Plaintiff-appellant Denise Coutin,1 flush with victory after winning an employment discrimination suit, encountered disappointment when the district court awarded her only a fraction of the attorneys' fees to which she believed herself entitled under the Fees Act, 42 U.S.C. § 1988 (1994). Coutin appeals. Because the district court employed a flawed methodology and relied on impermissible criteria, we vacate its order and remand for further proceedings.I. BACKGROUND

On December 30, 1993, the appellant sued her former employer, defendant-appellee Young & Rubicam of Puerto Rico, Inc. (Y & R), an advertising agency, for over $1,500,000 in compensatory and punitive damages. Her complaint advanced one substantive federal claim: that Y & R had violated Title VII of the Civil Rights Act of 1964, 42 U.S.C. §§ 2000e to 2000e-17 (1994), including the Pregnancy Discrimination Act, 42 U.S.C. § 2000e(k) (1994), by (1) assigning Coutin (who was then pregnant) to tasks that were detrimental to her physical and emotional health, (2) requiring her to work under unsafe conditions, (3) condoning (or, at least, neglecting to curb) her coworkers' disparaging comments about her gravidity, and (4) constructively discharging her. The complaint also included several claims under local law, the elements of which were subsumed, without exception, under the broader federal claim.

Y & R denied Coutin's allegations and defended the suit with considerable vigor. Along the way, the parties attempted to reach an accord, but they came no closer than a demand of $150,000 as against an offer of $15,000. At trial, the appellant (who had secured and retained other employment) offered no evidence of lost income, and that aspect of her original claim was pretermitted. The case went to the jury, which found that Y & R had intentionally discriminated against, and constructively discharged, the appellant, thus violating both federal and Puerto Rico law. The jury awarded the appellant and her spouse a total of $44,000 in compensatory damages, plus an additional $1,538 in severance pay under Law 80, P.R. Laws Ann. tit. 29, § 185a (1985). The jury rejected the appellant's prayer for punitive damages.

Y & R, which had moved unsuccessfully for judgment as a matter of law on several occasions during the trial, renewed that motion and asked alternatively for a new trial. See Fed.R.Civ.P. 50, 59. The district court refused relief. In turn, the appellant petitioned under 42 U.S.C. § 1988 for an award of $52,793.75 in counsel fees and related expenses. The fee application contained a sworn statement delineating her lawyer's two decades of experience in personal injury, labor, and discrimination cases, as well as extensive, contemporaneous billing records that detailed the lawyer's work over four calendar years. At the bottom line, the reckoning reflected out-of-court time (250.25 hours) billed at $175 per hour and in-court time (45 hours) billed at $200 per hour.

Despite this meticulous proffer, the judge eschewed any discussion of either the hours spent or the billing rates assigned and instead awarded the appellant a mere $5,000 in fees. The judge purported to base his decision entirely on the "plaintiffs' limited success on their claims, the plaintiffs' willingness to go to trial despite the defendant's earnest efforts to settle the case for a reasonable sum, and the equities involved."2 This appeal ensued.

We review fee awards deferentially, according substantial respect to the trial court's informed discretion. See Brewster v. Dukakis, 3 F.3d 488, 492 (1st Cir.1993). We will disturb such an award only for mistake of law or abuse of discretion. See United States v. Metropolitan Dist. Comm'n, 847 F.2d 12, 14 (1st Cir.1988). In this regard, an abuse of discretion occurs "when a material factor deserving significant weight is ignored, when an improper factor is relied upon, or when all proper and no improper factors are assessed, but the court makes a serious mistake in weighing them." Foster v. Mydas Assocs., Inc., 943 F.2d 139, 143 (1st Cir.1991) (internal quotation marks and citation omitted).

Although our analytical posture is respectful, we nonetheless must engage the district court's decision critically. To facilitate this perlustration, we require the lower court to explain its actions. See id. at 141. The explanation need not be painstaking, and, sometimes, it may even appear by implication, but at a bare minimum, the order awarding fees, read against the backdrop of the record as a whole, must expose the district court's thought process and show the method and manner underlying its decisional calculus. See Blum v. Stenson, 465 U.S. 886, 898, 104 S. Ct. 1541, 1548-49, 79 L. Ed. 2d 891 (1984); Hensley v. Eckerhart, 461 U.S. 424, 437, 103 S. Ct. 1933, 1941, 76 L. Ed. 2d 40 (1983).

This principle is especially important when the fee award departs substantially from the contours shaped by the application. "As a general rule, a fee-awarding court that makes a substantial reduction in either documented time or authenticated rates should offer reasonably explicit findings, for the court, in such circumstances, 'has a burden to spell out the whys and wherefores.' " Brewster, 3 F.3d at 493 (quoting Metropolitan Dist. Comm'n, 847 F.2d at 18). An appellate court deprived of meaningful insight into the trial court's thinking frequently will be unable to conduct an adequate review of a significantly adjusted fee award, and thus will be compelled to remand for further findings. See, e.g., Riley v. City of Jackson, 99 F.3d 757, 760 (5th Cir.1996); Freeman v. Franzen, 695 F.2d 485, 494 (7th Cir.1982).

The lodestar method is the strongly preferred method by which district courts should determine what fees to award prevailing parties in actions that fall within the ambit of section 1988. See Lipsett v. Blanco, 975 F.2d 934, 937 (1st Cir.1992). This approach contemplates judicial ascertainment of "the number of hours reasonably expended on the litigation multiplied by a reasonable hourly rate" as the starting point in constructing a fee award. Hensley, 461 U.S. at 433, 103 S.Ct. at 1939. While the lodestar method is a tool, not a straitjacket--as we have acknowledged, some deviation from an orthodox application of the method is permissible in highly unusual situations, see Metropolitan Dist. Comm'n, 847 F.2d at 15-16--a fee-awarding court shuns this tried-and-true approach at its peril. See Segal v. Gilbert Color Sys., Inc., 746 F.2d 78, 87 (1st Cir.1984). As we have said, the lodestar method is a tool, but it is not merely a tool. The method is also a device which enables courts to pay homage to the fundamental reason that Congress passed the Fees Act: its resolve that certain types of wrongs, such as discrimination on account of sex, should not be countenanced, and that private suits aimed at redeeming such abuses should be encouraged. See City of Riverside v. Rivera, 477 U.S. 561, 574-75, 106 S. Ct. 2686, 2694-95, 91 L. Ed. 2d 466 (1986) (plurality opinion).

To say that a trial court mulling a fee request ordinarily must fashion a lodestar is not to say that the court is in thrall to an attorney's time records. The court can segregate time spent on certain unsuccessful claims, see, e.g., Hensley, 461 U.S. at 435, 103 S.Ct. at 1940, eliminate excessive or unproductive hours, see, e.g., Lipsett, 975 F.2d at 937, and assign more realistic rates to time spent, see, e.g., Brewster, 3 F.3d at 492. In these and other ways, the trial court, though adhering to the time-and-rate-based method of fee calculation, may fashion a lodestar which differs substantially from the fee requested by the prevailing party. Moreover, the trial court retains the authority to adjust the lodestar after initially computing it--but it must do so in accordance with accepted principles. See Hensley, 461 U.S. at 429-31, 103 S.Ct. at 1937-38 (citing the legislative history of the Fees Act and observing that it is appropriate to adjust fees in accordance with the twelve factors set forth in Johnson v. Georgia Highway Express, Inc., 488 F.2d 714, 717-19 (5th Cir.1974)).3IV. ANALYSIS

In this instance, Coutin submitted the documentation needed to permit the district court to follow the conventional approach, but the court scarcely mentioned that proffer and, in all events, did not engage in any lodestar analysis whatever. While such a departure from preferred practice will not necessarily be fatal, spurning all consideration of a lodestar places a substantial burden upon the district court to account for its actions. See Berg v. Gackenbach, 966 F.2d 731, 732 (2d Cir.1992); Metropolitan Dist. Comm'n, 847 F.2d at 12, 15.

Here, the district court did not cite any reason for abjuring the lodestar method. Still, the court did make an effort to explain its decisionmaking process, indicating that it had premised its decision on three factors: the appellant's limited success, the parties' abortive efforts to settle the case, and the equities of the situation. But the court's reliance on the first of these factors is at best insufficiently explained, and its reliance on the other two factors is plainly wrong. Moreover, none of these factors justifies the court's failure to compute (and then adjust, if necessary) a lodestar.

A. Limited Success.

The district court's conclusion that the appellant enjoyed only "limited success" (and, thus, deserved less in the way of counsel fees) is too much of a stretch. To be sure, as the court pointed out, there was a chasmal gulf between the damages requested in the complaint and the damages awarded.4 The court had a right to keep this discrepancy in mind, but it cannot amount to more than one element in the constellation of factors that the court considers when determining the quality of the results obtained. Because this phenomenon is sometimes misunderstood, we take some pains to explain it.

As Judge Casellas correctly noted, the Supreme Court has identified results obtained as a preeminent consideration in the fee-adjustment process. See Hensley, 461 U.S. at 432, 440, 103 S. Ct. at 1938-39, 1943. But the term "results obtained" has a variety of meanings. It can refer to a plaintiff's success claim by claim, or to the relief actually achieved, or to the societal importance of the right which has been vindicated, or to all of these measures in combination. We think that the last meaning is the best choice, and that, as a consequence, all three types of "results" potentially bear upon the amount of an ensuing fee award. See generally Norman v. Housing Auth. of Montgomery, 836 F.2d 1292, 1302 (11th Cir.1988).

Although all three measures of success must be factored into the fee-reduction calculus, they do not lend themselves to identical treatment. On the one hand, to the extent that fee adjustments are intended to reflect the success or failure of severable claims, they are relatively easy to calculate because, although some overlap may muddy the waters, a court usually can determine the extent to which a plaintiff has prevailed on her claims merely by perusing the docket (e.g., the complaint, the verdict form, etc.), and can then filter out time spent on unsuccessful claims. See, e.g., Lipsett, 975 F.2d at 940-41. On the other hand, a fee reduction in response to a scanty damage award or a shortfall in other relief entails a subjective evaluation of damages awarded and nonmonetary relief obtained, and is substantially more difficult to quantify.5 Seen in this light, the computational principles applicable to claims-based fee reductions are relatively simple and straightforward, whereas the computational principles applicable to relief-based fee reductions are highly ramified and, in some respects, operate at cross purposes. To visualize how these sometimes competing principles may affect a district court's effort to determine a "reasonable" fee, it may be helpful to catalog the several possible configurations in which the issue may arise.

1. If a plaintiff prevails on only some of multiple claims, then a fee reduction may be in order. To guide decisionmaking in this situation, the Justices have suggested two relevant questions: "First, did the plaintiff fail to prevail on claims that were unrelated to the claims on which he succeeded? Second, did the plaintiff achieve a level of success that makes the hours reasonably expended a satisfactory basis for making a fee award?" Hensley, 461 U.S. at 434, 103 S.Ct. at 1940.

When different claims for relief are not interconnected--that is, when the claims rest on different facts and legal theories--they are by definition severable and unrelated. Attorneys' fees normally should not be awarded for time spent in litigating (or preparing to litigate) unsuccessful, severable claims. See id. at 435, 103 S. Ct. at 1940; Lipsett, 975 F.2d at 940.

2. If a plaintiff prevails on an insubstantial subset of her interrelated claims and obtains only limited relief, the trial court has discretion to shrink fees to reflect that inferior result. See Hensley, 461 U.S. at 436, 103 S.Ct. at 1941; Andrade v. Jamestown Hous. Auth., 82 F.3d 1179, 1191 (1st Cir.1996). Withal, a plaintiff who has limited success from a claim-by-claim standpoint, but who nevertheless obtains substantial compensation or other important relief, usually will fare much better in the fee wars, even though some of her claims failed. See, e.g., Hensley, 461 U.S. at 440, 103 S.Ct. at 1943 ("Where a lawsuit consists of related claims, a plaintiff who has won substantial relief should not have his attorney's fee reduced simply because the district court did not adopt each contention raised.").

3. If a prevailing party is successful on all (or substantially all) of her claims, and receives complete (or near-complete) relief, it goes without saying that reasonable fees should be paid for time productively spent, without any discount for limited success.

4. If a prevailing party succeeds on all (or substantially all) of her claims, but receives no significant relief (e.g., the jury awards only nominal damages), the trial judge sometimes may deny fees altogether because this scenario often "highlights the plaintiff's failure to prove actual, compensable injury." Farrar v. Hobby, 506 U.S. 103, 115, 113 S. Ct. 566, 575, 121 L. Ed. 2d 494 (1992) (denying fees in a case in which the plaintiff sought $17,000,000 in damages and received $1); see also id. at 114, 113 S.Ct. at 574 (affirming that "the 'technical' nature of a nominal damages award or any other judgment ... does bear on the propriety of fees awarded under § 1988"). Farrar, then, signifies that fees need not be bestowed if the plaintiff's apparent victory is "purely technical or de minimis." Id. at 117, 113 S.Ct. at 576 (O'Connor, J., concurring).6

5. Sometimes, the plaintiff will prevail on all her claims, but will receive limited (though not insubstantial) redress. In such circumstances, it is appropriate for a trial court to consider the skimpiness of the relief when adjusting the lodestar figure. See Rivera, 477 U.S. at 574, 106 S.Ct. at 2694. But though a meager damage award may be taken into consideration, the Court has squarely disclaimed "the proposition that fee awards under § 1988 should necessarily be proportionate to the amount of damages a civil rights plaintiff actually recovers." Id. (approving $245,456.25 in fees in a section 1983 action that resulted in a judgment for $33,350); see also id. at 585, 106 S.Ct. at 2699-700 (Powell, J., concurring) (noting that "[n]either the decisions of this Court nor the legislative history of § 1988" lend credence to a "rule of proportionality between the fee awarded and the damages recovered in a civil rights case"); Foley v. City of Lowell, 948 F.2d 10, 19 (1st Cir.1991) (holding that although a trial court is "entitled to take into account the relative size of the damage award and the fee award," the former "does not constitute a dispositive criterion, or even a ceiling" on the latter).

It is readily apparent that some tension exists between these principles: while a judge may not automatically reduce a fee award in proportion to a judgment that is significantly less than the plaintiff sought, the judge can take that small judgment into reasonable account in massaging the lodestar. This dissonance makes it all the more crucial that a nisi prius court provide a clear explanation when limited relief furnishes the ostensible justification for a departure from the lodestar.

Conscious of these differing configurations, we turn to the case at hand. The focus of our inquiry is the lower court's determination that the appellant's success was limited (and, thus, justified a fee reduction).

We start by scrutinizing claims-based success. Y & R asserts that Coutin's victory was less than complete both because the jury declined to grant punitive damages and because the appellant did not pursue her original prayer for lost income. We believe that this riposte blurs the distinction between claims and damages. In the fee-shifting context, a "claim" is an allegation of a legal injury comprised of various elements and equivalent to a cause of action, whereas "damages" are the compensation awarded to the plaintiff who has suffered a legal wrong and who therefore has a valid claim against the defendant. In this case, punitive damages and loss of income (no matter how they are denominated in the complaint) are not failed claims, but are categories of relief that the jury and the appellant's subsequent employment history, respectively, have denied her.

We need not linger. From a claim-by-claim standpoint, the appellant prevailed up and down the line. She triumphed on every substantive claim asserted under both federal and Puerto Rico law. In so doing, she achieved a 100% success rate--and complete success is hardly "limited." Consequently, a claims-based, results-obtained fee reduction is wholly inappropriate.

From the standpoint of relief obtained, the situation is more ambiguous. After all, it remains within the district court's discretion to reduce a fee award in response to limited relief even in the presence of complete claims-based success. See, e.g., Cartwright v. Stamper, 7 F.3d 106, 109-10 (7th Cir.1993) (declining to award fees where plaintiff succeeded on all claims, but received only nominal damages for each). Here, however, the damage award is substantial in absolute terms--over $45,000--and equals roughly three times the appellant's annual salary. On its face, such relief does not seem "limited" in any relevant sense.

Moreover, to the extent--if at all--that the ratio of the damages requested to the judgment received may be taken into account in fixing an appropriate award, see Foley, 948 F.2d at 19-20 ("Often, when the amount sought is large but the actual recovery is small, fees may be reduced somewhat."); see also Loggins v. Delo, 999 F.2d 364, 369 (8th Cir.1993), this proportion may be used only as one facet of the trial court's determination of the quality of the results obtained. The court may not employ the derived ratio as an independent justification for a fee reduction. See Rivera, 477 U.S. at 574, 106 S.Ct. at 2694. Rather, in the absence of special circumstances, the court must evaluate the data submitted by the fee-seeker, compute a lodestar, consider the totality of the adjustment factors approved by Congress and the Court, see supra note 3, and make specific, reasoned adjustments if it is to arrive at a reduced fee award.

In the instant case, the court did not analyze the appellant's time-and-rate data; it ignored the appellant's broad claims-based success; and it failed to explain why the sum upon which it settled--$5,000--was itself reasonable in relation to counsel's efforts, even given a perceived shortfall in the relief received. Thus, the more than 90% fee reduction that the court imposed cannot be justified on the basis of limited success.7

B. Settlement Prospects.

In other instances, inquiry into the course of settlement negotiations may yield information that is useful in determining fees. See Marek v. Chesny, 473 U.S. 1, 11-12, 105 S. Ct. 3012, 3017-18, 87 L. Ed. 2d 1 (1985) (applying Fed.R.Civ.P. 68 in a civil rights context). In the case at bar, however, the defendant did not invoke Rule 68 and, in any event, the judgment that the plaintiff obtained more than trebled the highest settlement offer available to her. This success validates the appellant's rejection of the tendered settlement and immunizes her from detrimental consequences based upon that rejection. See Corder v. Gates, 947 F.2d 374, 380-81 (9th Cir.1991).

Policy considerations militate strongly against relaxing this rule. Permitting a district court to reduce a fee award for failure to settle when the eventual judgment exceeds the best settlement offer previously made by the losing party would put too large a club in the district court's hands. In the bargain, endorsing that praxis would create inordinate pressure on plaintiffs to accept low settlement offers. This result would inhibit the bringing of civil rights actions, and, in the end, frustrate Congress's manifest intention that the Fees Act facilitate the prosecution of private actions aimed at deterring civil rights abuses. See Rivera, 477 U.S. at 574-75, 106 S.Ct. at 2694-95. We therefore hold that it is a mistake of law to reduce an award of attorneys' fees in a civil rights case in response to a plaintiff's rejection of a defendant's settlement offer when the subsequent judgment exceeds that offer.8

C. Windfall.

The district court's opinion suggests that a fee reduction is appropriate because the appellant had a shaky case and did not deserve to prevail on the merits. See supra note 2. Wholly apart from the accuracy vel non of the district court's assessment, this criterion is not a proper element of the fee-award calculus.

Congress intended the Fees Act to effect attorney compensation in virtually all cases involving successful civil rights claims.9 See generally S.Rep. No. 94-1011 (1976), reprinted in 1976 U.S.C.C.A.N. 5908; see also Williams v. Hanover Hous. Auth., 113 F.3d 1294, 1300 (1st Cir.1997). After a jury's verdict has been rendered and has withstood whatever barrage of post-trial motions may ensue, the time for debate has expired. It is an abuse of discretion for the trial court thereafter to vent its skepticism about the claimant's right to recover by reducing the fee award to which the prevailing party is entitled. See Stefan v. Laurenitis, 889 F.2d 363, 370-71 (1st Cir.1989) (refusing to allow district courts to balance equities of this sort when considering whether to award attorneys' fees under section 1988); see also deJesus v. Banco Popular, 918 F.2d 232, 235 (1st Cir.1990). In other words, the time for a trial judge to express his doubts about the viability of a claim occurs when the judge rules upon the full panoply of motions for judgment as a matter of law and/or for a new trial. Once a case has scaled those barriers--and this case has--trimming attorneys' fees cannot be employed as a palliative to assuage lingering doubts about the legal viability of the claim. See deJesus, 918 F.2d at 235. Indeed, if a plaintiff has a thin case but nonetheless manages, as here, to secure a verdict for three times the largest settlement offer, such a template suggests skillful advocacy, perhaps worthy of an award of full fees.

D. Local Law.

Wholly apart from the district court's rationale, Y & R has a fallback position. It posits that, whatever the infirmities of the $5,000 fee award under federal law, the award comports with Puerto Rico law and should be sustained on that basis. This thesis, which proposes that Puerto Rico law should govern in respect to fees because the appellant prevailed on her non-federal claims and recovered double damages by operation of Puerto Rico law, fails for two reasons.

First, under section 1988 "the plaintiff is entitled to fees for hours worked not only on the successful civil rights claims, but also on other claims involving a 'common core of facts' or 'related legal theories,' " and, therefore, a "plaintiff should receive significant fees when he has won a partial victory on a civil rights claim while receiving substantially the relief he there sought, though the jury awards it on a factually or legally related pendent state claim." Aubin v. Fudala, 782 F.2d 287, 291 (1st Cir.1986) (quoting Hensley, 461 U.S. at 435, 103 S.Ct. at 1940).10 Here, where the elements of the various claims under Puerto Rico law are subsumed by the Title VII claim, the claims are unquestionably interrelated. Hence, the fact that the appellant also recovered under Puerto Rico law is irrelevant vis-a-vis her section 1988 recovery.

Second, as a general matter, a plaintiff who prevails on congruent federal and state claims and qualifies for fee-shifting under two or more statutes may recover fees under whichever fee-shifting regime she chooses. See Freeman v. Package Mach. Co., 865 F.2d 1331, 1347 (1st Cir.1988). To constrain the plaintiff's choice would withhold from her the deserved fruits of her victory and would discourage potential claimants from redeeming their civil rights.

We need go no further. The court below offered no plausible reason for eschewing the lodestar method, and no such reason springs spontaneously from the record. It was, therefore, error to forgo the lodestar. In addition, the court relied on impermissible criteria in making its non-lodestar fee award. Consequently, we vacate the order appealed from and remand for further proceedings consistent with this opinion. Costs on appeal shall be taxed in favor of the appellant. Upon the timely filing of a supplemental application in suitable form, the district court shall include in its new fee award a sum sufficient to compensate the appellant's counsel for services rendered in connection with the successful prosecution of this appeal.

It is so ordered.