OPINION OF THE COURT
GIBBONS, Circuit Judge.
This appeal is before us pursuant to the Interlocutory Appeals Act of 1958, 28 U.S.C. § 1292(b). The appellant Carte Blanche Corporation (Carte Blanche) seeks review of an order of the district court which granted the pretrial motion of the plaintiff Katz that his suit be maintained as a class action under Rule 23(b)(3) of the Federal Rules of Civil Procedure and directed that notice be sent to the alleged members of the class.
Carte Blanche is a national credit card company. Its members pay an annual .membership fee of $15.00 which gives them the privilege of charging items at the company’s associated establishments. Carte Blanche purchases the charge slips from the associated establishments at a discount, and invoices the members for the face amount. Katz, who after intervening was substituted for a previously named plaintiff, Stept, seeks statutory damages and costs of suit, including a reasonable attorneys fee, for Carte Blanche’s alleged failure to make pretransaction and transaction disclosures regarding the computation of finance charges to its members in violation of the Truth in Lending Act (TILA), 15 U.S.C. §§ 1601-1681L The three allegedly improperly disclosed finance charges are:
(1) The $15 annual membership charge;1
(2) The late charge assessed on overdue unpaid balances ;2 and
(3) The finance charge on an extended payment plan for purchase of airline tickets.
TILA’s disclosure provisions apply only to extensions of consumer credit. Thus persons using a Carte Blanche credit card for business are not protected by the Act from the allegedly inadequate disclosures. Katz claims to be a nonbusiness card user, and to represent all other [751]*751such users. He seeks recovery pursuant to 15 U.S.C.A. § 1640(a):
“Except as otherwise provided in this section, any creditor who fails in connection with any consumer credit transaction to disclose to any person any information required under this part to be disclosed to that person is liable to that person in an amount equal to the sum of
(1) twice the amount of the finance charge in connection with the transaction, except that the liability under this paragraph shall not be less than $100 nor greater than $1000; and
(2) in the case of any successful action to enforce the foregoing liability, the costs of the action together with a reasonable attorney’s fee as determined by the court.”
Katz, or any other card holder, in order to establish Carte Blanche’s liability under § 1640(a), must show: (1) that a given charge was a finance charge within the meaning of TILA; (2) that the disclosure with respect to that charge was not in compliance with the applicable TILA regulations; and (3) that the charge was imposed in connection with a consumer credit transaction, not a business transaction. The first two showings focus on the uniform conduct of Carte Blanche. The third showing focuses on the separate conduct, and possibly the subjective intention, of each Carte Blanche member. In the district court Katz moved for a determination that the case proceed as a class action in which he represented a class consisting of all Carte Blanche credit card holders, since all such holders had been charged the $15 annual membership fee, and such a class would include all who had been assessed an overdue payment charge or an extended payment airline ticket charge. That class would include somewhere between 717,000 and 800,000 members. If Katz succeeded in establishing liability to every class member for the minimum recovery specified in § 1640 (a)(1) the liability would be substantially in excess of Carte Blanche’s net worth. The record does not establish how many would be included in a class restricted to members who had been assessed late charges since the effective date of TILA. Nor does it establish the number of persons who have paid extended payment charges for airline tickets, although Katz estimates this number at 40,000.
The district court first considered whether Katz’ complaint satisfied the prerequisites of Fed.R.Civ.P. 23(a). The court readily concluded that the class was too numerous for joinder of all members, that there were questions of law or fact common to the class, and that Katz would fairly and adequately represent the class. Thus three of the prerequisites of rule 23(a) were satisfied. In examining the fourth prerequisite, “the claims ... of the representative parties are typical of the claims ... of the class,” Fed.R. Civ.P. 23(a)(3), the court was aware that each class member in order to recover would have to show that the transaction on which he based his claim was a consumer and not a business transaction. But concluding that Katz’ claims •were coincident and not in conflict with those of the class, it found this requirement was satisfied as well. The district court had doubts, initially, that it could make the findings required by rule 23(b)(3) that questions of law or fact common to the class predominate over questions affecting only individual members and that a class action is superior to other available methods for the fair and efficient adjudication of the controversy. See Katz v. Carte Blanche Corp., 52 F.R.D. 510 (W.D.Pa.1971). After further briefing and argument, however, the district court, over Carte Blanche’s strenuous objection, overcame its doubts and entered an order (1) granting Katz’ motion that the action proceed as a class action on behalf of all authorized holders of one of Carte Blanche’s credit cards since July 1, 1969, the effective date of TILA, and (2) directing that a notice of the pendency of the action be sent to [752]*752each class member, the initial cost, estimated at |37,500, to be borne by Katz. See Katz v. Carte Blanche Corp., 53 F. R.D. 539 (W.D.Pa.1971). The form of notice is reproduced at 53 F.R.D. 547. The district court simultaneously certified that the' order involves a controlling question of law as to which there is a substantial ground for a difference of opinion and that an immediate appeal from the order may materially advance the ultimate termination of the litigation. A panel of this court permitted an appeal to be taken.
Appealability of Rule 23(b)(3) Orders
We are met at the outset with the contention that the appeal should be dismissed because permission to appeal pursuant to 28 U.S.C. § 1292(b) was improvidently granted, in that an appeal under that section is not appropriate to review a district court’s favorable class action determination under rule 23(a) and 23(b)(3). That contention is based upon a misapprehension of the intended purpose of the Interlocutory Appeals Act and a misunderstanding of the role of the district court judge in granting or denying leave to proceed as a class action under rule 23(b) (3).
A class action determination, affirmative or negative, is not in this circuit a final order appealable under 28 U.S.C. § 1291. Hackett v. General Host Corp., 455 F.2d 618 (3d Cir.), cert. denied, 407 U.S. 925, 92 S.Ct. 2460, 32 L.Ed.2d 812 (1972). Compare id. with Eisen v. Carlisle & Jacquelin, 370 F.2d 119 (2d Cir. 1966), cert. denied, 386 U.S. 1035, 87 S.Ct. 1487, 18 L.Ed.2d 598 (1967). If a district court denies class action treatment, even on discretionary grounds, and certifies, pursuant to rule 54(b), that the order dismissing the complaint against the absent class members is a final judgment, such an order is appealable. See Hayes v. Sealtest Foods Division of National Dairy Products Corp., 396 F.2d 448 (3d Cir. 1968). But rule 54(b) does not provide a means for reviewing the grant of class action treatment. If the court in entering a class action order acts outside its jurisdiction, Shutte v. Armco Steel Corp., 431 F.2d 22 (3d Cir. 1970), cert. denied, 401 U.S. 910, 91 S.Ct. 871, 27 L.Ed.2d 808 (1971), or in disregard of appropriate procedural safeguards, Swindell-Dressler Corp. v. Dumbauld, 308 F.2d 267 (3d Cir. 1962), the order may be reviewable by mandamus. But if the court has acted within its jurisdiction pursuant to appropriate procedural safeguards and in a nonarbitrary manner, mandamus will not lie. Solomon v. Continental American Life Insurance Co., 472 F.2d 1043, 1046 (3d Cir. 1973); Hackett v. General Host Corp., supra, 455 F.2d at 626; Weight Watchers of Philadelphia, Inc. v. Weight Watchers International, Inc., 455 F.2d 770, 775 (2d Cir. 1972); Interpace Corp. v. City of Philadelphia, 438 F.2d 401, 404 (3d Cir. 1971). Thus if there is any route open for the interlocutory review of the grant of class action treatment under rule 23(b)(3) in this circuit, it is only pursuant to 28 U.S.C. § 1292(b).
We have, in Kauffman v. Dreyfus Fund, Inc., 434 F.2d 727, 734 (3d Cir. 1970), cert. denied, 401 U.S. 974, 91 Ct. 1190, 28 L.Ed.2d 323 (1971), in an appeal pursuant to § 1292(b), reviewed and reversed the grant of class action treatment. See also Wilcox v. Commerce Bank of Kansas City, 474 F.2d 336 (10th Cir. 1973), Zahn v. International Paper Co., 469 F.2d 1033 (2d Cir. 1972), aff’d, 414 U.S. 291, 94 S.Ct. 505, 38 L.Ed.2d 511 (1973), and Johnson v. Georgia Highway Express, Inc., 417 F.2d 1122 (5th Cir. 1969), affirming the denial of class action motions on § 1292 ' appeal. Obviously, then, some class action determinations are reviewable under § 1292(b).
Both Kauffman v. Dreyfus Fund, Inc. and Zahn v. International Paper Co. are arguably distinguishable from this case, 1 however, since both involved nondiscretionary reasons for rejecting class action treatment. In Kauffman the plaintiff, not being a member of the class, could not be a class representative. See 434 F.2d at 734. In Zahn the issue was [753]*753whether each of the class members had to meet the jurisdictional amount requirement to bring his claim within the district court’s subject matter jurisdic-' tion. The contention against appealability in this case is that because the district court’s decision allowing the case to proceed as a class action in contrast with Kauffman and Zahn involved the exercise of discretion, an appeal under § 1292(b) should not lie. Two circuit courts have in § 1292(b) appeals reviewed those aspects of a district court’s class action determination which involve some exercise of discretion. Wilcox v. Commerce Bank of Kansas City, supra; Johnson v. Georgia Highway Express, Inc., supra. In fact in Wilcox, without discussing the appropriateness of its review, the Eighth Circuit examined one of the issues before us, the superiority of a rule 23(b) (3) class action for adjudication of a TILA claim. It found that the district court acted within its sound discretion in denying the motion.
The source of the contention that discretionary aspects of a rule 23 determination are not reviewable in a § 1292(b) appeal may be traced to language in cases such as Standard v. Stoll Packing Corp., 315 F.2d 626 (3d Cir. 1963), holding that a grant of or refusal to grant a change of venue authorized by 28 U.S.C. § 1404(a) will not be reviewed under § 1292(b). See A. Olinick & Sons v. Dempster Bros., Inc., 365 F.2d 439, 443 (2d Cir. 1966); cf. Akers v. Norfolk & Western Railway, 378 F.2d 78, 79 (4th Cir. 1967). Some commentators, relying on the change of venue authorities, have suggested that matters entrusted to the discretion of the district court are not properly reviewable in a § 1292(b) appeal. See, e. g., 9 Moore’s Federal Practice If 110.22[2], at 261 (2d ed. 1973); C. Wright, Law of Federal Courts § 102, at 463 (2d ed. 1970). But to extrapolate from the change of venue cases, or even from the commentators’ discussion of those cases, a general rule that no district court order involving some exercise of discretion may be reviewed in a § 1292(b) appeal is to misread that statute and to disregard its legislative history.
Section 1292(b) was the result of dissatisfaction with the prolongation of litigation and with harm to litigants uncorrectable on appeal from a final judgment which sometimes resulted’ from * strict application of the federal final judgment rule. In the federal courts prior to 1891 no interlocutory appeals ; were possible. In that year Congress authorized interlocutory appeals from orders granting or continuing injunctions. Act of March 3, 1891, ch. 517, § 7, 26 Stat. 828. In 1895 it authorized interlocutory appeals from the denial of injunctive relief. Act of Feb. 18, 1895, ch. 96, 28 Stat. 666 (repealed, Act of June 6, 1900, ch. 803, 31 Stat. 660; reenacted, Act of March 3, 1911, § 129, • 36 Stat. 1134). In 1900 it included orders in receivership cases. Act of June 6, 1900, ch. 803, 31 Stat. 660. In 1928 it included orders determining liability in admiralty cases prior to the assessment of damages. Act of April 3, 1926, ch. 102, 44 Stat. 233. In 1927 it included orders adjudging infringement in patent cases before an accounting. Act of Feb. 28, 1927, ch. 228, 44 Stat. 1261.3 These four exceptions to the final judgment rule are included in § 1292(a). All four were in effect when the Judicial Conference of the United States began its consideration of broadening the availability of interlocutory appellate review. That consideration began with a suggestion by Judge Jerome Frank to the 1951 meeting of the Judicial Conference of the United States, which referred the matter to a committee chaired by Judge John J. Parker. Although Judge Frank’s particular proposal was rejected, after extensive investigation the committee reported in favor of a draft statute in the form of § 1292(b) and the Judicial Conference approved the report at [754]*754its 1953 meeting. The draft statute was transmitted to the House of Representatives in 1957. See 3 U.S.Code Cong. & Admin.News, 85th Cong., 2d Sess., pp. 5258-5263 (1958). Judge Parker and Judge Albert B. Maris testified in its support. See Hearings Before Subcommittee No. 3 of the House Committee on the Judiciary on H.R.6238', 85th Cong., 2d Sess., ser. 11 (1958) (hereinafter House Hearings). Since the bill was enacted exactly as submitted, the Judicial Conference Committee Report and the testimony of representatives of the Judicial Conference before Congress as to what they were attempting to accomplish are the best indication of the legislative intention.
In assessing both the text of § 1292(b) and its legislative history, it must be kept in mind that the draftsmen were not writing on a blank slate. Congress had already, in the four exceptions now embodied in § 1292(a), expressed certain value judgments as to when there should be a departure from the final judgment rule. In the case of grants or denials of injunctive relief or grants or denials of applications for the appointment of receivers, it had focused upon the problem of serious harm to the litigant pendente lite from an erroneous interlocutory order. In the admiralty and patent exceptions it had focused upon avoiding the wasted effort of a possibly protracted litigation over damages where there might be no liability.4 Thus the draftsmen knew what considerations had historically moved Congress toward relaxation of the final judgment rule. Those historical categories obviously bear upon the interpretation of § 1292(b). The section probably was intended to include orders having similar potential for harm to the litigant pendente lite or similar potential for causing a wasted protracted trial if it could early be determined that there might be no liability. That at least the latter potential was a major concern of the draftsmen is borne out by the examples given in support of the bill by its proponents. They referred specifically to eases in which a long trial" results from a pretrial order erroneously overruling a defense going to the right to maintain the action, to cases involving prolonged assessment of damages after determination of liability, to cases where the disposition of motions for impleader might induce voluntary nonsuit or settlement, and to cases where venue is claimed to have been transferred without proper authority. House Hearings at 8, 9.
The statute imposes three criteria for the district court’s exercise of discretion to grant a § 1292(b) certificate. The order must (1) involve a “controlling question of law,” (2) offer “substantial ground for difference of opinion” as to its correctness, and (3) if appealed immediately “materially advance the ultimate termination of the litigation.” Significantly, while the district judge must certify that the order satisfies the three criteria, the discretion to grant leave to appeal at the circuit level is not limited by any specific criteria. Judge Maris, testifying in support of the bill, analogized the appellate court’s discretion to that of the Supreme Court on certiorari. House Hearings at 21. Denial of permission to appeal may be based upon a different assessment than that of the district court as to any of the three criteria. But leave to appeal may be denied for entirely unrelated reasons such as the state of the appellate docket or the desire to have a full record before considering the disputed legal issue. -And once leave to appeal is granted the court of appeals is not restricted to a decision of the question of law which in the district judge’s view was controlling. Johnson v. Alldredge, 488 F.2d 820 (3d Cir. 1973).
There can be little difficulty over the criterion of difference of opinion as to correctness of the order. The likelihood is remote that a district judge would [755]*755make such a certification frivolously with respect to his own order. Certainly the instant case involves an order over which a difference of opinion might exist since it took two district court opinions to arrive at a decision. Nor can there be much difficulty over the requirement of a likelihood of materially advancing the ultimate termination of the litigation. The district court’s opinion about settlement possibilities, about the potential length of a possibly avoidable trial, and similar matters, will in most cases,' including this case, be at least as informed as that of a panel of this court. The “controlling question of law” requirement is the source of whatever difficulty may arise as to the propriety of an interlocutory appeal.
A controlling question of law must encompass at the very least every order which, if erroneous, would be reversible error on final appeal. If the statute were interpreted to exclude any such order that interpretation would be inconsistent with the clear intention of the sponsors to avoid a wasted trial. Nor need the order be determinative of any plaintiff’s claim on the merits, since a dismissal for want of jurisdiction is within § 1292(b). Zahn v. International Paper Co., supra; S.Rep.No.2434, 85th Cong., 2d Sess. 2 (1958) (reprinted in 3 U.S.Code Cong. & Admin.News, p. 5256 (1958)). Nor need a reversal of the order terminate the litigation, since with impleader and transfer of venue orders, two of the sponsors’ examples,5 the lawsuit could continue regardless of the interlocutory determination although an erroneous decision might cause a reversal on appeal from a final order. What remains is the question whether in order for a question to be “controlling” must it be one which if decided erroneously would lead to a reversal on appeal? Certainly Judge Maris, testifying in favor of the bill on behalf of the Judicial Conference, did not think so. His testimony suggests that “controlling” means serious to the conduct of the litigation, either practically or legally. House Hearings at 21; see Note, Discretionary Appeals of District Court Interlocutory Orders: A Guided Tour through Section 1292(b) of the Judicial Code, 69 Yale L. J. 333, 343 (1959). And on the practical level, saving of time of the district court and of expense to the litigants was deemed by the sponsors to be a highly relevant factor. See S.Rep.No.2434, 85th Cong., 2d Sess. 2 (1958) (reprinted in 3 U.S.Code Cong. & Admin.News, p. 5256 (1958)).
The clear case of a controlling question of law — one which would result in a reversal of a judgment after final hearing — serves to focus upon the ' limited precedential value of statements about the exercise of discretion appearing in § 1292(b) cases dealing with orders granting a change of venue. .If an order changes venue to a place where there would be no jurisdiction, no discretion is involved and the certain reversal of the judgment on jurisdictional grounds after trial should be avoided. See 28 U.S.C. § 1404(a); cf. Swindell-Dressler Corp. v. Dumbauld, supra. If there are two districts, either of which would have jurisdiction, but one of which is in the opinion of the district court superior, it is extremely unlikely that a change of venue would result in a reversal after final judgment. As to most such orders it matters little whether one says that they do not decide a controlling question of law or that their review will not materially advance the litigation. However they are decided the litigation will go forward and a resulting final judgment will not, on the basis of the order, be disturbed on appeal. But a district court could so- abuse the discretion vested in it by 28 U.S.C. § 1404(a) as to commit error reversible after final judgment. It could, for example, in a ease in which credibility of witnesses was crucial, transfer to a venue in which the critical witness on one side was beyond the reach of a subpoena [756]*756and available for testimony only by deposition. That would be an exercise of discretion. It might also be a ground for the grant of a new trial after final hearing. Obviously such a result should be preventable by a'§ 1292(b) appeal if the district judge anticipating, the problem certified the change of venue order for interlocutory appeal. The key consideration is not whether the order involves the exercise of discretion, but whether it truly implicates the policies favoring interlocutory appeal. The determination of what orders are properly reviewable under § 1292(b) must be made by a practical application of those policies, not by a mechanical application of labels such as “discretionary” or “nondiscretionary.” Those policies, both before and since the enactment of § 1292(b) have included the avoidance of harm to a party pendente lite from a possibly erroneous interlocutory order and the avoidance of possibly wasted trial time and litigation expense. In this case, as subsequent discussion will disclose, there exists by virtue of the order appealed from both the possibility of prejudice to a party pendente lite and the possibility of considerable avoidable wasted trial time and litigation expense.
Commentators on the rule 23 amendments support the view that appeals certified pursuant to § 1292(b) are appropriate for testing whether suitability of class action treatment has been correctly determined. See Kaplan, Continuing Work of the Civil Committee: 1966 Amendments of the Federal Rules of Civil Procedure (I), 81 Harv.L.Rev. 356, 390 n.131 (1967); Report of the American Bar Association Special Committee on Federal Rules of Procedure, 38 F.R.D. 95, 104-105 (1965). As the foregoing discussion indicates, we agree with that view. The order appealed from is properly before us.
Scope of Review
There remains the question of scope of review of rule 23 class action determinations. That will depend to some extent upon which facet of the multi-faceted determination is challenged on appeal. The district court must determine if the four prerequisites for a class action listed in rule 23(a) have been met. These are mandatory requirements, and our review decides whether the mandates have been met. E.g., Kauffman v. Dreyfus Fund, Inc., supra; cf. Zahn v. International Paper Co., supra. The district court must also determine whether the class action is maintainable under rule 23(b)(1) or (2) so that it may proceed without notice to or identification of the class members. Our review will determine whether the -court properly classified the type of class action in reaching its decision as to class action treatment. E.g., Yaffe v. Powers, 454 F.2d 1362, 1366 (1st Cir. 1972). If the class action is maintainable only under rule 23(b)(3) the district court must make two additional findings: (1) that the questions of law or fact common to the members of the class predominate over any questions affecting only individual members, and (2) that a class action is superior to other available methods for the fair and efficient adjudication of the controversy. Both findings require the exercise of an informed judgment as to the application of defined legal standards.
The predominance finding requires at a minimum the identification of the legal and factual issues, common and diverse, and an identification of the class members to which those relate. We must determine whether the district court has properly identified the factual or legal issues, and has properly identified those which are common. If the district court has properly identified the issues common and diverse, we would undoubtedly defer in most instances to its conclusion as to predominance, since that requirement relates to the conservation of litigation effort, and the trial court’s judgment probably will be as good as ours. If the district court has applied the correct criteria to the facts of the case, then, it is fair to say that we will ordinarily defer to its exercise of discretion. But if it has not properly iden[757]*757tified the issues, and not properly evaluated which are common, the order is not entitled to such deference.
The superiority finding requires at a minimum (1) an informed consideration of alternative available methods of adjudication of each issue, (2) a comparison of the fairness to all whose interests may be involved between such alternative methods and a class action, and (3) a comparison of the efficiency of adjudication of each method. Since, as will be pointed out hereinafter, a single ease resulting in a judgment which will be given collateral estoppel effect against a losing defendant is in most rule 23(b)(3) class actions an alternative available method, the comparative fairness and comparative efficiency of that method must be considered. As with the determination of predominance, our review looks first at whether the district court properly applied the relevant criteria to the facts of the case. If this has been done it is fair to say that we will ordinarily defer to its exercise of discretion. But if, for example, the district court has in rejecting alternative available methods of adjudication disregarded possible unfairness of the class action to a particular defendant, its determination is not entitled to such deference. See Wilcox v. Commerce Bank of Kansas City, 474 F.2d at 339 & n. 7.
Professor Moore summarizes the scope of review:
“In determining whether an action -brought as a class action is to be so maintained the trial court should carefully apply the criteria, set forth in Rule 23 . . ., to the facts in the case; and if it fails to do so its determination is subject to reversal by the appellate court when the issue is properly before the latter court. On the other hand, where the trial court does apply the Rule’s criteria to the facts of the case, the trial court has broad discretion in determining whether the action may be maintained as a class action and its determination should be given great respect by a reviewing court.” 3B Moore’s Federal Practice fl 23.50 at 1104-05 (2d ed. 1969).
Cf. Greenfield v. Villager Industries, Inc., 483 F.2d 824 (3d Cir. 1973).
Carte Blanche’s Objections to
Class Action Treatment Carte Blanche urges that the district court erred both in its finding of superiority and in its findings of predominance. It points out, and Katz concedes,6 that the claim that Carte Blanche’s $15.00 annual fee is an improperly disclosed finance charge is of extremely doubtful validity. Yet the order appealed from directs that notice of this claim be sent to all of its card holders with the suggestion that they may be able to recover the $100 minimum recovery specified in 15 U.S.C. § 1640(a)(1). Carte Blanche points out that almost all of its card holders are also its account debtors, and it foresees the effect, possibly catastrophic to it, that a substantial number of the account debtors will upon receipt of the notice withhold payments. It points out, as well, that at any given time a good many of its account debtors are delinquent entirely apart from delinquency induced by a class action notice. As to [758]*758these account debtors, if after notice they fail to opt out, Carte Blanche will be required under Rule 13(a) of the Federal Rules of Civil Procedure to file a compulsory counterclaim. This course, it urges, would be disruptive of its normal collection practices and of its relations with its account debtors. All those disruptive effects, it urges, will take place even if it prevails in establishing its own full compliance with TILA. It urges that the district court, in making its superiority determination, disregarded these substantial elements of unfairness to it. And the court also disregarded the fact that Congress has afforded to individual litigants a remedy of attorneys fees if they are relegated to individual lawsuits, thus making such relegation not unfair to the individual class members. Moreover, it contends the district court erred in making its predominance finding, since as to each alleged class member it will be entitled to a hearing on the fundamental liability issue of whether the challenged finance charge was imposed in connection with a consumer rather than a business transaction. These individual factual issues, it is urged, heavily predominate over the legal and factual issues relating to Carte Blanche’s conduct.
The Test Case as an Available Alternative Method.
In rejecting Carte Blanche’s contention that notice should not be given to its account debtors until its noncompliance with TILA disclosure requirements had been established, the district court discussed and rejected the test case as an available alternative, writing:
“Some confusion surrounded the meaning of the term test case. A classic test case with binding legal effects on all those affected by its outcome is not possible, of course, without the express consent of all those affected, both parties and non-parties. In the case sub judiee in which the obtaining of the express consent of non-nominal parties is an impracticability, the procedural device by which to bind all non-nominal parties to the outcome of the case is the class action.3 The term test case as applicable here, therefore, means a test case for practical purposes. In other words, if the plaintiff’s motion for determination of a class action is denied the instant action becomes, in practical effect, a case not legally binding upon anyone other than Mr. Katz, and it is the superiority of that alternative which we must evaluate.
There is in Kahan v. Rosenstiel, 424 F.2d 161, 169 (3d Cir. 1970), language similar to the district court’s footnote 3, but in the footnote it is taken out of context. We said in Kahan v. Rosenstiel that an alleged class action should be treated as such for purposes of dismissal or compromise until there is a full determination that a class action is not maintainable. We did not, as the district judge apparently believed, rule in Kahan v. Rosenstiel on the question when in a rule 23(b)(3) case the class action determination must be made. Rule 23(c) (1) provides that the determination should be made as soon as practicable after the commencement of the action. But “[a]s soon as practicable” must mean as soon as practicable in the light of the relevant rule 23(b)(3) factors of predominance and superiority. Where the defendant' seeks a postponement of that determination on the ground of fairness, as soon as practicable does not necessarily mean at the outset of the lawsuit.
The district court points out that if Katz’ case went forward alone it would not be binding upon anyone other than Mr. Katz. That is only partly true. Judgment against Katz would not protect Carte Blanche against other class [759]*759members, but judgment for Katz would bind Carte Blanche in suits by other class members. In making its superiority determination the district court overlooked that factor, and disregarded the fact that Carte Blanche was content to take its chances on stare decisis rather than res judicata. No reference was made to the Supreme Court's decision in Blonder-Tongue Laboratories, Inc. v. University of Illinois Foundation, 402 U.S. 313, 91 S.Ct. 1434, 28 L.Ed.2d 788 (1971), or to the well-developed jurisprudence in this circuit flowing from Judge Hastie’s well-known decision in Bruszewski v. United States, 181 F.2d 419 (3d Cir.), cert. denied, 340 U.S. 865, 71 S.Ct. 87, 95 L.Ed. 632 (1950). To understand the relevance of Blonder-Tongue and Bruszewski to the superiority finding one must start with a consideration of case law under the old rule 23. Prior to the 1966 amendments, cases which would now be covered by rule 23(b)(3) were commonly referred to as spurious. In York v. Guaranty Trust Co., 143 F.2d 503, 529 (2d Cir. 1944), rev’d on grounds not here relevant, 326 U.S. 99, 65 S.Ct. 1464, 89 L.Ed. 2079 (1945), Judge Frank for the Second Circuit and in Pentland v. Dravo Corp., 152 F.2d 851, 856 (3d Cir. 1945), Judge Goodrich for this court suggested that it would be proper to make the class action determination and permit class members to intervene after the defendant’s liability had been determined in the single lawsuit. In Union Carbide & Carbon Corp. v. Nisley, 300 F.2d 561, 588-589 (10th Cir. 1961), petition for cert, dismissed, 371 U.S. 801, 83 S.Ct. 13, 9 L. Ed.2d 46 (1963), the Tenth Circuit held that such “one-way” intervention after a judgment adverse to the defendant was proper. Many commentators objected that one-way intervention had the effect of giving collateral estoppel effect to the judgment of liability in a case where the estoppel was not mutual. This was thought to be unfair to the defendant. To meet the point'-that one-way intervention was unfair to the defendant, the Advisory Committee on the Federal Rules concluded that class members should be brought in prior to the determination of defendant’s liability, thus making the estoppel mutual. See Advisory Committee’s Note to Proposed Amendments to Rule 23, 39 F.R.D. 69, 105 (1966); Kaplan, Continuing Work of the Civil Committee: 1966 Amendments of the Federal Rules of Civil Procedure (I), 81 Harv.L.Rev. 356, 385 (1967). To make joinder at an early stage practically achievable, the “opting out” mechanism was devised. The literature on the development of rule 23(b)(3) makes it quite clear that the early notice requirement was directly related to dissatisfaction with the lack of mutuality of the estoppel which resulted from the York, Pentland, Nisley rule.
Of course the 1966 amendment to rule 23(b) (3) was also based upon the underlying assumption that class action disposition of multiple similar claims, especially small consumer claims, is socially desirable. But the socially desirable effects, looked at from the point of view of anyone other than the defendant, were as achievable under the York, Pentland, Nisley rule as under a rule prohibiting one-way intervention. A major social policy in favor of the § 23(b)(3) class action is the prevention of multiple relitigation of the factual and legal issues as to a defendant’s liability. The Nisley result accommodated that policy perfectly in any case where liability was established. In cases where the defendant prevailed, other class members were, of course, free to try again. But practically speaking, stare decisis provided a substantial deterrent in most instances. Another major social policy in favor of the § 23(b)(3) class action is its encouragement of private enforcement on behalf of small claimants. Again the Nisley result accommodated that policy, since notice could be given to all potential claimants once the defendant’s liability was established. Moreover a fee award could be made in light of the value of the benefit conferred on the entire class. Lindy Bros. Builders, Inc. v. American [760]*760Radiator & Standard Sanitary Corp., 487 F.2d 161 (3d Cir. 1973); Kahan v. Rosenstiel, supra. The objection to the Nisley result was not any conflict with the policies favoring class action litigation, but a conflict with Bigelow v. Old Dominion Copper Mining & Smelting Co., 225 U.S. 111, 32 S.Ct. 641, 56 L.Ed. 1009 (1912). That case had announced “[i]t is a principle of general elementary law that the estoppel of a judgment must be mutual.” Id. at 127, 32 S.Ct. at 642. Judge Hastie by 1950 in Bruszewski did not think the principle was quite so elementary as it had seemed in 1912. In 1970 the Supreme Court unanimously rejected the authority of the Bigelow case. Blonder-Tongue v. University of Illinois, supra. Ratio est legis anima; mulata legis ratione mutatur et lex.
In this case Carte Blanche seeks not to have conferred upon it what it perceives to be the dubious benefit of mutual estoppel. It is perfectly willing to run the risk that if it loses the class may later be enlarged rather than, as Katz suggests, contracted.7 We think the Blonder-Tongue decision requires that a new look be taken at the alternative of a test case in lieu of an early class action determination. That new look must, as with all rule 23(b)(3) superiority determinations, take into account several different interests. Superiority must be looked at from the point of view (1) of the judicial system, (2) of the potential class members, (3) of the present plaintiff, (4) of the attorneys for the litigants, (5) of the public at. large and (6) of the defendant. The listing is not necessarily in order of importance of the respective interests. Superiority must also be looked at from the point of view of the issues.
Turning to the issues in this case there is legally and factually a clear line between the question whether Carte Blanche has made the disclosures required by TILA and the question whether any given card holder is a consumer or a business user. Whether or not class action treatment is given at the outset, the judicial system will not, since the Blonder-Tongue case, be required to relitigate after a finding that Carte Blanche did not make the required disclosures. If there is a finding for the defendant on liability, the judicial system will be protected from relitigation of the essentially legal issue of compliance with TILA regulations by the discouragement of relitigation afforded by stare decisis. Where, as here, the defendant makes a nonfrivolous claim that it will be prejudiced by the mailing of notice, it is hardly fair to say that the judicial system must insist on res judicata rather than collateral estoppel or stare decisis. Especially is this so when it is recalled, that the plaintiff, not the judicial system, controls whether or not to ask for class action treatment. As to the issues relating to consumer versus business use, a postponement of those issues until violation is decided will actually protect the judicial system from time and expense which might have been wasted on them in a nonliability case. Moreover the judicial system may, by postponing class action treatment until a violation has been proved, avoid whatever time and expense it, as distinguished from the parties, might have incurred in the preparation of the notice.
Postponement of class action determination until the violation has been proved in no way prejudices the potential class members. If Katz loses his case on violation they will not be bound. If Katz establishes the violation they can be afforded the same opting out option, but the notice will advise them that there is a judgment establishing violation, and their decision will be more informed than if the notice was sent early in the proceedings. The only real potential for harm to the absent class mem[761]*761bers, the running of the statute of limitations, once thought to be a possible problem, has been eliminated by the holding in American Pipe & Construction Co. v. Utah, 414 U.S. 538, 94 S.Ct. 756, 38 L.Ed.2d 713 (1974), that the filing of a class action complaint tolls the running of the statute of limitations for the entire class. The Court has thus adopted the same holding of the Tenth Circuit in the Nisley case. It might be suggested that the seventh amendment gives the absent class members a right to a unitary trial as to all issues of liability and damages before a single jury. This can hardly be the case as to the absent class members, however, since it is under rule 23 discretionary with the court whether they will be included in the case at all. If the seventh amendment presents any problem at all, it is a problem related to the rights of the defendant. And more practically, even the order appealed from contemplates a non-unitary determination of liability and damage issues relating to the individual class members.
As to the original plaintiff, Katz, he is not affected with respect to the amount of his recovery or with respect to the trial of the case. He may be protected from the expense, in this case $37,500, of mailing out a notice to parties who have no rights. If he first establishes a violation, the district court may conclude that the defendant rather than he should bear the initial cost of the mailing. The only possible adverse effect upon him of postponing the class action determination until after he proves a violation is that it may be more difficult for him to obtain an attorney willing to pursue his claim. Further reference will be made to the effect of postponing the class action determination upon the willingness of attorneys to undertake such cases in the next paragraph. Here, we point out that in TILA Congress has provided for the award of reasonable attorneys fees. 15 U.S.C. § 1640(a)(2). It seems unlikely that actions under TILA or any other federal regulatory statute providing for the award of reasonable attorneys fees are likely to go begging for want of an attorney.
We have no doubt that the maximization of potential recoveries in a single lawsuit which has resulted from the 1966 amendments to rule 23 has produced the desirable effect of interesting outstanding attorneys in the pursuit of consumer protection lawsuits. The encouragement of that interest is a legitimate consideration in the district courts’ administration of rule 23(b)(3). It is not, however, the sole consideration. Fairness is an explicit criterion for a superiority determination, and fairness to the defendant must be balanced against any disincentive for class action litigation which might result from the possibility of postponing the class action determination until a violation has been proved. %The major incentive to the bar is the possibility of aggregating small claims. That incentive is no different, since the same class members will be opted in once a violation .has been proved, and the attorney’s fee will be measured to the same extent as heretofore by the benefit conferred upon the class. Lindy Bros. Builders, Inc. v. American Radiator & Standard Sanitary Corp., supra; Kahan v. Rosenstiel, supra. Another incentive to the bar is the increased settlement leverage which results from aggregating small claims. That incentive is no different, for the potential for aggregation if a violation is proved still remains, and the defendants will be under no less compulsion to make a settlement offer to the class, when confronted with a legitimate claim of violation, than in the past. Indeed we have observed that not infrequently settlement negotiations do precede class action determinations. See Greenfield v. Villager Industries, Inc., supra. What the attorneys would be deprived of by postponement of class action determination until a violation has been proved is the additional settlement leverage which results from the disruption or injury which may occur to a defendant’s business relationships regardless of the mer[762]*762its of the claim by the mere sending of the notice. That leverage is said by the instant defendants to be considerable. It is, we think, less than a legitimate use of a procedural device. If its absence provides some disincentive to attorneys’ pursuit of class actions, that disincentive must be tolerated in the interest of fairness.
The public at large has an interest in the fact that the class action device provides supplementary enforcement of regulatory statutes. What we have said about disincentives to attorneys’ pursuit of those actions applies as well to the public interest. There is incentive enough in the possibility of aggregating small claims without adding to the leverage resulting from aggregation any untoward effects resulting from notice in possibly groundless actions.
Finally, turning to the defendant’s interest, it must be understood that we are dealing only with the defendant who declines the protection against one-way intervention after a violation has been proved which rule 23(b)(3) was designed to afford. If a class action defendant insists upon early class action determination and notice, he is, under the rule, entitled to it. But where he makes a nonfrivolous claim that his business will be harmed or disrupted by the notice, and is willing to run the risk that the determination of liability, if he loses, will be given effect in favor of the class, with notice in the event of such determination, the district court must seriously consider that alternative, and should, absent other compelling circumstances, pursue that course. A defendant taking such a position would, of course, be deemed to have waived any seventh amendment right to a unitary trial before a single jury on all issues, and would realize that the statute of limitations was tolled in favor of the class. He would lose, as well, the collateral estoppel effect against the class of a favorable judgment on liability.8
Manageability.
As a separate ground of appeal Carte Blanche contends that the district court erred in concluding that it could manage the problem of determining, separately as to each, which of 800,000 potential class members were nonbusiness credit card users. Although the literature and case law under rule 23 often refer to a finding of “manageability” as a separate requirement for a class action determination, it is actually one of the nonexclusive factors which is to be taken into account in making the required findings of predominance and superiority. We have concluded that the order appealed from must be reversed because the superiority finding was in error. With the procedure we have suggested, the class, if any, ultimately to be included and receive notice may be considerably smaller than 800,000. The management problem later in the case may be considerably different than appears at this juncture since at a later stage the content of the notice may be considerably more specific. If suffices to say at this point that determining which of 800,000 claimants was a consumer rather than a businessman, if the defendant chooses, as it may, to insist on proof as to each, appears on the surface to present formidable management problems and will undoubtedly be reconsidered if the district court concludes that [763]*763Katz has proved all three claimed violations.9
Class Actions and the Truth in Lending Act
Carte Blanche urges as an additional objection to the district court’s order that in enacting 15 U.S.C. § 1640(a) and in particular the penalty provision in § 1640(a)(1), Congress intended that the penalty could be recovered only in an individual and not in a class action. It points to the attorneys fee provision in § 1640(a)(2) as suggesting such an interpretation. Neither of the district court opinions explicitly passed upon that contention. Carte Blanche urges, however, that it is implicit in the district court’s superiority determination. It does not appear to have been decided even implicitly for, as Carte Blanche concedes in its petition for rehearing, there is no suggestion that Congress intended to prohibit the recovery of actual damages in [764]*764a class action. The district court has not yet determined whether it will award statutory penalties as well as damages. Its superiority determination would have been no different on the assumption that class members might recover actual damages only. Our reversal does not turn on the difference between actual damages and statutory penalties. We cannot exercise appellate jurisdiction over an issue not yet. decided in the district court.
The order appealed from will be reversed and the case remanded to the district court for further proceedings consistent with this opinion.