OPINION

TROTT, Circuit Judge:

Most cases in our judicial system never make it to trial. Litigants often find it advantageous to secure a resolution more quickly by settling the case and negotiating a result the parties can tolerate, even though neither side can call it a total win. Normally, that is the end of the story, and the parties walk away — not entirely happy, but not entirely unhappy either.

In a class action, however, any settlement must be approved by the court to ensure that class counsel and the named plaintiffs do not place their own interests above those of the absent class members. In this false advertising case, we confront a class action settlement, negotiated prior to class certification, that includes cy pres distributions of money and food to unidentified charities. It also includes $2 million in attorneys' fees — which breaks down to a $2,100 hourly rate — while offering class members a sum of (at most) $15.

After carefully reviewing the class settlement, we conclude that it must be set aside for two reasons. First, the district court did not apply the correct legal standards governing cy pres distributions and thus abused its discretion in approving the settlement. The settlement neither identifies the ultimate recipients of the cy pres awards nor sets forth any limiting restriction on those recipients, other than characterizing them as charities that feed the indigent. To the extent that we can meaningfully review such distributions where the parties fail to identify the recipients, we hold that the cy pres portions of the settlement are not sufficiently related to the plaintiff class or to the class's underlying false advertising claims. Second, even if the cy pres distributions did comply with our cy pres standards, the settlement would still fail because the negotiated attorneys' fees are excessive. We therefore reverse the district court's approval of the settlement, vacate the judgment, and remand for further proceedings consistent with this opinion.

I

BACKGROUND

In January 2008, Kellogg Co., the maker of Frosted Mini-Wheats cereal, began a marketing campaign that claimed the cereal was scientifically proven to improve children's cognitive functions for several hours after breakfast. Obviously aimed at parents of school-age children, Kellogg's advertisements allegedly included the following statements:

According to a declaration submitted by lead counsel for the plaintiff class, counsel began investigating these marketing claims and, in April and May 2009, drafted a class action complaint on behalf of Ohio resident Jon Koz, alleging violations of Ohio consumer protection laws. Around the same time, another law firm was investigating the same marketing claims on behalf of California resident Harry Dennis. Although Mr. Koz never filed his Ohio complaint, Mr. Dennis filed suit in August 2009 against Kellogg in the United States District Court for the Southern District of California, alleging violations of that state's Unfair Competition Law (UCL) and asserting a claim of unjust enrichment.

Sometime prior to January 2010, counsel for Koz and counsel for Dennis discovered they were involved in similar activities and decided to join forces. Because informal settlement attempts were unsuccessful, counsel for the consumers and for Kellogg participated in a day-long mediation session with Martin Quinn of JAMS, a well-established alternative dispute resolution firm. As a result of this mediation session and numerous other settlement discussions, the parties agreed, in principle, to settle the case.

Meanwhile, the Dennis lawsuit had been gathering dust. On June 22, 2010, the district court notified the parties of its intent to dismiss the case for lack of prosecution. Koz and Dennis immediately filed a joint amended class action complaint.

In their amended complaint, the named plaintiffs asserted that Kellogg's marketing claims regarding the effect of Frosted Mini-Wheats on children's attentiveness were false, that the study upon which these results were based did not support the company's claims, and that the study was not scientifically valid. The plaintiffs asserted unjust enrichment, claims under the UCL and California's Consumer Legal Remedies Act (CLRA), and claims under "similar laws of other states."

Over the next three months, the parties continued to work out the details of their settlement. Ultimately, they agreed to settle the case on the following terms:

Together with notice and administrative costs approximated at $391,500, the parties value the settlement at $10,641,500.

On the plaintiffs' motion, the district court certified the class, granted preliminary approval of the settlement, and approved the proposed class notice. Because Kellogg sells its products to wholesalers, not directly to consumers, there was no way to identify each member of the class. Therefore, the class notice was published in Parents magazine and other "targeted sources based on market research about consumers who purchased the products," including 375 websites.

Two class members objected to the settlement: Stephanie Berg and Omar Rivero (Objectors). As relevant to this appeal, the Objectors argued that the settlement's use of cy pres relief was improper because "the only relationship between this lawsuit and feeding the indigent is that they both involve food in some way." They argued also that the cy pres distributions would benefit class counsel and Kellogg, but not the class members, because class members "have no idea how their funds might be used or in whose hands their monies will end up." Finally, the Objectors argued that the attorneys' fees — which represented approximately 19% of a common fund allegedly worth over $10.64 million — were excessive. The district court approved the class settlement and dismissed the case with prejudice. In doing so, however, the court did not address the Objectors' argument that the cy pres distributions were too remote from the class members and were not sufficiently related to their UCL and CLRA claims. The court also approved the requested attorneys' fees, stating that the fees were

II

STANDARD OF REVIEW

The settlement of a class action must be fair, adequate, and reasonable. Fed.R.Civ.P. 23(e)(2). "We review a district court's approval of a proposed class action settlement, including a proposed cy pres settlement distribution, for abuse of discretion. A court abuses its discretion when it fails to apply the correct legal standard or bases its decision on unreasonable findings of fact." Nachshin v. AOL, LLC, 663 F.3d 1034, 1038 (9th Cir.2011) (internal citations omitted).

Appellate review of a settlement agreement is generally "extremely limited." Hanlon v. Chrysler Corp., 150 F.3d 1011, 1026 (9th Cir.1998). But where, as here, class counsel negotiates a settlement agreement before the class is even certified, courts "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." In re Bluetooth Headset Prods. Liab. Litig., 654 F.3d 935, 947 (9th Cir.2011). In such a case, settlement approval "requires a higher standard of fairness" and "a more probing inquiry than may normally be required under Rule 23(e)." Hanlon, 150 F.3d at 1026. "To survive appellate review, the district court must show it has explored comprehensively all factors," id., and must give "a reasoned response" to all non-frivolous objections, Officers for Justice v. Civil Serv. Comm'n, 688 F.2d 615, 624 (9th Cir.1982).

III

DISCUSSION

A

The Cy Pres Distributions of Food and Unclaimed Funds

Although the cy pres doctrine originated in the area of wills as a way to effectuate the testator's intent in making charitable gifts, federal courts now frequently apply it in the settlement of class actions "`where the proof of individual claims would be burdensome or distribution of damages costly.'" Nachshin, 663 F.3d at 1038 (quoting Six Mexican Workers v. Ariz. Citrus Growers, 904 F.2d 1301, 1305 (9th Cir.1990)). Used in lieu of direct distribution of damages to silent class members, this alternative allows for "aggregate calculation of damages, the use of summary claim procedures, and distribution of unclaimed funds to indirectly benefit the entire class." Six Mexican Workers, 904 F.2d at 1305. To ensure that the settlement retains some connection to the plaintiff class and the underlying claims, however, a cy pres award must qualify as "the next best distribution" to giving the funds directly to class members. Id. at 1308(internal quotation marks omitted).

Not just any worthy recipient can qualify as an appropriate cy pres beneficiary. To avoid the "many nascent dangers to the fairness of the distribution process," we require that there be "a driving nexus between the plaintiff class and the cy pres beneficiaries." Nachshin, 663 F.3d at 1038. A cy pres award must be "guided by (1) the objectives of the underlying statute(s) and (2) the interests of the silent class members," id. at 1039, and must not benefit a group "too remote from the plaintiff class," Six Mexican Workers, 904 F.2d at 1308. Thus, in addition to asking "whether the class settlement, taken as a whole, is fair, reasonable, and adequate to all concerned," we must also determine "whether the distribution of the approved class settlement complies with our standards governing cy pres awards." Nachshin, 663 F.3d at 1040 (internal quotation marks omitted).

A review of our relevant precedent reveals that the settlement here fails to satisfy those standards. In Six Mexican Workers v. Arizona Citrus Growers, a class of undocumented Mexican farm workers sued various companies for violations of the Farm Labor Contractor Registration Act. 904 F.2d at 1303. After a bench trial, the district court found the defendants liable for over $1.8 million, which we later reduced to $850,000, in statutory damages. Id. at 1303-04, 1310. The district court identified the Inter-American Fund, which provided humanitarian aid in Mexico, as the cy pres recipient of any unclaimed funds. Id. at 1304.

We held that the cy pres distribution was an abuse of discretion because there was "no reasonable certainty" that any class member would benefit from it, even though the money would go "to areas where the class members may live." Id. at 1308. The choice of charity and its relation to the class members and class claims — or lack thereof — figured heavily in our analysis. The purpose of the statute was to compensate victims of unscrupulous employers and to deter future violations, but the Inter-American Fund was "not an organization with a substantial record of service nor is it limited in its choice of projects," and any distribution would therefore have required court supervision "to ensure that the funds are distributed in accordance with the goals of the remedy." Id. at 1309. Because "the district court's application[of the cy pres doctrine] was inadequate to serve the goals of the statute and protect the interests of the silent class members," we reversed the cy pres distribution. Id. at 1312.

We recently came to a similar conclusion in Nachshin v. AOL, LLC. In that case, AOL was accused of violating a number of statutes, including the UCL and the CLRA, by wrongfully inserting commercial footers into the plaintiffs' outgoing emails. 663 F.3d at 1036. Because damages would be small and distribution to the class prohibitively expensive, AOL agreed, as part of a class settlement, to make substantial donations to three charities: the Legal Aid Foundation of Los Angeles, the Federal Judicial Center Foundation, and the Los Angeles and Santa Monica chapters of the Boys and Girls Club of America. Id. at 1037.

We held that the cy pres distribution "fail[ed] to target the plaintiff class, because it d[id] not account for the broad geographic distribution of the class." Id. at 1040. The class included over 66 million AOL users across the country, but two-thirds of the donations were slated for Los Angeles charities. Further, although the donation to the Federal Judicial Center Foundation "at least conceivably benefit[ed] a national organization," the Foundation "ha[d] no apparent relation to the objectives of the underlying statutes, and it [wa]s not clear how this organization would benefit the plaintiff class." Id. We noted, however, that it would not be difficult for the parties to come up with an appropriate charity if they wished to do so:

Id. at 1041. In approving the cy pres distribution to charities that had no relation to the class or to the underlying claims, the district court "applied the incorrect legal standard" and abused its discretion. Id. at 1040.

The cy pres award in the settlement here is likewise divorced from the concerns embodied in consumer protection laws such as the UCL and the CLRA. As California courts have stated, "[t]he UCL is designed to preserve fair competition among business competitors and protect the public from nefarious and unscrupulous business practices," Wells v. One2One Learning Found., 116 Cal.App.4th 515, 10 Cal.Rptr.3d 456, 463-64 (2004), rev'd in part on other grounds, 39 Cal.4th 1164, 48 Cal.Rptr.3d 108, 141 P.3d 225 (2006), and the purpose of the CLRA is similarly "to protect consumers against unfair and deceptive business practices," Cal. Civ.Code § 1760. Although there is no way to identify the cy pres beneficiaries from this record, we do know that according to the settlement, any charity to receive a portion of the cy pres distributions will be one that feeds the indigent. This noble goal, however, has "little or nothing to do with the purposes of the underlying lawsuit or the class of plaintiffs involved." Nachshin, 663 F.3d at 1039.

At oral argument, Kellogg's counsel frequently asserted that donating food to charities who feed the indigent relates to the underlying class claims because this case is about "the nutritional value of food." With respect, that is simply not true, and saying it repeatedly does not make it so. The complaint nowhere alleged that the cereal was unhealthy or lacked nutritional value. And no law allows a consumer to sue a company for selling cereal that does not improve attentiveness. The gravamen of this lawsuit is that Kellogg advertised that its cereal did improve attentiveness. Those alleged misrepresentations are what provided the plaintiffs with a cause of action under the UCL and the CLRA, not the nutritional value of Frosted Mini-Wheats. Thus, appropriate cy pres recipients are not charities that feed the needy, but organizations dedicated to protecting consumers from, or redressing injuries caused by, false advertising.

Our concerns are not placated by the settlement provision that the charities will be identified at a later date and approved by the court — a decision from which the Objectors might again appeal. Our standards of review governing precertification settlement agreements require that we carefully review the entire settlement, paying special attention to "terms of the agreement contain[ing] convincing indications that the incentives favoring pursuit of self-interest rather than the class's interests in fact influenced the outcome of the negotiations." Staton v. Boeing Co., 327 F.3d 938, 960(9th Cir.2003). Cy pres distributions present a particular danger in this regard. "When selection of cy pres beneficiaries is not tethered to the nature of the lawsuit and the interests of the silent class members, the selection process may answer to the whims and self interests of the parties, their counsel, or the court." Nachshin, 663 F.3d at 1039. This record leaves open the distinct possibility that the asserted $5.5 million value of the cy pres award will only be of serendipitous value to the class purportedly protected by the settlement. The difficulty here is that, by failing to identify the cy pres recipients, the parties have restricted our ability to undertake the searching inquiry that our precedent requires. The cy pres problem presented in this case is of the parties' own making, and encouraging multiple costly appeals by punting down the line our review of the settlement agreement is no solution.

For the foregoing reasons, we conclude that the district court did not apply the correct legal standards for cy pres distributions as set forth in Six Mexican Workers and Nachshin. Therefore, the approval of the settlement was an abuse of discretion.

We do not have the authority to strike down only the cy pres portions of the settlement. "It is the settlement taken as a whole, rather than the individual component parts, that must be examined for overall fairness," and we cannot "delete, modify or substitute certain provisions. The settlement must stand or fall in its entirety." Hanlon, 150 F.3d at 1026 (internal quotation marks omitted). See also Jeff D. v. Andrus, 899 F.2d 753, 758 (9th Cir.1989) ("[C]ourts are not permitted to modify settlement terms or in any manner to rewrite agreements reached by parties."). Thus, we must reverse the district court's order approving the settlement and dismissing the case, vacate the judgment, and remand for further proceedings.

On remand, the parties are free to negotiate a new settlement or proceed with litigation. If they again decide to settle, they must consider correcting the additional deficiencies that we find in this settlement agreement. Not only does the settlement fail to identify the cy pres recipients of the unclaimed money and food, but it is unacceptably vague in other areas as well.

The settlement states only that Kellogg will donate "$5.5 million worth" of food, (emphasis added). But the settlement gives no hint as to how that $5.5 million will be valued. Is it valued at Kellogg's cost? At wholesale value? At retail? Kellogg stated at oral argument and in its briefs to the district court that it will value the food donation at wholesale, but the only legally-enforceable document — the settlement — says nothing of the sort. Additionally, the settlement fails to include any restrictions on how Kellogg accounts for the cy pres distributions. Can Kellogg use the value of the distributions as tax deductions because they will go to charity? And given that Kellogg already donates both food and money to charities every year — which is unquestionably an admirable act — will the cy pres distributions be in addition to that which Kellogg has already obligated itself to donate, or can Kellogg use previously budgeted funds or surplus production to offset its settlement obligations? Again, the settlement is silent, and we have only Kellogg's statements as to its future intentions. All of this vagueness detracts from our ability to determine the true value of the common fund.

Neither class counsel nor Kellogg offers any credible reason for the mysteries in the current settlement. To approve this settlement despite its opacity would be to abdicate our responsibility to be "particularly vigilant" of pre-certification class action settlements. In re Bluetooth Headset Prods. Liab. Litig., 654 F.3d at 947.

B

Attorneys' Fees

Even if the cy pres distributions were sufficiently related to the class members and their false advertising claims, which they are not, we would still vacate the settlement because the $2 million award of attorneys' fees is unreasonable. Indeed, because the settlement grants counsel "a disproportionate distribution of the settlement" compared with the benefit to the class, it is possible the settlement was "driven by fees." Hanlon, 150 F.3d at 1021.

We review an award of attorneys' fees for abuse of discretion. Rodriguez v. West Publ'g Corp., 563 F.3d 948, 967 (9th Cir.2009). "[C]ourts have an independent obligation to ensure that the [fee] award, like the settlement itself, is reasonable, even if the parties have already agreed to an amount." In re Bluetooth Headset Prods. Liab. Litig., 654 F.3d at 941. In common fund cases like this one, "where the settlement or award creates a large fund for distribution to the class, the district court has discretion to use either a percentage or lodestar method." Hanlon, 150 F.3d at 1029.

Under a percentage analysis, attorneys' fees are calculated as a percentage of the common fund, the real value of which is simply illusive here. "This circuit has established 25% of the common fund as a benchmark award for attorney fees." Id. But not every fee award under 25% is necessarily reasonable. "[F]or example, where awarding 25% of a `megafund' would yield windfall profits for class counsel in light of the hours spent on the case, courts should adjust the benchmark percentage or employ the lodestar method instead." In re Bluetooth Headset Prods. Liab. Litig., 654 F.3d at 942.

The lodestar method of determining fees "begins with the multiplication of the number of hours reasonably expended by a reasonable hourly rate." Hanlon, 150 F.3d at 1029. In a common fund case, the court can then "apply a risk multiplier," which is "a number, such as 1.5 or 2, by which the base lodestar figure is multiplied in order to increase (or decrease) the award of attorneys' fees on the basis of such factors as the risk involved and the length of the proceedings." Staton, 327 F.3d at 967-68. Even where a court uses the percentage of the fund analysis, it can be helpful to apply a "lodestar cross-check." Vizcaino v. Microsoft Corp., 290 F.3d 1043, 1050 (9th Cir.2002). This "[c]alculation of the lodestar, which measures the lawyers' investment of time in the litigation, provides a check on the reasonableness of the percentage award. Where such investment is minimal, as in the case of an early settlement, the lodestar calculation may convince a court that a lower percentage is reasonable." Id.

Under either method, reasonableness remains the focus. Several factors are at play in the determination of a reasonable award of attorneys' fees. These factors include (1) the results achieved for the class, (2) the risk in bringing the suit, particularly with respect to the contingent nature of the fee, (3) the length of the litigation and the burdens borne by counsel throughout that litigation, and (4) any other relevant circumstance that might affect the amount of the award. Id. at 1048-51. See also In re Bluetooth Headset Prods. Liab. Litig., 654 F.3d at 942 (identifying "a host of `reasonableness' factors").

In this case, the district court applied a percentage of the fund analysis and approved the full $2 million in fees, which — according to the parties — represents approximately 19% of the common fund. Because the Objectors did not ask the district court to apply the lodestar analysis instead of the percentage analysis, they have forfeited that argument. Nevertheless, we will consider the lodestar cross-check as one factor in determining whether the fees are reasonable.

We conclude they are not. Considering that (1) the parties moved for settlement approval only three months after class counsel filed the amended complaint, (2) the settlement results in vaporous benefit to the class members and is flawed at its core, and (3) class counsel's financing of the litigation and investment of time were rather limited, we hold that the district court's reasonableness finding is implausible. See United States v. Hinkson, 585 F.3d 1247, 1263 (9th Cir.2009) (en banc).

We recognize that the abbreviated nature of legal proceedings does not always require a lower fee. We also understand the necessity of allowing counsel "a premium over their normal hourly rates for winning contingency cases" to account for the risk of non-payment if the case is unsuccessful. Vizcaino, 290 F.3d at 1051. However, the most important factors in this case — at least at this juncture — are the results achieved for the class and the lawyers' limited investment of time and money.

The settlement yields little for the plaintiff class. As discussed above, there is no reasonable certainty that the cy pres distributions as currently structured will benefit the class. The injunctive relief, prohibiting Kellogg from using the 20% attentiveness advertisements, lasts only three years. And class members, assuming they were aware of the litigation and submitted claims, will each receive the paltry sum of $5, $10, or $15.

In comparison, the $2 million award is extremely generous to counsel — even if we were to accept their assertion that the value of the common fund is $10.64 million. At the time the plaintiffs moved for settlement approval, class counsel had spent 944.5 hours working on the case. If the case had been litigated on an hourly basis at the attorneys' ordinary and uncontested rates, the total fees would have come to $459,203. The requested award, however, is about 4.3 times this lodestar amount. Although under the parties' valuation the award is below the 25% benchmark, a lodestar multiplier of 4.3 is quite high, particularly in a case that was not heavily litigated. Because the attorneys' investment was so minimal — as was the relief they claim to have obtained for the class — the lodestar cross-check leads us to the inescapable conclusion that the $2 million award is not reasonable. Cf. Fischel v. Equitable Life Assurance Soc'y of U.S., 307 F.3d 997, 1008 (9th Cir.2002) (lodestar multiplier of 1.5 was not unreasonably low given that the case settled early); Vizcaino, 290 F.3d at 1050-51 (lodestar cross-check of a 3.65 multiplier was not unreasonably high given that litigation extended over eleven years). Rather, it "yield[s] windfall profits for class counsel in light of the hours spent on the case." In re Bluetooth Headset Prods. Liab. Litig., 654 F.3d at 942.

This is so even assuming that the fund is actually worth what the parties say it is, an allegation very much in doubt. Moreover, once the alleged $5.5 million cy pres distribution is removed from the equation, the $2 million fee award becomes 38.9% of the remaining value of the fund. This figure is well above our presumptive benchmark, making the unreasonableness of the fee award all the more blatant.

Finally, let us not forget that the $2 million fee award breaks out to just over $2,100 per hour. Not even the most highly sought after attorneys charge such rates to their clients. Class counsel contends that the requested fees are reasonable because counsel have continued to represent the class on appeal and will do so throughout the administration of the settlement. But one reason why those counsel had to defend this appeal is because they negotiated a deficient settlement agreement. We do not believe it appropriate to reward counsel for failing to follow our cy pres precedent.

In overruling the objections to the attorneys' fees, the district court recited the Vizcaino factors but did not "adequately explain" its determination that those factors justified the fees as fair and reasonable. Powers v. Eichen, 229 F.3d 1249, 1257 (9th Cir.2000). Indeed, the court did not explain its determination at all. Given the high amount of the negotiated attorneys' fees, the court "needed to do more to assure itself — and us — that the amount awarded was not unreasonably excessive in light of the results achieved." In re Bluetooth Headset Products Liab. Litig., 654 F.3d at 943. If and when the issue of fees is again before the district court, the court shall consider all of the circumstances of the case as they exist at that time, including time wasted in preparing a stillborn settlement, in finally determining a reasonable award of attorneys' fees.

IV

CONCLUSION

Class counsel and Kellogg ask us for the impossible — a verdict before the trial. They essentially say, "Just trust us. Uphold the settlement now, and we'll tell you what it is later." But that is not how appellate review works. The settlement provides no assurance that the charities to whom the money and food will be distributed will bear any nexus to the plaintiff class or to their false advertising claims and therefore violates our well-established standards governing cy pres awards. Moreover, the attorneys' fees are impermissibly high considering what the defective settlement provides the class. The district court's contrary conclusions were an abuse of discretion.

REVERSED, JUDGMENT VACATED, and CASE REMANDED.