fee shifting statute exists, California Business and Professions Code § 17082. We conclude that the district court correctly used the lodestar method in gauging the fairness of the attorneys' fee award.

The district court also correctly calculated the lodestar amount and reasonably concluded that the agreed upon award, $994,800, was appropriate. In its analysis, the district court noted that the case had been contentiously litigated for over two years, and that the report submitted by Plaintiffs' counsel showing that over 4,500 hours had been billed by six attorneys, a paralegal, and a law clerk was fair and accurate. Using that number, the district court calculated that the lodestar amount reached almost $3 million. At a third of the lodestar amount, the district court soundly concluded that the attorneys' fee award of $994,800 was reasonable.

Moreover, the district court prudently cross-checked the award amount against the alternative percentage-of-recovery method. We have "encouraged courts to guard against an unreasonable result by cross-checking their calculations against a second method." Bluetooth, 654 F.3d at 944-45. Generally, courts use a benchmark figure of 25% to gauge the reasonableness of an award under the percentage-of-recovery method, which is most appropriate in common fund settlement cases. Id. at 942; In re Mercury Interactive Corp. Sec. Litig., 618 F.3d 988, 992 (9th Cir.2010); Six (6) Mexican Workers v. Ariz. Citrus Growers, 904 F.2d 1301, 1311 (9th Cir.1990). Here, the district court correctly noted that the value of the settlement, and particularly its injunctive terms, was disputed. While Singh's figure of $56,525 is clearly incorrect because it gives no value to the injunctive terms of the settlement, Plaintiffs may certainly be overstating the value of the settlement at $20 million. The district court reasonably surmised that even if the value of the settlement was $4 million — only a part of the amount claimed by Plaintiffs — the attorneys' fee award would still be within the normal bounds of reasonableness. The district court was within its discretion to find the attorneys' fee award to be fair, reasonable, and adequate because it was both significantly below the lodestar amount and represented an unobjectionable percentage of recovery once the value of injunctive relief was considered.

Singh's main argument against the reasonableness of the attorneys' fee award is that the actual value of the settlement, which he characterizes as primarily the amount of the cash payments, is so low that the award is unreasonable. Singh correctly notes that "the benefit obtained for the class" is important in determining whether to adjust the lodestar amount and by how much, Hanlon, 150 F.3d at 1029, but any such adjustment is equitable and squarely at the discretion of the district court, Hensley v. Eckerhart, 461 U.S. 424, 437, 103 S.Ct. 1933, 76 L.Ed.2d 40 (1983), and Singh presents no evidence that the district court abused its discretion in declining further adjustment from the lodestar amount. Moreover, the district court acted within its proper discretion when it found that the settlement contains significant benefits for Plaintiffs beyond the cash recovery, and thus that the award, at about a third of the lodestar amount, was reasonable.

IV

Turning to the settlement as a whole, Federal Rule of Civil Procedure 23(e) "requires the district court to determine whether a proposed settlement is fundamentally fair." Hanlon, 150 F.3d at 1026. As a general rule, a district court must consider the following factors when examining the fairness of a proposed settlement:

The district court also agreed with Plaintiffs that California employment law would likely make obtaining class certification particularly difficult. Following the Supreme Court's decision in Wal-Mart Stores, Inc. v. Dukes, ___ U.S. ___, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011), there was a real prospect that California's multi-factor test for employee classification would have proven fatal to any eventual class certification in this case. As an additional risk, the district court found that the poor financial health of Coverall seriously increased the chance that Plaintiffs would be left with nothing if they continued to litigate their claims. Rounding out its analysis, the district court noted that no governmental entity had weighed in on the matter, that Plaintiffs' attorneys had significant experience and had demonstrated skill and diligence throughout the litigation, and that only two class members had opted out of the agreement.

Further, the district court reasonably found that the settlement would yield significant benefits for Plaintiffs given the risks and costs of continuing litigation. The district court found that, beyond the cash for former franchisees, "assignment of customer accounts and pledges for programmatic changes are significant victories." The district court elaborated that "once franchises are assigned, franchisees will own a valuable business they can choose to sell or continue to operate." Viewed together, the district court determined that the Churchill factors supported the conclusion that the settlement was fair, reasonable, and adequate.

Singh claims that the district court was "under a special obligation to make clear, fact-based findings regarding the value of the non-monetary terms of the settlement," by which he seems to contend that the district court should have assigned a monetary value to the non-monetary terms of the settlement. But we have never required a district court to assign a monetary value to purely injunctive relief. To the contrary, we have stated that courts cannot "judge with confidence the value of the terms of a settlement agreement, especially one in which, as here, the settlement provides for injunctive relief." Staton, 327 F.3d at 959. Monetary valuation of injunctive relief is difficult and imprecise.1 As such, we "put a good deal of stock in the product of an arms-length, non-collusive, negotiated resolution, and have never prescribed a particular formula by which that outcome must be tested." Rodriguez, 563 F.3d at 965 (internal citations omitted). The district court has no obligation to make explicit monetary valuations of injunctive remedies, and it did not abuse its discretion in applying the Churchill factors to this case.

Singh also argues that the district court abused its discretion in not exercising heightened review given the alleged presence of "warning signs" indicating collusion. Although we must be particularly vigilant in the pre-class certification context, when there are typically more opportunities for attorney collusion, Hanlon, 150 F.3d at 1026, we will rarely overturn an approval of a compromised settlement "unless the terms of the agreement contain convincing indications that the incentives favoring pursuit of self-interest rather than the class's interests in fact influenced the outcome of the negotiations and that the district court was wrong in concluding otherwise," Staton, 327 F.3d at 960.

In Bluetooth, we outlined three settlement arrangements that could indicate collusion because they may improperly favor counsel at the expense of the plaintiffs. 654 F.3d at 946-47. These are:

Id. at 947 (internal quotation marks and citations omitted). Although the district court did not explicitly outline these "warning signs," contrary to Singh's assertion, it did specifically address two of the three. The district court found the third sign, the presence of a reversion clause, to "not [be] a preferable result," but balanced it with the overall benefits of the settlement to Plaintiffs and the fact that the cash payment represented a small amount of those benefits.

The district court's analysis, balancing the reversion clause against the overall strength of the settlement, was adequate. The first warning sign was not present because the district court correctly concluded that the attorneys' fee award was entirely reasonable. This conclusion has broad implications for the district court's obligation to ensure "that the settlement is not the product of collusion among the negotiating parties." Id. (internal quotation marks, citation, and alteration omitted). Because collusion is the product of attorneys pursuing their self-interest to the detriment of the class's interests, one would expect primarily to find collusion where attorneys disproportionately benefitted from the settlement. Id. at 947-48; see Staton, 327 F.3d at 964 ("If fees are unreasonably high, the likelihood is that the defendant obtained an economically beneficial concession with regard to the merits provisions [of the settlement]."). A settlement may still be the product of collusion when attorneys' fees are reasonable. But when, as in this case, the fee award is clearly reasonable as viewed through the appropriate application of either the lodestar or percentage-of-recovery methods, the chance of collusion narrows to a slim possibility. In these cases, it is sufficient that a district court recognizes and balances potentially collusive provisions, such as the reversion to defendants of unclaimed funds, against the other terms of the settlement agreement. See Bluetooth, 654 F.3d at 948 ("But these factors did not obviate the need to examine the fee provision in light of the rest of the agreement."); Staton, 327 F.3d at 961 ("[I]t will be rare that we will reverse a district court's approval ... unless the fees and relief provisions clearly suggest the possibility that class interests gave way to self-interest."); Hanlon, 150 F.3d at 1026 ("[I]t is the settlement taken as a whole, rather than the individual component parts, that must be examined for overall fairness.").

Plaintiffs faced real dangers in proceeding on their case in light of menacing precedents from the United States Supreme Court. At the same time, the class gained significant benefits from the settlement, and Plaintiffs' lawyers received fees that are overall reasonable. The district court correctly examined the settlement agreement and did not abuse its discretion in finding the agreement to be fair, reasonable, and adequate.

V

The district court's decision to grant discovery is reviewed for abuse of discretion. Hallett v. Morgan, 296 F.3d 732, 751 (9th Cir.2002). As a threshold matter, Singh has standing to appeal the settlement's deposition requirement because the district court ordered that objectors comply with the relevant settlement provision, and he was later required to sit for a deposition as the sole objector.

The district court also did not abuse its discretion in approving the settlement term that objectors be available for depositions. Federal Rule of Civil Procedure 30(a)(1) allows a party to conduct depositions, and courts commonly require objectors to make themselves available for deposition given the power held by objectors. See, e.g., In re Netflix Privacy Litig., 2013 WL 6173772 (N.D.Cal.2013) (finding that objectors "have voluntarily inserted themselves into this action, and as such, depositions ... are relevant and proper"); In re TFT-LCD (Flat Panel) Antitrust Litig., 289 F.R.D. 548 (N.D.Cal. 2013) (holding objectors in contempt for refusing to be deposed); In re Cathode Ray Tube (CRT) Antitrust Litig., 281 F.R.D. 531, 533 (N.D.Cal.2012). The district court considered the totality of the circumstances when it concluded that a deposition of Singh was appropriate, including the deposition's utility to the court and the scant possibility that Singh would be harassed or intimidated by giving a deposition, and the district court was well within its discretion in so concluding.

VI

Finally, the district court did not abuse its discretion when it approved the settlement agreement consistent with the Class Action Fairness Act ("CAFA") notice requirement described in 28 U.S.C. § 1715(b) and (d). CAFA requires that defendants in a class action suit send notice to all relevant state and federal authorities where class members reside. 28 U.S.C. § 1715(b). A court may not issue an order giving final approval of a proposed settlement until 90 days have passed since the relevant authorities were served with notice. 28 U.S.C. § 1715(d). When violations of the 28 U.S.C. § 1715(b) notice requirement occur, "[a] class member may refuse to comply with and may choose not to be bound by a settlement agreement or consent decree." 28 U.S.C. § 1715(e)(1). Rather than asking to be exempt from the settlement agreement, Singh demands relief beyond the scope of 28 U.S.C. § 1715 — the rejection of the entire settlement agreement. Because Singh requests no relief tied to § 1715, he cannot show that a "favorable decision will provide redress," and thus he lacks standing on this claim. See Knisley v. Network Assocs., Inc., 312 F.3d 1123, 1126 (9th Cir.2002). And even if Singh had standing to pursue this claim, it is without merit because the class was clearly limited to "individuals in the State of California," and Coverall properly notified the Attorney General of California.

VII

The district court did not abuse its discretion in finding the settlement agreement and its attorneys' fee award to be fair, reasonable, and adequate. The district court also appropriately exercised its discretion in ordering that Singh be available for a deposition.2

AFFIRMED.

CHEN, District Judge, dissenting:

With reluctance, I dissent from the majority decision affirming the district court's approval of the proposed class action settlement. The record below is bereft of crucial information — information without which the district court could not fully review either the adequacy of the settlement or the reasonableness of the fee award. In particular, the record is silent as to two essential matters: (1) the portion of the class eligible to receive the chief non-monetary benefit of the settlement (i.e., the assignment of customer accounts to current franchisees) and (2) the value of the monetary relief to the class (and whether there is a justification for imposing a claims process with a reverter of unclaimed funds back to the defendant). The case should be remanded for fuller development of the record. I also believe this case affords this Court an opportunity to provide additional guidance to the district courts in their assessment of proposed class action settlements.

I am well aware that there are good reasons for a district judge to afford deference to a settlement reached by agreement between the parties. See Lane v. Facebook, Inc., 696 F.3d 811, 819 (9th Cir.2012) (noting that one factor for a court to consider in determining the fairness of a class action settlement is the experience and views of counsel). The parties know more about the case and the factors that lead to settlement than the trial judge does; often there are facts beyond the record (such as the finances or future business plans of the defendant or infirmities with the plaintiff's ability to prosecute the case) which can have a substantial and legitimate influence on settlement negotiations. I also recognize that there are good reasons for an appellate court to defer to the trial judge's approval of a settlement. See Hanlon v. Chrysler Corp., 150 F.3d 1011, 1026 (9th Cir.1998) (stating that "[o]ur review of the district court's decision to approve a class action settlement is extremely limited"). The trial judge generally is in a better position to ferret out the relevant factors and to discern the subtle dynamics of the litigation warranting approval of a negotiated settlement. See id. (stating that "the decision to approve or reject a settlement is committed to the sound discretion of the trial judge because he is `exposed to the litigants, and their strategies, positions and proof'").

Nevertheless, the district court has an important and meaningful role to play in the settlement of a putative class action. In particular, the district judge has a fiduciary duty to safeguard the interests of the absent and putative class members. See, e.g., Sullivan v. DB Invs., Inc., 667 F.3d 273, 319 (3d Cir.2011) (stating that "`trial judges bear the important responsibility of protecting absent class members,' and must be `assur[ed] that the settlement represents adequate compensation for the release of the class claims'"); Maywalt v. Parker & Parsley Petroleum Co., 67 F.3d 1072, 1078 (2d Cir.1995) (noting that "the district court has a fiduciary responsibility to ensure that the settlement is fair and not a product of collusion, and that the class members' interests were represented adequately") (internal quotation marks omitted). That duty is especially important when the interests of the class and its counsel negotiating on its behalf are not aligned. See Reynolds v. Beneficial Nat'l Bank, 288 F.3d 277, 279-80 (7th Cir.2002) (stating that the problem that class counsel "may, in derogation of their professional and fiduciary obligations, place their pecuniary self-interest ahead of that of the class ... requires district judges to exercise the highest degree of vigilance in scrutinizing proposed settlements of class actions"). Moreover, where the settlement agreement is negotiated prior to final class certification, "There is an even greater potential for a breach of fiduciary duty owed the class during settlement." In re Bluetooth Headset Products Liability Litigation, 654 F.3d 935, 946 (9th Cir.2011). Thus, in this case, an "even higher level of scrutiny than that ordinarily required under Rule 23(e)" is warranted. Id.

The terms and structure of the proposed settlement herein warrant meaningful scrutiny. Its approval must be supported by the record. Unfortunately, though the district judge plainly took this duty seriously, I believe the requisite scrutiny was not possible given the deficiencies in the record. The district court's task was compounded by the lack of clarity in this Circuit's law on particular issues discussed below.

I.

Although the majority addresses the reasonableness of the fee award before addressing the reasonableness of the settlement's substantive terms, review ordinarily begins with the adequacy of the settlement itself. See Bluetooth, 654 F.3d at 941. Therefore, I address the adequacy of the settlement first.

Under Federal Rule of Civil Procedure 23(e), a class action may be settled only with the approval of the district court, and "[a] district court's approval ... must be accompanied by a finding that the settlement is `fair, reasonable, and adequate.'" Lane, 696 F.3d at 818. Here, a higher standard of fairness is required because the settlement took place before formal class certification. See id. at 819.

A district court must consider a number of factors in determining whether a proposed settlement is fair and adequate; these include:

Id. See Churchill Vill., L.L.C. v. Gen. Elec., 361 F.3d 566, 575 (9th Cir.2004).

In the case at bar, the district court adequately considered many of the Lane/Churchill factors, including the risk of further litigation. It properly found, for example, that there was a risk class claims could have been forced into individual arbitration pursuant to AT & T Mobility v. Concepcion, ___ U.S. ___, 131 S.Ct. 1740, 179 L.Ed.2d 742 (2011). There was also a risk in obtaining class certification in light of Wal-Mart Stores, Inc. v. Dukes, ___ U.S. ___, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011). Many of the other factors supported approval.

However, in my view, the district court's order — although thorough in most respects — did not adequately assess a critical factor: the amount offered in settlement. Although Lane does not expressly order the importance of the assessment factors, inescapably, the core of any settlement is "the amount offered in settlement." In this regard, the district court's order did not determine whether the most significant terms of the settlement were real or largely illusory. See In re Dry Max Pampers Litig., 724 F.3d 713, 721 (6th Cir.2013) (stating that "[t]he relief that this settlement provides to unnamed class members is illusory" and therefore finding the settlement unfair); cf. Dennis v. Kellogg Co., 697 F.3d 858, 868 (9th Cir.2012) (stating that dollar value of the cy pres fund should not be fictitious).

Under the settlement, there were two classes that would receive benefits: (1) current franchisees and (2) former franchisees. Current franchisees obtained certain equitable relief — most notably, the assignment of customer accounts which would confer upon the franchisees economic value that could not summarily be taken away. As for former franchisees, each was entitled to $475 in cash and a credit of $750 (in effect, a coupon) which could be used to purchase a new Coverall franchise. For the reasons discussed below, both aspects of the settlement are problematic, and a fuller record is required in order to properly assess the fairness, reasonableness, and adequacy of the settlement.

A.

Current Franchisees: Assignment of Customer Accounts

For current franchisees, the primary benefit of the settlement was the assignment of customer accounts. The centrality of this benefit is clear: a major claim asserted by Plaintiffs in this action was a breach of contract based on Coverall's "churning" of customers accounts. Churning involves taking a customer account away from a franchise owner for a pretextual reason in order to resell the customer account to a different franchise owner. If accounts were assigned to franchise owners, that would prevent the practice of churning.

The importance of the assignment provision is underscored by the district court's order as well as the parties' briefs presented to this Court. The district court stated in its order that "[t]he [Settlement] Agreement provides separate benefits for current and former franchisees. Principally, Coverall pledges to assign customer accounts to current franchisees." ER 23 (emphasis added); see also ER 25 (stating that "assignment of customer accounts and pledges for programmatic changes are significant victories"1). Similarly, in Plaintiffs' appellate brief, the first benefit highlighted is the assignment of accounts, with Plaintiffs touting that "[t]he value of the assignment ... is estimated to be worth $18 to $20 million per year in California alone."2 Pls.' Resp. Br. at 6. Defendants did the same in their own appellate brief. See Defs.' Resp. Br. at 36-40.

Because the assignment was the primary benefit of the settlement for current franchisees, the district court had a duty to ensure that this benefit had real value. Notably, the burden of proving that the assignment had real value lay with the settling parties; it was not Mr. Singh's burden to disprove it. See In re Dry Max Pampers Litig., 724 F.3d at 719 ("[T]o the extent the parties here argue that the settlement was fair because the refund program has actual value for consumers, it was the parties' burden to prove the fact, rather than [the objector's] burden to disprove it.").

As to whether the assignment provision of the settlement would actually benefit current franchisees, that posed a troubling question because, under the proposed settlement, the assignment is conditional. See ER 73. That is, until a current or new franchise owner pays for a customer account in full (i.e., the full franchise fee, including any financed portion), there is no actual assignment, and customer accounts can continue to be taken away pursuant to the old just or good cause standard.3 But the record contains no information as to how many or what percentage of franchise owners have paid their fees in full. In particular, there is no information about how many franchise owners still owe money on their franchise fees (and if so, how much); how long it takes an average franchise owner to pay off a franchise fee; and whether franchise owners typically leave before they are able to pay off their franchise fees in full.4 This is significant because Mr. Singh, the objector, asserted in argument before this Court that a substantial number of franchisees finance their franchise fees. See also Pls.' Op. Br. at 10 n.3 (claiming that, "[t]ypically, Coverall workers make a down payment toward [the franchise] fee [which ranges from approximately $10,000 to $32,000] and then finance the remaining portion of the fee by borrowing it from Coverall for a two or three-year period (at 12% interest)"; adding that, "[w]hen workers lose accounts and want to obtain more work, they may do so by then paying fees for `additional business,'" and these fees must also be paid off before any assignment). On the current record, we do not know whether 5%, 50%, or 95% of the class will or will likely receive the benefit of the assignment of accounts.

Moreover, even if a significant portion of current franchisees were actually able to pay off their franchise fees and thus get an assignment, the parties' claim that the value of the assignment was $18-20 million is still problematic. The parties asserted this value because the gross billing of customer contracts in California in the year 2010 was $18-20 million. See Pls.' Resp. Br. at 6. But under the settlement agreement, certain customer accounts are deemed not assignable. See ER 73 (providing that "certain accounts, including National Accounts, customer prepared contracts that preclude assignment, and local multiple location customer contracts are not assignable"). There does not appear to be any information in the record as to how many customer contracts fall within this category, or their worth. Thus, the claimed $18-20 million value of the assignment may well be inflated. The district court's order did not address this issue.

Furthermore, although the district court acknowledged that the assignment of accounts may not have the full $18-20 million value claimed by the parties because of the conditional nature of the assignment, see ER 30 ("Indeed, it is unclear whether the assignment of all accounts to franchisees will reach the $18-20 million Plaintiffs claim, especially because assignments are only conditional until franchises are paid for in full."), it did not estimate — indeed, had no basis in the record to estimate — what portion of the class would actually benefit from the assignment. The district court could only state that, "once franchisees are assigned, franchisees will own a value business they can choose to sell or continue to operate." ER 30.5 The majority correctly states that the district court ordinarily need not necessarily place a specific monetary value on injunctive relief. See, e.g., Staton v. Boeing Co., 327 F.3d 938, 959 (9th Cir.2003) (stating that courts cannot "judge with confidence the value of the terms of a settlement agreement [in which] the settlement provides for injunctive relief"). Still, given the importance of the "amount offered in settlement" in assessing the fairness and adequacy of a class settlement, valuation of benefits should be "examined with great care." Dennis, 697 F.3d at 868 (finding the settlement valuation "unacceptably vague and possibly misleading" and noting that "[t]he issue of the valuation ... must be examined with great care to eliminate the possibility that it serves only the `self-interests' of the attorneys and the parties, and not the class, by assigning a dollar number to the fund that is fictitious"). In any event, while it is true that equitable relief sometimes is not capable of quantitative valuation (for example, where equitable relief is directed to dignitary interests such as privacy rights), where that relief has some calculable monetary value and is central to the relief obtained, some meaningful effort should be made to determine the "amount offered in settlement." Cf. Fed.R.Civ.P. 23(h), 2003 advisory committee notes (stating that "[s]ettlements involving nonmonetary provisions for class members also deserve careful scrutiny to ensure that these provisions have actual value to the class").

In the case at bar, the assignment of accounts is capable of quantitative valuation, even if that value cannot be precisely determined. See, e.g., Staton, 327 F.3d at 978 (noting that value of benefits from injunctive relief may be included in common fund calculation where value can be accurately ascertained — e.g., in a prior case, there was clearly a measurable benefit of one replacement latch for each minivan owned, and thus the court could, "with some degree of accuracy, value the benefits conferred").6 Indeed, as noted above, the parties themselves valued the assignment at approximately $18-20 million.7 See Pls.' Resp. Br. at 6; see also ER 74 (providing that "Coverall represents that, in 2010, the gross billing of the customer contracts in the State of California was approximately $20 million"). The factor that is missing from the calculus (and missing from the record) is the number or percentage of current franchisees who actually will or, at least, will likely receive the benefit.

Absent anything in the record which would shed light on this crucial variable, the district court was not in a position to conclude that the settlement was fair, reasonable, and adequate. Although "[w]e review the district court's approval of a class-action settlement for abuse of discretion," Radcliffe v. Experian Info. Solns., 715 F.3d 1157, 1162 (9th Cir.2013), and, under such review, "[w]e are not permitted to `substitute our notions of fairness for those of the district court and the parties to the agreement,'" Class Plaintiffs v. Seattle, 955 F.2d 1268, 1276 (9th Cir.1992), a district court's approval cannot be affirmed where its "findings of fact were ... without support in the record."8 Radcliffe, 715 F.3d at 1162. Here, the record was not sufficiently developed on what should be a central factor in assessing the fairness, reasonableness, and adequacy of the class settlement — "the amount offered in settlement." Lane, 696 F.3d at 819.

B.

Former Franchisees: Monetary Relief

With respect to the monetary portion of the settlement, former franchisees are each entitled to $475 in cash and a credit of $750 towards a purchase of a new Coverall franchise. There were approximately 750 former franchisees in the class. Therefore, Defendants would, in theory, pay a maximum of $918,750 to former franchises.

The monetary terms of the settlement, however, are suspect. Although the $918,750 class fund was identified as a potential pay-out, neither side expected Defendants would actually pay anywhere close to this sum under the structure of the proposed settlement. This is because, under the settlement, (1) pay-outs would be made only to those class members who submitted claims, and (2) any unclaimed funds would revert back to Defendants rather than being redistributed to claiming class members or distributed to a cy pres beneficiary.

As district courts have pointed out, "in a reversionary common fund or a claims-made settlement, the defendant is likely to bear only a fraction of the liability to which it agrees." In re TJX Cos. Retail Sec. Breach Litig., 584 F.Supp.2d 395, 405 (D.Mass.2008) (emphasis added); see also Sylvester v. CIGNA Corp., 369 F.Supp.2d 34, 52 (D.Me.2005) (stating that parties' evidence indicated that "`claims made' settlements regularly yield response rates of 10 percent or less"; adding that "the parties' expectations of a low response rate gives the reverter clause real value for Defendants"). Researchers and commentators have also confirmed that, while results vary from case to case depending on a number of factors, it is common that only a fraction of the class in class action settlements actually submit claims. See, e.g., Max Helveston, Promoting Justice Through Public Interest Advocacy in Class Actions, 60 Buff. L.Rev. 749, 782-83 (2012) (noting that there are "a variety of different explanations" for low claims rates — e.g., "the difficulty of notifying class members, the overly complex and technical notifications that class members receive, the effort required to pursue a claim, the lack of interest of class members in the types of relief available, and the failure of fund designers to design claim procedures in ways that take into account cognitive biases" — but adding that, at the end of the day, "class members in large class action suits typically file claims at rates that are much lower than one would expect"); Mayer Brown LLP, Do Class Actions Benefit Class Members?, available at http://www.mayerbrown.com/files/uploads/Documents/PDFs/2013/December/DoClassActionsBenefitClassMembers.pdf (last visited Apr. 30, 2014) (stating that "extremely small claim-filing rates [in several cases] are consistent with the few other reports of claim rates in class action settlements that have come to light").

That is precisely, and predictably, what happened here. With respect to the $475 cash payment, only 119 out of 750 former franchisees actually submitted claims. This amounts to a response rate of about 16%, for a total of $56,525 out of $356,250. As for the $750 coupon, of the 750 former franchisees, only 34 made a claim for the coupon — a response rate of about 4.5%. Thus, instead of a true pay-out of $562,500 (with respect to coupons), Defendants would be obligated to pay (or credit) at most $25,500, and this assumes that all 34 former franchisees will actually use the coupon once received.

At the end of the day, the purported settlement value of $918,750 for former franchisees (covering both the $475 pay-out and the $750 coupon) amounted to, at most, a true value of only $82,025 (or a response rate of approximately 9%) — and likely even less because some of the coupons claimed probably will not be used. The unclaimed funds did not benefit the class as they would revert back to Defendants.

This result was entirely predictable once the parties stipulated to a claims process. This then begs the question why the parties chose to require a claims process in the first place. While there are situations in which a claims process is unavoidable (such as a consumer class action where there is no readily accessible record of purchases to identify class members and their contact information), in this case, it is not obvious why a claims process was necessary. Compare In re Wells Fargo Loan Processor Overtime Pay Litig., No. C-07-1841 EMC, 2011 WL 3352460, at *8, 2011 U.S. Dist. LEXIS 84541, at *22-23 (N.D.Cal. Aug. 2, 2011) (explaining why "there is a valid and substantial justification for the claims process"). As noted in a Federal Judicial Center publication, a claims process is not necessary in all cases:

De Leon v. Bank of Am., N.A., No. 6:09-cv-1251-Orl-28KRS, 2012 WL 2568142, at *20, 2012 U.S. Dist. LEXIS 91124, at *61 (M.D.Fla. Apr. 20, 2012) (quoting the 2010 version of the "Judges' Class Action Notice and Claims Process Checklist and Plain Language Guide" produced by the Federal Judicial Center).

Here, the necessity of a claims process is not apparent from the record. There was a set sum to be paid to each former franchisee, so a claims process was not needed in order for the parties to determine the appropriate amount of damages for each former franchisee. No proof of claim was needed to identify class members because Defendants already had within their possession information identifying the former franchisees. And presumably Defendants had the last best address for each former franchisee: how else could they send the claim forms to the franchisees? The record fails to indicate why it was not feasible to send checks and coupons directly to class members rather than requiring a claims process. See De Leon, 2012 WL 2568142, at *19, 2012 U.S. Dist. LEXIS 91124, at *61 (finding a claim form procedure unreasonable because, inter alia, defendant had databases that "enabled it to identify approximately 500,000 cardholders who were assessed late fees on timely payments made"); see also Mayer Brown LLP, Do Class Actions Benefit Class Members? (noting that automatic distribution settlements are possible where "class members whose eligibility and alleged damages could be ascertained and calculated — such as retirement-plan participants in ERISA class actions").

The decision to use coupons (or credit) as a benefit for former franchisees towards the price of a new franchise also raises questions about the adequacy of the settlement. As noted in the Manual for Complex Litigation, coupons are often illusory monetary benefits. See Manual for Complex Litig., 4th § 21.61, at 310 (noting that a judge "should be wary" of a proposed settlement that "grant[s] class members illusory nonmonetary benefits, such as discount coupons for more of defendants' products, while granting substantial monetary attorney fee awards"). The illusory value of the coupon is particularly acute here: each former franchisee was entitled to a $750 coupon to be used towards purchase for a new Coverall franchise, but a franchise costs at least $10,000 and can be as high as $32,000. See ER 138. It was unlikely that many of the $750 coupons would ever be claimed or used. It is not surprising that only 4.5% of former franchisees filed a claim for the coupon.

Finally, even if the claims process were justified, that does nothing to justify the reverter of any unclaimed funds to Defendants. The parties could have agreed (as many settlements have) to a second distribution of residual funds to those former franchisees who did submit claims and/or a distribution to a cy pres beneficiary. See, e.g., Garner v. State Farm Mut. Auto. Ins. Co., No. CV 08 1365 CW (EMC), 2010 WL 1687832, at *18, 2010 U.S. Dist. LEXIS 49477, at *49 (N.D.Cal. Apr. 22, 2010) (noting that, under the settlement, uncashed settlement checks and unspent portions of the administrative cost reserve will be redistributed to the class or given to a cy pres beneficiary). At least this would have guaranteed a meaningful and substantial payment to the class.

In sum, the monetary benefits to the former franchisees was vastly overstated and largely illusory, thus raising significant questions that deserve a meaningful degree of scrutiny by the district court. Cf. In re Classmates.com Consolidated Litig., No. C 09-45 RAJ, 2011 WL 744664, at *8, 2011 U.S. Dist. LEXIS 17761, at *24 (W.D.Wash. Feb. 23, 2011) (stating that, "where class counsel points to an illusory $9.5 million benefit as justification for its own fee award, without acknowledging that counsel expected the benefit to be dramatically smaller, it illustrates the danger of deferring to counsel's view of a settlement that misaligns the interests of class members and class counsel"). This is particularly troubling given that Plaintiffs' counsel was guaranteed (subject only to court approval where Defendants agreed not to oppose the motion) nearly $1 million regardless of the amount of money actually claimed by the class.

The requirement of a claims process (without a substantial justification) combined with a reversion of unclaimed class funds back to the defendant was not a factor presented or expressly discussed in Bluetooth. This Court in Bluetooth identified three "subtle signs that class counsel have allowed pursuit of their own self-interests and that of certain class members to infect the negotiations" — including an arrangement for attorney fees not awarded to revert to the defendant rather than to the class fund. Bluetooth, 654 F.3d at 947. But for the same reason why a reverter of unawarded attorney fees is a risk factor indicative of potential collusion, the use of a claims process coupled by a reversion of unclaimed class funds likewise presents such a risk. Indeed, where the magnitude of the reverter of class funds is great, it should raise even more serious questions than reverter of unawarded fees.

II.

Because the reasonableness of the settlement's substantive terms could not be adequately determined based on the record before the district court, the reasonableness of the fee award also could not be appropriately assessed.

As noted above, where a case is settled prior to formal class certification, a higher standard of fairness is required. This higher standard is necessary precisely because, before formal class certification, "there is an even greater potential for breach of fiduciary duty owed the class" — i.e., there is an even greater chance of collusion or conflicts of interest on the part of the named plaintiffs and their attorneys. Bluetooth, 654 F.3d at 946. Moreover, it is important to take into account that,

In re Dry Max Pampers Litig., 724 F.3d at 717 (citations omitted).

In Bluetooth, this Court underscored that "[c]ollusion may not always be evident on the face of a settlement, and courts therefore must be particularly vigilant not only for explicit collusion, but also for more subtle signs that class counsel have allowed pursuit of their own self-interests and that of certain class members to infect the negotiations." Bluetooth, 654 F.3d at 947. As indicated above, Bluetooth identified three such subtle signs:

Bluetooth, 654 F.3d at 947.

In the case at bar, the second and third factors above are clearly applicable, as the majority recognizes. This puts a premium on the district court's evaluation of the first Bluetooth factor. However, the district court could not evaluate whether the $1 million fee award constituted a "disproportionate" distribution of the settlement without first having a fair sense of the value of the overall settlement, particularly the settlement's key terms discussed above.

Furthermore, although not necessarily a subtle sign of collusion in and of itself, the fact that the fee award here was not tied to the magnitude of relief actually obtained by the class amplifies the need for scrutiny. Here, class counsel had no particular incentive to negotiate a settlement that would ensure substantial benefits would actually accrue to the class. Counsel's fee was a stipulated sum unaffected by the actual pay-out to the class. To the extent the parties based the fee award in part upon a claimed settlement fund of over $900,000 and the promised assignment of accounts, that basis is misleading given that the class fund was largely illusory and it was uncertain whether a substantial portion of the class will benefit from the assignment. Although this Circuit has sanctioned using the full class fund theoretically available as a basis for evaluating the reasonableness of a negotiated fee award, see Williams v. MGM-Pathe Communs. Co., 129 F.3d 1026, 1027 (9th Cir.1997) (holding that the lower court had "abused its discretion by basing the fee on the class members' claims against the fund rather than on a percentage of the entire fund or on the lodestar"), such a practice tends to divorce the class counsel's incentives from the best interests of the class. It is perhaps for this reason the Fifth Circuit takes a different approach. See Strong v. BellSouth Telecomms., Inc., 137 F.3d 844, 852-53 (5th Cir.1998) (indicating that fee awards may be based on actual pay-out to class rather than reversionary fund). Where fees are tied to actual pay-out to the class, counsel has a strong incentive to negotiate a real, rather than illusory, class fund. Under this Circuit's decision in Williams, however, that incentive may be absent, and thus there is a need for meaningful scrutiny of the settlement and fee award.

That scrutiny is particularly warranted here where, as noted above, two of the three Bluetooth factors were clearly met (i.e., the clear sailing and reverter arrangements). Without a full record as to the value of the settlement, the district court could not ensure that the fee award was not disproportionate compared to the benefits to the class as required by the third Bluetooth factor.

For example, the district court's lodestar analysis was problematic without a more fully developed record. In the settlement, the parties had agreed to a fee award of approximately $1 million. According to the district court, this was a reasonable award because class counsel had incurred approximately $3 million in fees in the course of litigating the action. But even accepting that counsel had legitimately incurred $3 million in fees (based on hours spent and rates charged), that does not reflect a full lodestar analysis. "[T]he fairness of the lodestar amount should be gauged against the overall class recovery, adjusting the lodestar fees upward or downward as necessary to ensure their reasonableness. In doing so, the district court must weigh the requested lodestar figure against a variety of factors, foremost among them the results obtained for the class, monetary and non-monetary alike." In re HP Inkjet Printer Litig., 716 F.3d 1173, 1190 (9th Cir.2013). Here, even if we assume that the district court did an adjustment based on results obtained (i.e., a downward adjustment from $3 million to $1 million), the district court did not have a complete basis for making the adjustment without a fuller understanding of the value of the assignment of accounts — the central benefit of the settlement.

The district court also performed a percentage method analysis as a cross-check on the fee award. According to the district court, "even if the cash settlement [for former franchisees], the assignment of accounts [for current franchisees], and the pledged programmatic changes [for current franchisees] ... were worth only about $4 million, the requested fee award [of $1 million] would fit with[in] normal bounds of class action recovery." ER 31. Presumably, the district court's statement was informed by case law from this Court using 25% of the common fund as a benchmark under the percentage method. See Bluetooth, 654 F.3d at 942 (stating that, under the percentage method, "courts typically calculate 25% of the fund as the `benchmark' for a reasonable fee award"); cf. Staton, 327 F.3d at 945-46 (noting that injunctive "relief should generally be excluded from the value of a common fund when calculating the appropriate attorneys' fee award, as the benefit of that relief to the class members is most often not sufficiently measurable," but "[t]he fact that counsel obtained injunctive relief in addition to monetary relief for their clients is ... a relevant circumstance to consider in determining what percentage of the fund is reasonable as fees"). However, there is nothing in the record upon which the district court could have based an assumption that the benefits of the settlement could be worth $4 million. The value of the assignment to the class was on this record entirely unknown since there is no information as to the number or percentage of franchise owners who are actually able or likely to receive the assignments of accounts. While the district court is entitled to make estimates in evaluating the reasonableness of the fee award, its assumptions must have some basis in the record. Cf. Radcliffe, 715 F.3d at 1162 (stating that a district court's approval of a class action settlement cannot be affirmed where its "findings of fact were ... without support in the record"). In this case, that basis is absent. Cf. Bluetooth, 654 F.3d at 947 (district court must guard against risk that a high fee award was traded for "lower monetary payments to class members or less injunctive relief for the class than could otherwise have been obtained").

III.

On the record before this Court, I cannot conclude that the proposed settlement was fair, reasonable, and adequate and the fee award reasonable. Crucial information about the value of the benefits obtained for the class (i.e., how many franchisees will benefit from the assignment of accounts) is missing. Also lacking is the justification for requiring former franchisees to submit claims and providing for a reverter of undistributed funds back to Coverall — while guaranteeing a $1 million fee award to counsel.9

The district court's task in cases such as this could be made more manageable were this Court to provide more specific guidance on: (1) the centrality of the Lane/Churchill factor "the amount offered in settlement" in assessing the adequacy of class action settlements, and (2) the explicit inclusion as a factor in Bluetooth the unjustified use of a claims process and/or reverter of unclaimed class funds back to the defendant — particularly if Williams continues to allow fees to be based on size of the potential class fund rather than actual pay-out to the class. This case affords the Court an opportunity to provide that clarity.

I do not express any opinion whether ultimately the substantive terms of the settlement or the fee award may in fact be fair, reasonable and adequate, as the majority concludes. Rather, I diverge from the majority because I believe there are substantial questions about the fairness, reasonableness, and adequacy of the proposed settlement that cannot be answered without a fuller record and more complete analysis.

For the foregoing reasons, I respectfully dissent.