Before SMITH, GRAVES, and HO, Circuit Judges. JERRY E. SMITH, Circuit Judge: A district court must engage in a “rigorous analysis” when it certifies a class action. In the absence of that rigor, we vacate the certification order. Plan Benefit Services, Inc., Fringe Insurance Benefits, Inc., and Fringe Benefit Group (collectively “FBG” or “the company”) market and administer retirement and health benefit plans to various employers. FBG offers the plans through two trusts, and there are many plan options, which (FBG asserts) vary in fees and structures. Employers can sign up to provide their employees with benefits through those offerings; or they can retain FBG merely to keep records and supply administrative services. The named plaintiffs are current and former employees of a company that contracted with FBG for various services. They sued FBG under the Employee Retirement Income Security Act of 1974 (“ERISA”), 29 U.S.C. § 1001 et seq. The thrust of their complaint is that FBG has acted as a fiduciary and breached its duties. They charge that the company accepted excessive fees, handpicked providers to maximize its profits, controlled disbursements from the trusts for its own benefit, and unlawfully procured indirect compensation. In short: They allege garden-variety fiduciary misbehavior. The plaintiffs seek to represent a class of “all participants in and beneficiaries of employee benefit plans that provide benefits through [the trusts], . . . from July 6, 2011 until the time of trial.” The proposed group involves some 90,000 individuals and implicates many employers and plans.[1] FBG opposes certification with predictable vigor. It maintains that the central issues—such as whether FBG was a fiduciary or charged excessive fees—necessarily turn on the diverse features of each plan, so a class involving so many of them is improper. After a hearing, the court certified the class. It settled on a mandatory Federal Rule of Civil Procedure 23(b)(1)(B) class and found that Rule 23(a)’s prerequisites—numerosity, commonality, typicality, and adequacy—are met. Despite the complexity of the certification issues, the sweeping scope of the proposed class, and FBG’s numerous case-specific objections, the court’s certification order has about five pages of substantive analysis. We accepted this interlocutory appeal, see Fed. R. Civ. P. 23(f), and vacate. II. A. Given the impact of certification, district courts must analyze Rule 23 with special attention. Certification is proper only where “the trial court is satisfied, after a rigorous analysis,”[2] that the Rule’s requirements are met. Put another way, “a district court must detail with sufficient specificity how the plaintiff has met the requirements of Rule 23.” Vizena v. Union Pac. R.R., 360 F.3d 496, 503 (5th Cir. 2004) (per curiam). “Rule 23 does not set forth a mere pleading standard.” Dukes, 564 U.S. at 350. Instead, “[a] party seeking class certification must affirmatively demonstrate his compliance with the Rule—that is, he must be prepared to prove that there are in fact sufficiently numerous parties, common questions of law or fact,” and so on. Id. As a result, in weighing certification, the court will often have “to probe behind the pleadings,” Falcon, 457 U.S. at 160, because “[t]he class determination generally involves considerations that are enmeshed in the factual and legal issues” of the case, Dukes, 564 U.S. at 351. So the court should seek to “understand the claims, defenses, relevant facts, and applicable substantive law in order to make a meaningful determination[.]“ Flecha v. Medicredit, Inc., 946 F.3d 762, 766 (5th Cir. 2020). “If some of the determinations . . . cannot be made without a look at the facts, then the judge must undertake that investigation.” Spano v. Boeing Co., 633 F.3d 574, 583 (7th Cir. 2011). The judge cannot merely “review a complaint and ask whether, taking the facts as the party seeking the class presents them, the case seems suitable for class treatment.” Id. (emphasis added). Much more is needed. Thus, to satisfy the rigor requirement, a district court must detail with specificity its reasons for certifying. Vizena, 360 F.3d at 503. It must explain and apply the substantive law governing the plaintiffs’ claims to the relevant facts and defenses, articulating why the issues are fit for classwide resolution.[3] The court should respond to the defendants’ legitimate protests of individualized issues that could preclude class treatment.[4] And its analysis must stay close to the facts and law of the case, spurning reliance on generalizations about what types of disputes may be fit for a class.[5] The court must rigorously consider both Rule 23(a)’s prerequisites[6] and the Rule 23(b) class type.[7] This “rigorous analysis” mandate is not some pointless exercise that we foist on this circuit’s hardworking and conscientious district judges, such as the judge in this case. It matters. A “class action is an exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only,” Behrend, 569 U.S. at 33 (quotation marks removed), and creative uses are perilous. It is no secret that certification “can coerce a defendant into settling on highly disadvantageous terms regardless of the merits of the suit.”[8] And the existence of a class fundamentally alters the rights of present and absent members, particularly for mandatory classes such as the one here.[9] No less than due process is implicated, so a careful look is necessary. See Unger v. Amedisys Inc., 401 F.3d 316, 320 (5th Cir. 2005). A rigorous analysis also ensures effective appellate review, which we are unable to engage in here. See Stukenberg, 675 F.3d at 842. Indeed, in addressing a certification, we search only for abuse of discretion, recognizing “the essentially factual basis of the certification inquiry and . . . the district court’s inherent power to manage and control pending litigation.” Id. at 836. Appellate judges are not finders of fact,[10] and we play no role in managing a district court’s docket. So it’s up to the district judge to find the facts. B. FBG complains that the court analyzed Rule 23 superficially. We agree. 1. We begin with commonality, which requires the plaintiff to show that “there are questions of law or fact common to the class.” Fed. R. Civ. P. 23(a)(2). Under Dukes—which “heightened the standards”[11]—”all of the class member’s claims [must] depend on a common issue . . . whose resolution will resolve an issue that is central to the validity of each one of the class member’s claims in one stroke.” Stukenberg, 675 F.3d at 840 (cleaned up). To satisfy the rigor requirement, the district court must explain how that standard is met. Id. at 841. It should do so “with specific reference to the claims, defenses, relevant facts, and applicable substantive law raised by the class claims,” id. at 843 (quotation marks removed), and it must “address actual or potential differences in purported class members’ individual circumstances and claims,” Ward, 753 F. App’x at 246. Here, however, the court’s analysis—for this huge putative class—is fleeting: In this case, Plaintiffs allege prohibited self-dealing and fiduciary breaches stemming from Defendants’ exertion of discretionary control over [the trusts]. Plaintiffs further allege Defendants’ actions affected all plans participating in [the trusts]. Because Defendants’ status as fiduciaries with discretionary control over [the trusts] presents a common question capable of classwide resolution, Plaintiffs’ proposed class satisfies the commonality requirement (docket citations removed). Our caselaw demands more than this brevity. First, the order does not identify the common question with any specificity.[12] And, having defined the question vaguely, the court then analyzes it conclusionally. There is no reference to ERISA.[13] Nor does the court explain why clarifying FBG’s status as a fiduciary will in one stroke resolve an issue that is central to the claims of each one of the class members.[14] Most noticeably, the order neglects to consider asserted differences among class members that could prevent the suit from generating “common answers apt to drive the resolution of the litigation.”[15] As the briefs reveal, liability apparently will turn on whether FBG (1) was a fiduciary as to each plan[16] and (2) received too much compensation.[17] But FBG informs us—just as it told the district court—that differences in fees, structure, and assignment of responsibility among the many plans negate the possibility of a common answer to those issues. The plaintiffs reply that the documents governing each plan are materially indistinguishable, such that the analysis pertaining to one plan will apply to all. FBG retorts that there are fundamental and dispositive differences among the plan documents. Faced with those warring factual contentions, the district court needed to resolve whether there were relevant differences among the many plans— and, if so, to explain why they did not prevent classwide resolution of the common issue.[18] Instead, the court offered only its unsupported assurance that FBG’s fiduciary status is “capable of classwide resolution.” It is an abuse of discretion to find commonality on such a thin survey.[19] An issue is not “capable of classwide resolution” just because the district court, without explanation, says it is. 2. The district court’s analysis of the class type is similarly—and reversibly—breezy. The court certified under Rule 23(b)(1)(B), which provides for a mandatory class with no opting out,[20] where prosecuting separate actions by or against individual class members would create a risk of . . . adjudications with respect to individual class members that, as a practical matter, would be dispositive of the interests of the other members not parties to the individual adjudications or would substantially impair or impede their ability to protect their interests [.] Fed. R. Civ. P. 23(b)(1)(B). The district court’s reasoning needs more facts and fewer generalizations. The court noted that the plaintiffs sought “restitution, an accounting for profits, and an order that Defendants make good to the plans the losses stemming from Defendants’ exercise of discretion and control with respect to [the trusts]” (quotation marks removed). It then quoted dictum in Ortiz, which stated that Rule 23(b)(1)(B) would often encompass “actions charging a breach of trust by a[] . . . fiduciary similarly affecting the members of a large class of beneficiaries,” because such actions “requir[e] an accounting or similar procedure to restore the subject of the trust.” Ortiz, 527 U.S. at 834 (quotation marks removed). The proposed group fell within that “[c]lassic example,” the district court opined, because the plaintiffs sought an ERISA accounting that would make good any losses that resulted from FBG’s fiduciary breaches. So Rule 23(b)(1)(B) applied. Despite relying on Ortiz, the court missed its central premise. Even if a proposed class appears, at first glance, to be an historical example of a 23(b)(1)(B) class, the court must look closely at the facts to ensure that separate adjudications as to each class member would indeed impair nonparties’ ability to protect their interests.[21] In other words, the Rule does not tolerate a fleeting analysis that assumes that a given class is appropriate just because it appears to be a classic (b)(1)(B) class. See id. at 848-61. Applying that principle, Ortiz held that a proposed class did not qualify, because it did not sufficiently resemble a “limited fund.” See id. In so doing, the Court applied a “limiting construction” to Rule 23(b)(1)(B) designed “to stay close to the historical model[s]” of such classes. Id. at 842. It spoke warily of the “likelihood of abuse” of that subsection and counseled against its “adventurous application.”[22] The Court wished to avoid “potential conflict with the Rules Enabling Act” and “serious constitutional concerns raised by the mandatory class resolution of individual legal claims.” Id. Ortiz shows that the district court’s order does not analyze the class type with requisite rigor. The court notes that, just as in Ortiz, id. at 834, the plaintiffs’ case relates to one of the historical models—namely, an action against a fiduciary seeking an accounting to restore the subject of the trust (in this case, benefits plans). But, parting ways with Ortiz, the court’s analysis begins and ends there. It fails to examine the facts of this specific class to ensure that it qualifies. That failure is reversible error.[23] FBG vigorously asserts that the class in no way resembles a typical ERISA class belonging under (b)(1)(B). Among other things, the company contends—as it did at the district court—that even assuming that the relief sought for one individual’s claim would be dispositive of another individual’s claim in the same plan, that relief would not affect, much less be dispositive of, claims under the other . . . plans. Different plans require their own liability and damage [s] analysis, and their own respective accountings or other relief. The order has no response to that fact-bound contention. It instead assumes that the desired relief—an ERISA accounting—automatically entitles the plaintiffs to certification via (b)(1)(B). The court’s nonspecific, categorical reasoning contradicts Ortiz and fails to demonstrate a rigorous analysis. It follows that we cannot do this work in the district court’s stead. The Rule requires a case-specific inquiry into whether, “as a practical matter,” one class member’s relief in an adjudication would impair the interests of non-parties.[24] But here, the court bypasses that required analysis, offering instead only its conclusional reassurance that the desired “relief would, as a practical matter, dispose of the interests of the other putative class members.” A more searching analysis is required. “[C]areful attention to the requirements of [Rule] 23 remains . . . indispensable.” Rodriguez, 431 U.S. at 405. The court must explain, in the factual context of this case, why adjudications for individual class members would prejudice nonparty members.[25] Such will necessarily entail considering and responding to FBG’s contention that the existence of many different plans prevents (b)(1)(B) from applying. C. The plaintiffs, maybe sensing the inadequacies, point out that the court had plenty of evidence before it and that the hearing showed that the court was thoughtfully considering the issues. All true. But the rigor requirement does not turn on how much material the parties submit or on how carefully the court seems to weigh the issues at a hearing. It instead asks whether the court’s order sufficiently analyzes Rule 23. E.g., Stukenberg, 675 F.3d at 842-44. For reasons described, the one here does not. The certification order is VACATED. We express no view on whether a class should be certified, and we place no limit on the matters the court may consider in further proceedings.
