HARTZ, Circuit Judge.
After settlement of a class action for royalties from gas wells, the United States District Court for the Western District of Oklahoma awarded attorney fees to class counsel and an incentive award to the lead plaintiff to be paid out of the common fund shared by class members. The court rejected claims by two objectors, and they appealed. Exercising jurisdiction under 28 U.S.C. § 1291, we reverse and remand. The district court failed to compute attorney fees under the lodestar method, as required by Oklahoma law in this diversity case, and the incentive award is unsupported by the record.
I. BACKGROUND
The underlying class action alleged underpayment of royalties by the defendants on gas from wells in Oklahoma. The parties reached a settlement for a cash payment of $52 million, to be distributed pro rata to the class members after payment of expenses and fees. Class counsel moved for attorney fees in the amount of 40% of the settlement fund, plus interest; and the lead plaintiff, Chieftain Royalty Company, requested an incentive award of 1% of the fund. Appellants Charles David Nutley and Danny George were class members who objected to these requests. After a hearing on the settlement and fee requests, the court awarded class counsel 83 1/3% of the fund ($17,333,333.33) as attorney fees and awarded Chieftain 1/2% of the fund ($260,-000) as an incentive award. The objectors appealed each award. We address them in turn.
II. ATTORNEY FEE
There are two primary methods for determining attorney-fee awards in common-fund class-action cases. The first is the percentage-of-the-fund method, which awards class counsel a share of the benefit achieved for the class. See Newberg on Class Actions § 15:63 (5th ed. 2016) (Newberg). Many courts, including this circuit, consider 12 factors to determine the appropriate percentage. See Gottlieb v. Barry, 43 F.3d 474, 482 & n.4 (10th Cir. 1994). [1186]*1186These factors were first set forth in Johnson v. Georgia Highway Express, Inc., 488 F.2d 714, 717-19 (5th Cir. 1974), which was not a common-fund case. We have stated the factors as:
the time and labor required, the novelty and difficulty of the question presented by the case, the skill requisite to perform the legal service properly, the preclusion of other employment by the attorneys due to acceptance of the case, the customary fee, whether the fee is fixed or contingent, any time limitations imposed by the client or the circumstances, the amount involved and the results obtained, the experience, reputation and ability of the attorneys, the “undesirability” of the case, the nature and length of the professional relationship with the client, and awards in similar cases.
See Gottlieb, 43 F.3d at 482 n.4. The second method is the lodestar approach. The court first determines the lodestar by multiplying the number of hours reasonably spent on the litigation by a reasonable hourly rate. See Anchondo v. Anderson, Crenshaw & Assocs., LLC, 616 F.3d 1098, 1102 (10th Cir. 2010). This “produces a presumptively reasonable fee,” but it “may in rare circumstances be adjusted to account for the presence of special circumstances.” Id.
This court has approved both methods in common-fund cases, although expressing a preference for the percentage-of-the-fund approach. See Gottlieb, 43 F.3d at 483 (“In our circuit, following Brown [v. Phillips Petroleum Co., 838 F.2d 451 (10th Cir. 1988),] and Uselton [v. Commercial Lovelace Motor Freight, Inc., 9 F.3d 849 (10th Cir. 1993) ], either method is permissible. in common fund cases; however, Uselton implies a preference for the percentage of the fund method.”). Our approach also “has been called a ‘hybrid’ approach, combining the percentage fee method with the specific factors traditionally used to calculate the lodestar.” Id. at 482-83.
The district court chose the percentage-of-the-fund analysis, explaining that this is “[t]he preferred method of determining a reasonable attorney fee award in common fund cases.” JA at 523 (Dist. Ct. Order). It overruled the objectors’ argument that the lodestar approach should govern and that the fee is excessive under that analysis. The court stated that “[b]oth state and federal cases recognize and/or permit a percentage of fund recovery under the common fund doctrine.” Id. It added that in any event, the result.would not have differed under a lodestar analysis, citing Newberg on Class Actions § 14.6 at 551 (4th ed. 2002), for the proposition that empirical studies show that the average fee award is about one-third of the recovery, whichever method is used. See id. at 524.
Although a contingency-fee agreement allowed class counsel to recover 40% of any common-fund recovery, the district court ruled that “in fairness and consistent with the best interest of the Class,” counsel should recover 33 1/3% of the settlement. Id. at 523. It stated that an award of that percentage was not unusual, pointing out that “[t]he Tenth Circuit has previously identified the typical fee range as 23.7% to 33.7%.” Id. at 526 (citing Brown, 838 F.2d at 455 n.2). It then recited the Johnson factors and found that “most, if not all, ... support Class Counsel’s fee request, as reduced by the Court.” Id. It explained:
Class Counsel has conducted the Litigation and achieved the Settlement with skill, perseverance and diligent advocacy;
The Litigation involved complex factual and legal issues and was actively prosecuted for over four years;
[1187]*1187Had Class Counsel not achieved the Settlement, there would remain a significant risk that Class Representative and the other members of the Settlement Class may have recovered less or nothing from the Settling Parties;
Class Counsel devoted substantial time and resources to achieve the Settlement.
Id. (internal numbering removed). The court added that its award was informed by “(t]he market rate for Class Counsel’s legal services.” Id. at 528.
We review a district court’s award of attorney fees for abuse of discretion. See Gottlieb, 43 F.3d at 486. This includes review de novo of the legal principles underlying the fee award — such as the choice of whether to apply state or federal law. See Rosenbaum v. MacAllister, 64 F.3d 1439, 1444 (10th Cir. 1995); see also Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 405, 110 S.Ct. 2447, 110 L.Ed.2d 359 (1990) (district court abuses its discretion in Rule 11 determination if ruling is based on an erroneous Anew of the law).
Appellants argue that Oklahoma law governs the award of attorney fees in this case and requires using the lodestar approach rather than a percentage-of-the-fund analysis. We agree.1
Because federal jurisdiction in this common-fund case is based on the diversity of the parties, see 28 U.S.C. § 1332, the doctrine established in Erie R.R. Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938), requires us to apply Oklahoma law governing the award of attorney fees in common-fund cases. Under Erie, “federal courts in diversity cases must respect the definition of state-created rights and obligations by the state courts.” Byrd v. Blue Ridge Rural Elec. Coop., 356 U.S. 525, 535, 78 S.Ct. 893, 2 L.Ed.2d 953 (1958). In other words, the federal courts must recognize state-created substantive rights. Those rights include “the rules of decision by which [the] court will adjudicate [substantive] rights.” Mississippi Pub. Corp. v. Murphree, 326 U.S. 438, 446, 66 S.Ct. 242, 90 L.Ed. 185 (1946) (describing prohibition in Rules Enabling Act, 28 U.S.C. § 2072(b), against federal rules of procedure that modify substantive rights); see Shady Grove Orthopedic Assocs., P.A. v. Allstate Ins. Co., 559 U.S. 393, 407, 130 S.Ct. 1431, 176 L.Ed.2d 311 (2010) (plurality opinion); Scottsdale Ins. Co. v. Tolliver, 636 F.3d 1273, 1280 (10th Cir. 2011) (applying Oklahoma statute authorizing award of fees to defendant when judgment awarded is less than amount of offer of judgment). But the Erie doctrine sometimes also requires federal courts to apply state law that, in other contexts, might be deemed matters of procedure. “[T]he twin aims of the Erie rule [are] discouragement of forum-shopping and avoidance of inequitable administration of the laws.” Hanna v. Plumer, 380 U.S. 460, 468, 85 S.Ct. 1136, 14 L.Ed.2d 8 (1965). To advance those aims, at least when there is no contrary federal rule of civil procedure properly enacted under the Rules Enabling Act, see Shady Grove, 559 U.S. 393, 130 S.Ct. 1431 (2010), the Erie doctrine requires federal courts to “conform' as near as may be-r-in the absence of other considerations — to state rules even of form and mode where the state rules may bear substantially on the question whether the litigation would come out one way in the federal court and another way in the state court' if the feder[1188]*1188al court failed to apply a particular rule.” Byrd, 356 U.S. at 536-37, 78 S.Ct. 893.
We now apply this doctrine in the present context. To begin with, it is necessary to distinguish between two different types of attorney fees, depending on the basis for the fee award. In this circuit we have used the labels substantive and procedural to classify the two types of fees. Substantive fees are those that “are tied to the outcome of the litigation” and procedural fees are those that are “generally based on a litigant’s bad faith conduct in litigation.” Scottsdale, 636 F.3d at 1279. These labels are shorthand for those attorney fees that are governed by Erie (substantive fees) and those that are not (procedural fees). Substantive fees are part and parcel of the cause of action over which we have diversity jurisdiction. For example, if the cause of action is for bad-faith denial of insurance coverage and state law authorizes an award of attorney fees to a successful insured, then the right to an award of attorney fees is part of the state substantive right and the federal court must recognize it. In contrast, an attorney-fee award against bad-faith conduct in the litigation has nothing to do with the nature of the cause of action and does not derive in any way from state substantive law. “[F]ee-shifting here is not a matter of substantive remedy, but of vindicating judicial authority.” Chambers v. NASCO, Inc. 501 U.S. 32, 55, 111 S.Ct. 2123, 115 L.Ed.2d 27 (1991) (internal quotation marks omitted) (applying federal law in imposing sanctions for bad-faith litigation conduct in diversity case). One way to see that Erie deference to state law has no purchase in this context is to note that “there is no risk that [an award of sanctions for bad-faith conduct during litigation] will lead to forum-shopping.” Id. at 53, 111 S.Ct. 2123.
This appeal concerns substantive attorney fees. Whether to award counsel a fee out of a common fund is not based on whether counsel behaves properly during the litigation; rather, the award is “tied to the outcome of the litigation.” Scottsdale, 636 F.3d at 1279 (emphasis added) (fee award based on comparison of ultimate judgment to offer of judgment). In this context we have said, “In a diversity case, the matter of attorney’s fees is a substantive legal issue and is therefore controlled by state law.” N. Tex. Prod. Credit Ass’n v. McCurtain Cty. Nat’l Bank, 222 F.3d 800, 817 (10th Cir. 2000) (affirming denial of attorney fees because Oklahoma law did not authorize such an award for successful defense against claims of fraud and conspiracy). Our position finds support in leading authorities. Moore’s Federal Practice states that “when ... fees are connected to the substance of the case,” the question of whether a fee should be awarded is substantive. § 124.07[3][b] (3d ed. 2011) (Moore’s). And Wright & Miller endorses the view that “state law [is] controlling when it forbids an award of attorney’s fees” and that “when state law provides for the recovery of an attorney’s fee as a part of the claim being asserted[,] ... the federal court should permit an award of a fee on the theory that it is part of the substantive right in issue.” 10 Charles Alan Wright, Arthur R. Miller, Mary Kay Kane & Richard L. Marcus, Federal Practice and Procedures § 2669 at 263 (3d ed. 2014) (Wright & Miller).2
[1189]*1189There remains, however, the question whether the federal court must follow state law governing how to calculate the proper attorney fee. On that issue, there is no binding precedent in this circuit. But cf. Davis v. Prudential Prop. & Cas. Ins. Co., 145 F.3d 1345, 1998 WL 237255, at *3 (10th Cir. 1998) (unpublished disposition) (“The calculation of attorney’s fees in a diversity case is determined with reference to state law.” (emphasis added)). But there appears to be a consensus among those circuits that have considered the matter. We have found decisions from five other circuits. When state law governs whether to award attorney fees, all agree that state law also governs how to calculate the amount. See, e.g., In re Volkswagen & Audi Warranty Extension Litig., 692 F.3d 4, 15, 21-22 (1st Cir. 2012) (in diversity case “the issue of attorneys’ fees has long been considered for Erie purposes to be substantive and not procedural, and so state-law principles normally govern the award of fees,” and “[t]he district court should determine which method [lodestar or multifactor approach] Massachusetts would apply here”); Mathis v. Exxon Corp., 302 F.3d 448, 461 (5th Cir. 2002) (“State law controls both the award of and the reasonableness of fees awarded where state law supplies the rule of decision.”); Davis v. Mut. Life Ins. Co. of New York, 6 F.3d 367, 382-83 (6th Cir. 1993); Riordan v. Nationwide Mut. Fire Ins. Co., 977 F.2d 47, 53 (2d Cir. 1992); N. Heel Corp. v. Compo Indus., 851 F.2d 456, 475 (1st Cir. 1988).3
A leading treatise lends further support. Moore’s states, “When state law controls, state law governs not only the right to fees but also the method of calculating the fees.” Moore’s § 124.07[3][b] (emphasis added). The calculation of attorney fees is considered substantive law because “[t]he method of calculating a fee is an inherent part of the substantive right to the fee itself and reflects substantive state policy.” Id. (footnote omitted).
Here, the attorney-fee award was based on the outcome of the litigation not the district court’s power to discipline the litigants. State law therefore governs the propriety of granting a fee award. And we must also apply the State’s rules on how the amount of the fee is to be calculated because they are “rules of decision by which [the] court will adjudicate [the] right[ ] [to the fee].” Murphree, 326 U.S. at 446, 66 S.Ct. 242. To be sure, the ultimate standard for awarding a fee under either the lodestar or Johnson methodology is whether the fee is reasonable, so one might argue that the method of calculation to determine reasonableness is merely a matter of procedure. But it is state substantive law that cabins the meaning of reasonable. If state law declares that failure to wear a seatbelt while riding in a motor vehicle is never unreasonable (or is always unreasonable), there can be no question that a federal court hearing a [1190]*1190common-law negligence claim in diversity-must comply with that declaration. Similarly, a federal court in a diversity action must follow state law declaring that, absent extraordinary circumstances, a reasonable attorney fee is the fee computed under the lodestar method. We therefore see no reason to depart from what appears 'to be the consensus view that state law governs how to calculate a substantive attorney fee.
Chieftain nevertheless relies on Federal Rule of Civil Procedure 23(h), which governs class actions in federal court, arguing that “the method used to assess the reasonableness of [class counsel’s] fee is a procedural matter governed by Rule 23(h), not state procedural law.” Chieftain Br. at 18. We agree that this would be a more challenging issue if a federal rule of procedure said, for example, that fees should be calculated under the Johnson methodology. We would first have to determine whether the rule violated the prohibition in the Rules Enabling Act against rules that modify substantive rights. See 28 U.S.C. § 2072(b). And then assuming (although, for reasons' discussed above, we doubt the validity of the assumption) that the rule passed muster, we would probably have to apply it. See Shady Grove, 559 U.S. at 393, 130 S.Ct. 1431. But Rule 23(h) does not establish a rule of decision for assessing attorney fees.4 It provides that “[i]n a certified class action, the court may award reasonable attorney’s fees and nontaxable costs that are authorized by law or by the parties’ agreement.” And it outlines procedures to be followed before granting such an award: “A claim for an award must be made by motion ... at a time the court sets”; “[n]otice of the motion must be served on all parties and, for motions by class counsel, directed to class members in a reasonable manner”; “[a] class member, or a party from whom payment is sought, may object to the motion”; “[t]he court may hold a hearing and must find the facts and state its legal conclusions....”; and “[t]he court may refer issues related to the amount of the award to a special master or a magistrate judge....” In short, the rule addresses only how to conduct the court proceedings for determining the fee award. It does not provide any rules to guide the court on how to calculate the amount of the award.
Thus, we turn to Oklahoma law to determine how to compute the attorney fee in this case. The controlling precedent is Burk v. Oklahoma City, 598 P.2d 659 (Okla. 1979), a common-fund case. That decision directed that to enable a court to determine attorney fees, attorneys in Oklahoma must henceforth (the attorneys in that case were excused from this requirement) present “detailed time records showing the work performed and offer evidence as to the reasonable value for the services performed.” Id. at 663. This allows the court to determine the lodestar. [1191]*1191See id. Then other factors can be considered to provide an “incentive fee or bonus,” Id. at 661 (internal quotation marks omitted); see Hess v. Volkswagen of Am., Inc., 341 P.3d 662, 667 (Okla. 2014) (the lodestar “fee may be enhanced by application of certain factors”). The enhancement factors set forth in Burk have now been codified in a state statute. See 12 O.S. Supp. 2013 § 2023(G)(4)(e) (effective September 10, 2013, for motions filed after that date). The statutory factors are essentially the same as the Johnson factors, with the only material difference being that the statutory factors include the risk of recovery in the litigation. Compare id. with Johnson, 488 F.2d at 717-19.
Burk continues to be good law. See, e.g., Hess, 341 P.3d at 667; Spencer v. Okla. Gas & Elec. Co., 171 P.3d 890, 895 (Okla. 2007). We have been pointed to no contrary Oklahoma authority, nor have we found any. If anything, the Oklahoma Supreme Court has limited the application of the enhancement factors. In Hess it declared that “[tjhere is a strong presumption that the lodestar method, alone, will reflect a reasonable attorney fee.” 341 P.3d at 671. In support of that presumption the court cited Perdue v. Kenny A. ex rel. Winn, 559 U.S. 542, 552-54, 130 S.Ct. 1662, 176 L.Ed.2d 494 (2010), thereby signaling that it would follow the lead of the United States Supreme Court in greatly limiting departures from the lodestar figure.
The district court did not use the lodestar method to calculate class counsel’s fee in this case. Class counsel failed to provide the information necessary to apply that method. As already noted, in 1979 the Oklahoma Supreme Court stated that attorneys seeking fees must present “detailed time records” and “evidence as to the reasonable value for the services performed.” Burk, 598 P.2d at 663. Class counsel did not come close to performing this task. As the district court recognized, “Now, if I were to determine that the lodestar is applicable, then I think you will agree with me we just don’t have enough information in this case right now.” JA at 467-68. (Transcript of Fairness Hearing). Although class counsel claimed to have spent “much more” than 10,000 hours on the case, id. at 434, the firm acknowledged that “we don’t keep detailed time records on every hour we do in these cases,” id. at 421. Any time figures were mere estimates. To be sure, the district court stated that “utilization of either the common fund or lodestar method would not affect this Court’s ruling.” Id. at 524. .But 'this statement was not, and could not have been, based on computing the fee by both methods. Rather, the court was simply noting that empirical studies indicated that awards turned out to be about the same under both methods.
Therefore, we must set aside the attorney-fee award. The district court will have to decide in the first instance whether any award can be made in light of the absence of contemporaneous time records. It is unfortunate that class counsel did not do the necessary homework on Oklahoma law.5
[1192]*1192III. INCENTIVE AWARD
At the fairness hearing, class counsel sought an incentive award for the lead plaintiff, Chieftain, for its involvement in the litigation through its President, Robert Abernathy. Such awards are not uncommon. An empirical study published in 2006 reviewed 374 opinions in class actions from 1993 to 2002. See Theodore Eisenberg & Geoffrey P. Miller, Incentive Awards to Class Action Plaintiffs: An Empirical Study, 53 UCLA L. Rev. 1303 (2006) (Empirical Study). It reported that incentive awards were granted in about 28% of settled class actions; on average the incentive award was .16% of the class recovery, with a median of .02%. Id. at 1303. The average award per class representative was $15,992 and the median award was $4,357. Id. at 1348. More recent studies have shown a marked increase in the frequency of incentive awards, with the rate approaching 80% by 2011. See Newberg § 17:7. The average award dropped to $11,697 and the median increased to $5,250 in a study of cases from 2006 to 2011. See id. 17:8.
Counsel for Chieftain offered two grounds for an incentive award. One was the “risk or burden” Mr. Abernathy incurs as a result of his role as lead plaintiff. Counsel gave two examples. First, because he is litigating against a company called Merit, “he is not able to sit and talk to people at Merit like an ordinary royalty owner would because they know he’s our client and they know he is in these cases, so he has to go through a-little different procedure. And it’s very difficult for [Mr. Abernathy] to just do basic deals at times compared to your average royalty owner.” Id. at 440. Second, during this litigation he could not talk to the law firm that ordinarily does his estate planning, because that firm represents defendants in the underlying action. Counsel added that “there is a lot of hardship and risk when you’re in the State of Oklahoma where a lot of the buildings downtown are oil and gas companies and you live here and you’re going against-them in several different cases in ways that they don’t like.” Id.
Counsel for Chieftain also sought to justify an incentive award by describing Mr. Abernathy’s contribution to the case — his services rendered — as follows:
He is a licensed attorney. He has represented to me that he spent somewhere around 200 to 300 hours working in this case. And I know that without looking at any — what he does every day. I know that he has been directly involved in this case.
His deposition was taken in this case. He attended, I believe, two of the depositions in this case. He helps us prepare for all kinds of things during the settlement process. He reviews the pleadings in this case. He has been active in reviewing the pleadings for the settlement and the settlement documentation.
He is an advocate of royalty owners here in the state of Oklahoma. He has many, many friends that are royalty owners that he is in direct contact with about what happens in these cases and their rights in general.
As I mentioned, he told me a story recently, the OCC has had him go out and speak to out-of-state Oklahoma royalty owners. He was telling me — I think it was last year he was in Fort Myers and a lady came up to him and said, Are you the Chieftain from Chieftain vs. QEP1
He said, Yes.
She said, I want to give you a hug.
He said, Why?
She said, I got $40,000 in that case. Thank you for what you did.
So, you know, this isn’t, as the objectors try to insinuate, some guy who is [1193]*1193just out there and doesn’t know what he is doing and letting the lawyers run amuck over him. He is a real active helpful client.
Also, in this case, because there is this future component, somebody has got to look out and see if the defendants are actually complying with their future benefits under the settlement.
And what we will do and what Mr. Abernathy will do is we will look at his statements every month. We will probably have Ms. Ley and class counsel look at- them too. We will be doing this for the next three years to make sure that nothing changes on the way his payments are calculated.
We had an audit provision in the QEP case. We don’t have it here. But, based on that experience, I think Mr. Abernathy would say — when we talked — it will probably be another 50 hours to 100 hours doing that over the next three years. Hopefully, it’s not that much, but we have no way to know. So this is real time. He is a licensed attorney.
JA at 438-39 (Transcript of Fairness Hearing).
The district court agreed that Chieftain should receive an incentive award. It stated that “ ‘a class representative may be entitled to an award for personal risk incurred or additional effort and expertise provided for the benefit of the class,’ ” id. at 528 (Dist. Ct. Order) (quoting UFCW Local 880-Retail Food Emp’rs Joint Pension Fund v. Newmont Mining Corp., 352 Fed.Appx. 232, 235 (10th Cir. 2009)), and that “[c]ase contribution awards are meant to compensate class representatives for their work on behalf of the class, which has benefited from their representation,” id. (internal quotation marks omitted). It also observed that “[ijncentive awards are not uncommon and are particularly given where a common fund has been created for the benefit of the entire class.” Id. at 529. It concluded that it should consider “(1) the actions the class representative took to protect the interests of the class; (2) the degree to which the class has benefitted from those actions; and (3) the amount of time and effort the class representative expended in pursuing the litigation.” Id.
The district court ruled as follows:
A Case Contribution Award is appropriate in this case. Class Representative— through its President, Robert Abernathy — has been actively involved in this Litigation since its inception. Mr. Abernathy has contributed by reviewing draft pleadings and motions, searching for and producing records, reviewing filings, communicating regularly with Class Counsel, making himself available by telephone during the formal mediation, continuously monitoring the Litigation and settlement process, and approving the terms of the Settlement. Mr. Abernathy’s efforts helped lead to a settlement that greatly benefits the Class, and Class Representative should be rewarded for those efforts.
Id.
As for the amount of the award, the district court said that 1/2% of the fund was “fair” and “reasonable.” Id. at 530. In reducing the award from the amount sought (1%), the court pointed to the reduction of the attorney fee sought and stated that “a commensurate reduction of the requested [incentive reward] is appropriate.” Id. The court also pointed out that “[sjimilar [incentive] awards have been granted in similar cases.” Id. (citing opinions in other cases, apparently involving royalty litigation, in Oklahoma federal district court in which the incentive award was between 1% and 2% of the settlement amount).
[1194]*1194We review a district court’s grant of an incentive award for abuse of discretion. See Cobell v. Jewell, 802 F.3d 12, 23 (D.C. Cir. 2015); Montgomery v. Aetna Plywood, Inc., 231 F.3d 399, 408 (7th Cir, 2000); UFCW Local 880, 352 Fed.Appx. at 235 (“Generally, our standard of review for a district court’s award of attorney fees in the class-action context is abuse of discretion”; and “the district court’s familiarity with the parties and the proceedings supports an abuse-of-discretion standard” in reviewing an incentive award); Newberg § 17:21 (endorsing analysis in UFCW Local 880).
Appellant Nutley argues on appeal that incentive awards are unlawful per se in common-fund cases. He asserts that two U.S. Supreme Court decisions — Cent. R.R. & Banking Co. v. Pettus, 113 U.S. 116, 5 S.Ct. 387, 28 L.Ed. 915 (1885), and Trustees v. Greenough, 105 U.S. 527, 26 L.Ed. 1157 (1881) — “hold that a litigant whose efforts create a common fund may recover expenses reasonably incurred, including its reasonable attorneys’ fees, but not incentive awards for services rendered.” Aplt. Br. (Nutley) at 27. But Nut-ley forfeited any argument about the general legality of incentive awards by failing to raise it below. We cannot find any indication in the record that Nutley raised this argument in district court. On the contrary, he stated in his written objection to the award that incentive awards are sometimes proper. See JA at 319 (Nutley’s objection) (“The concept of a service award is not unfamiliar in this Circuit, and may be appropriate when necessary to induce people to become named representatives, or to compensate personal risk or ‘additional effort expended’ by the representative in prosecuting the suit.”). Nutley claims that he preserved this issue because “[a]t the fairness hearing, [his] counsel challenged Chieftain’s right to claim any award for Abernathy’s time on the case.” Aplt. Joint Reply Br. at 24-25. As that statement suggests, however, his counsel spoke at the fairness hearing only of the propriety of Chieftain’s incentive award — not of the legality of these awards in general. See JA at 459-62 (Transcript of Fairness Hearing). And the district court made no ruling on the broader issue. Because this argument was forfeited, we can review only under the plain-error standard. See Richison v. Ernest Grp., Inc., 634 F.3d 1123, 1130 (10th Cir. 2011). But ordinarily we will not even do that unless the party specifically argues plain error in its briefs, see id. at 1131, and Nutley’s briefs made no such argument.
We therefore turn to the specific award in this case. As stated above, Chieftain raised in district court two potential grounds for receiving an incentive award. One was the risk or burden incurred'by Chieftain and Mr. Abernathy. Courts have recognized that an award may be appropriate to provide an incentive to act as a named plaintiff. See In re Synthroid Mktg. Litig., 264 F.3d 712, 722-23 (7th Cir. 2001) (“Incentive awards are justified when necessary to induce individuals to become named representatives.... But if at least one [class member] would have stepped forward without the lure of an ‘incentive award,’ there is no need for such additional compensation.”); Cook v. Niedert, 142 F.3d 1004, 1016 (7th Cir. 1998). (“Because a named plaintiff is an essential ingredient of any class action, an incentive award is appropriate if it is necessary to induce an individual to participate in the suit.”); Newberg § 17:3 (incentive awards can “in-eentivize class members to step forward” despite risks such as the possibility that the class representative could be liable for the costs of the suit or might face retaliation).
The district court, however, did not justify the incentive award on the basis of any [1195]*1195risk or burden incurred by Chieftain or Mr. Abernathy or on the need for an incentive for Chieftain to act as a named plaintiff. Indeed it made no findings on the subject. This is understandable, given (1) that the notice to the class stated that the requested incentive award would be compensation for time and effort (with no mention of risk or burden) and (2) that the presentation of risk and burden at the hearing on the award was quite weak.6 Accordingly, there is no foundation for us to affirm, in whole or in part, the incentive award on this ground. The district court is free to decide in the first instance whether under the circumstances of this case — including the absence of an initial court finding on the matter — Chieftain should be given another opportunity on remand to make a better showing under the risk-or-burden rationale.
Hence, we need analyze only Chieftain’s second argument in support of an award — that it would fairly compensate for the services rendered by Mr. Abernathy. Our first task is to determine the governing law. We think it clear that, as with the attorney-fee award discussed above, Erie requires us to apply Oklahoma law. Unfortunately, the Oklahoma Supreme Court apparently has not addressed incentive fees, nor have we been directed to or found any opinions by lower courts of that state. Our task then is to make an informed prediction of what the State’s highest court would do. See United States v. Badger, 818 F.3d 563, 568-69 (10th Cir. 2016). When “no state cases exist on a point, we turn to other state court decisions, federal decisions, and the general weight and trend of authority.” Sender v. Simon, 84 F.3d 1299, 1303 (10th Cir. 1996) (internal quotation marks omitted). In conducting this analysis we have the advantage of a treatise on class actions that discusses and cites relevant judicial decisions.
To begin with, we note that courts regularly give incentive awards to compensate named plaintiffs for the work they performed — their time and effort invested in the case. See, e.g., Cobell v. Salazar, 679 F.3d 909, 922-23 (D.C. Cir. 2012) (district court did not err in finding that lead plaintiff’s “singular, selfless, and tireless investment of time, energy, and personal funds to ensure survival of the litigation [merited] an incentive award”); Rodriguez v. West Publ’g Corp., 563 F.3d 948, 958 (9th Cir. 2009) (“Incentive awards ... are intended to compensate class representatives for work done on behalf of the class.... ”). These services typically include “monitoring class counsel, being deposed by opposing counsel, keeping in[1196]*1196formed of the progress of the litigation, and serving as a client for purposes of approving any proposed settlement with the defendant.” Newberg § 17:3. The award should be proportional to the contribution of the plaintiff. See Phillips v. Asset Acceptance, LLC, 736 F.3d 1076, 1081 (7th Cir. 2013) (if the lead plaintiffs services are greater, her incentive award likely will be greater); Rodriguez, 563 F.3d at 960 (incentive award should not be “untethered to any service or value [the lead plaintiff] will provide to the class”); Newberg § 17:18.
In our view, however, the Oklahoma Supreme Court would not approve the award given here. The district court granted Chieftain a 1/2% incentive award of $260,000. Yet the weight of authority apparently disfavors percentage-based awards. See Newberg § 17:16 (“Percentage-based awards are disfavored, if not altogether forbidden.”). There are several reasons to reject the practice. Most importantly, “sealing those rewards according to the size of the common fund is at best a rough proxy in that the services and risks are not necessarily directly related to the size of the settlement.” Id. In addition, percentage awards “skew the class representatives’ incentives by encouraging them to hold out for greater recovery ... when in fact the class’s interest would be best served by a settlement”; “percentage awards privilege monetary recoveries over other remedies, such as injunctive relief, creating a potential conflict between the interest of the class representative and the class”; “percentage awards threaten to be excessive”; and “paying the class representatives a portion of the settlement fund is simply unseemly: it gives the appearance that the representative is either a professional plaintiff, or bounty hunter, not a servant for the class.” Id. (footnotes omitted). If a percentage calculation is to be made at all, it should be made only to “check a flat award for excessiveness by reference to the percentage of the fund it represents.” Id.
We therefore examine whether the award to Chieftain can be justified as payment at a reasonable rate for reasonable time expended on services rendered that were helpful to the litigation and did not duplicate what could be performed less expensively by counsel. Appellant George concedes that such payment would be proper.7 See Aplt. Br. (George) at 30. And this approach to measuring the value of the work is consistent with the Oklahoma Supreme Court’s lodestar approach to attorney-fee awards in common-fund cases. The question is what rate and what efforts are reasonable. The answer must be supported by sufficient evidence in the record. See Cook, 142 F.3d at 1016 (upholding district court’s incentive award after noting that findings justifying the award were “well-supported by the evidence”). This evidence might be provided through “affidavits submitted by class counsel and/or the [1197]*1197class representatives, through which these persons testify to the particular services performed, the risks encountered, and any other facts pertinent to the award.” New-berg § 17:12. “Courts may also receive this evidence by live testimony at the fairness hearing.” Id. They “regularly reject awards where the relevant facts are not sufficiently documented.” Id. (collecting cases).
The record before us is devoid of evidence from which a computation could be made. The district court was not provided supporting documents either when deciding to grant the incentive award or when determining the amount of the award. Instead, as can be seen in the excerpt from the transcript above, counsel spoke broadly about the tasks Mr. Abernathy had performed in the case and offered an anecdote about a class member in a previous suit who was grateful for his work there. When discussing the time Mr. Abernathy had expended on the case, counsel did not provide detailed contemporaneous records but offered only approximations and generalities. See JA at 438 (Transcript of Fairness Hearing) (“[Mr. Abernathy] spent somewhere around 200 to 300 hours working in this casé.”); id. at 439 (“[I]t will probably be another 50 hours to 100 hours [of work for Mr. Abernathy] over the next three years.”). Therefore, in keeping with our prediction of what the Oklahoma Supreme Court would command, we must reverse for abuse of discretion and remand for further fact-finding.
IV. CONCLUSION
We REVERSE the attorney-fee and incentive awards and REMAND for further proceedings consistent with this opinion.