Summary. A class action lets one lawsuit resolve claims for thousands of people who will never appear in court, which is either an indispensable tool for aggregating small harms or a machine for manufacturing settlement pressure, depending on which side of the caption you occupy. Both descriptions are sometimes true. This article explains how Rule 23 actually works, from the four prerequisites of Rule 23(a) through the three categories of Rule 23(b), with attention to the fights that decide most cases: commonality after Wal-Mart v. Dukes, the damages-model requirement of Comcast v. Behrend, representative proof after Tyson Foods, ascertainability and the circuit split over it, and the standing problem for uninjured class members after TransUnion v. Ramirez. It covers settlement approval under the amended Rule 23(e), objectors, interlocutory review under Rule 23(f), and federal jurisdiction under the Class Action Fairness Act. It then turns to strategy for both sides, including the arbitration and class waiver defense that has reshaped consumer litigation. It closes with checklists, a worked example, an FAQ, and related reading.

Suppose a mobile app overcharged 900,000 users by $2.40 each. No one is going to sue over $2.40. The cost of filing exceeds the recovery by three orders of magnitude, and the app's operator knows it. Without some aggregation device, a $2.16 million wrong produces zero enforcement.

Now suppose instead that a company's HR software had a bug that misrouted 900,000 applications, and a plaintiff seeks to certify a class of everyone who applied, alleging discrimination. Some applicants were qualified, some were not; some would have been hired, most would not; the software behaved differently across a dozen configurations. Aggregating those claims does not produce efficient justice. It produces settlement pressure disconnected from the merits.

Rule 23 is the line between those two cases, and drawing it well is one of the harder tasks in civil procedure.

The short answer

Every class must satisfy all four prerequisites of Rule 23(a):

  1. Numerosity: the class is so numerous that joinder is impracticable.
  2. Commonality: there are questions of law or fact common to the class.
  3. Typicality: the representative's claims are typical of the class.
  4. Adequacy: the representative will fairly and adequately protect the interests of the class.

And fit within one of the three categories of Rule 23(b):

  • (b)(1): separate actions would risk inconsistent adjudications for the party opposing the class, or would be dispositive of the interests of absent members (limited fund).
  • (b)(2): the party opposing the class has acted on grounds generally applicable to the class, so that final injunctive or corresponding declaratory relief is appropriate respecting the class as a whole.
  • (b)(3): questions of law or fact common to class members predominate over individual questions, and a class action is superior to other available methods.

Certification requires a "rigorous analysis" that frequently overlaps with the merits. Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350-51 (2011).

A (b)(3) class requires notice and an opt-out right; (b)(1) and (b)(2) classes are mandatory.

Part I: The Rule 23(a) prerequisites

Numerosity

There is no magic number. Classes of 40 or more are routinely found sufficient; classes under 20 usually are not. Courts also consider geographic dispersion, the ease of identifying members, the size of individual claims, and members' ability to sue individually.

Commonality after Dukes

Commonality was, for decades, a formality. Any shared question satisfied it.

Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011), changed that. The plaintiffs sought to certify a class of roughly 1.5 million female Wal-Mart employees, alleging that the company's policy of delegating pay and promotion decisions to local managers produced discrimination. The Court held commonality was not satisfied:

Commonality requires the plaintiff to demonstrate that the class members "have suffered the same injury." ... Their claims must depend upon a common contention ... of such a nature that it is capable of classwide resolution, which means that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke.

Id. at 349-50. And: "What matters to class certification ... is not the raising of common 'questions' ... but, rather the capacity of a classwide proceeding to generate common answers apt to drive the resolution of the litigation."

The Court also rejected "Trial by Formula," in which a sample of claims would be tried and the results extrapolated, holding it would deprive the defendant of its right to litigate individual defenses.

Dukes also confirmed that certification analysis may require examining the merits where the merits and the Rule 23 requirements overlap. That was contrary to a widespread misreading of Eisen v. Carlisle & Jacquelin, 417 U.S. 156 (1974).

Typicality and adequacy

Typicality asks whether the representative's claim arises from the same event or course of conduct and is based on the same legal theory. It fails where the representative is subject to a unique defense (an individual arbitration agreement, a release, a limitations problem) that will consume the litigation.

Adequacy has two components: the representative must have no conflicts with the class, and class counsel must be qualified and adequately resourced. Rule 23(g) now governs the appointment of class counsel and lists the factors.

Amchem Products, Inc. v. Windsor, 521 U.S. 591 (1997), and Ortiz v. Fibreboard Corp., 527 U.S. 815 (1999), are the leading adequacy decisions. Both rejected sprawling asbestos settlement classes on the ground that the interests of currently injured and future-injury claimants conflicted, and that subclasses with separate representation were required.

Part II: The Rule 23(b) categories

(b)(2): injunctive and declaratory classes

Available where the defendant's conduct applies generally to the class and a single injunction or declaratory judgment would provide relief to the class as a whole.

Dukes held that (b)(2) does not permit claims for individualized monetary relief, at least where monetary relief is not incidental to the injunctive relief. Back pay claims for 1.5 million employees could not ride along on a (b)(2) certification. Whether truly "incidental" monetary relief remains available under (b)(2) was left open, and courts have generally read Dukes to foreclose it.

(b)(2) classes are mandatory: no notice is required and no opt-out right exists, which is why the Court insisted on cohesion.

The category matters more after the recent narrowing of universal injunctions in public law cases, because a (b)(2) class is now a principal route to relief beyond the named plaintiffs. See Preliminary Injunctions and Temporary Restraining Orders.

(b)(3): predominance and superiority

The workhorse category, and the hardest to satisfy.

Predominance "tests whether proposed classes are sufficiently cohesive to warrant adjudication by representation." Amchem, 521 U.S. at 623. It is "far more demanding" than commonality.

Courts assess predominance element by element: which elements can be proved with common evidence, which require individualized proof, and how much of the case each represents.

  • Common: the existence of a uniform policy, the content of a standard form contract, whether a statement was made, whether a product had a defect, whether a security's price was inflated.
  • Individual: reliance (often), causation (sometimes), damages amount (usually), affirmative defenses tied to particular members, and statute of limitations where accrual varies.

Comcast Corp. v. Behrend, 569 U.S. 27 (2013), held that a plaintiff must present a damages model that is consistent with its liability theory. The plaintiffs there advanced four antitrust theories, only one of which survived, but their damages expert's model did not isolate damages from that theory. Certification was reversed.

Comcast is regularly overread by defendants as requiring damages to be provable on a classwide basis. Most circuits have held it does not: individual damages calculations alone do not defeat predominance, provided the model matches the liability theory and common issues otherwise predominate.

Amgen Inc. v. Connecticut Retirement Plans & Trust Funds, 568 U.S. 455 (2013), held that materiality in a securities fraud class need not be proved at certification, because it is a common question that will be resolved for the class either way. The principle generalizes: a merits question that is common does not need to be answered at certification; it needs to be shown to be common.

Tyson Foods, Inc. v. Bouaphakeo, 577 U.S. 442 (2016), permitted the use of representative statistical evidence (a time study of donning and doffing) where each individual plaintiff could have relied on that same evidence in an individual suit. The Court distinguished Dukes: the problem there was that the sample could not have established liability in any individual case. Tyson Foods also flagged, without resolving, the question of how to handle uninjured class members at the distribution stage.

Superiority considers class members' interest in individually controlling separate actions, the extent of litigation already begun, the desirability of concentrating in the forum, and manageability. Manageability is where defendants make their best practical arguments, particularly regarding individualized damages, choice of law across fifty states, and class member identification.

The choice of law problem

In a nationwide class asserting state-law claims, the court must apply the forum's choice of law rules, which may require applying the law of each class member's state. Where the applicable laws differ materially (as they do for consumer protection, unjust enrichment, and warranty claims), predominance and manageability often fail. Plaintiffs respond with statewide subclasses or by pleading a single state's law under a contractual choice of law clause.

Part III: Standing, ascertainability, and uninjured members

TransUnion and uninjured class members

TransUnion LLC v. Ramirez, 594 U.S. 413 (2021), is the most consequential class action decision of the last decade even though it is formally a standing case.

TransUnion had placed alerts on consumer files indicating a potential match to a terrorism watch list. For 1,853 class members, the misleading reports were disseminated to third parties. For the remaining 6,332, the information sat in an internal file and was never disclosed.

The Court held that only the 1,853 had standing, because "[n]o concrete harm, no standing." The remaining class members suffered no concrete injury from the mere existence of inaccurate information in a database. And critically:

Every class member must have Article III standing in order to recover individual damages.

The Court expressly reserved whether every class member must have standing at the certification stage as opposed to the judgment stage, and that question has produced substantial litigation.

The follow-on question is what happens when a certified class contains some uninjured members. The Ninth Circuit, sitting en banc in Olean Wholesale Grocery Cooperative, Inc. v. Bumble Bee Foods, LLC, 31 F.4th 651 (9th Cir. 2022), held that the presence of some uninjured members does not automatically defeat predominance, and that certification is improper only where the number is not de minimis and identifying them would require individual inquiries that overwhelm common issues. The D.C. Circuit and others have taken a stricter view.

The Supreme Court granted certiorari on the question in Laboratory Corp. of America Holdings v. Davis, No. 24-304, but dismissed the writ as improvidently granted in June 2025, leaving the split in place. Expect the issue to return.

Practical implication: class definitions should be drafted to exclude uninjured members where possible, and defendants should develop evidence on the proportion of uninjured members early.

Ascertainability

Most courts require that the class be defined by objective criteria so that membership can be determined. The circuit split concerns whether there is an additional requirement of an administratively feasible method of identifying members.

The practical stakes are highest in low-value consumer product cases where nobody keeps receipts. In feasibility circuits, those cases are much harder to certify.

Fail-safe classes are improper everywhere: a class defined so that membership depends on the merits ("all persons who were unlawfully charged") cannot be certified, because it would bind no one if the defendant wins.

Part IV: Federal jurisdiction and the Class Action Fairness Act

Before 2005, most class actions were filed in state court, where certification standards were often looser and removal was difficult. The Class Action Fairness Act of 2005 changed the landscape.

CAFA jurisdiction, 28 U.S.C. § 1332(d)(2): federal courts have original jurisdiction over class actions where:

  • the proposed class has 100 or more members;
  • the amount in controversy exceeds $5,000,000 in the aggregate, exclusive of interest and costs; and
  • minimal diversity exists (any class member is a citizen of a different state from any defendant).

Removal under 28 U.S.C. § 1453 is easier than ordinary removal: the one-year limit does not apply, unanimity among defendants is not required, and a defendant may remove even if it is a citizen of the forum state.

Dart Cherokee Basin Operating Co. v. Owens, 574 U.S. 81 (2014), held that a notice of removal need include only a plausible allegation of the amount in controversy; evidence is required only if the plaintiff contests it.

Standard Fire Insurance Co. v. Knowles, 568 U.S. 588 (2013), held that a named plaintiff cannot defeat CAFA jurisdiction by stipulating to damages under $5 million on behalf of a putative class, because a proposed representative cannot bind absent members before certification.

Exceptions exist for local controversies and home-state classes, §§ 1332(d)(3)-(4), and for certain securities and internal-affairs claims, § 1332(d)(9).

CAFA also imposed settlement reforms: notice to federal and state officials before settlement approval (§ 1715), limits on coupon settlements (§ 1712), and a prohibition on awards that disadvantage class members based on geographic proximity (§ 1714).

Part V: Settlement

The approval process

A class action may be settled only with court approval. Rule 23(e). The process, as restructured by the 2018 amendments:

  1. Preliminary approval. The parties must provide the court with information sufficient to determine whether it will likely be able to approve the settlement and certify the class. Rule 23(e)(1). The 2018 amendment deliberately raised the bar for this stage, which had become perfunctory.
  2. Notice to the class, "in a reasonable manner to all class members who would be bound," including electronic means. Rule 23(e)(1)(B), 23(c)(2)(B).
  3. Opt-out and objection period.
  4. Fairness hearing.
  5. Final approval upon a finding that the settlement is "fair, reasonable, and adequate."

The Rule 23(e)(2) factors

The 2018 amendments codified four core factors, intended to focus courts on substance rather than on the long, circuit-specific lists that had accumulated:

(A) the class representatives and class counsel have adequately represented the class; (B) the proposal was negotiated at arm's length; (C) the relief provided for the class is adequate, taking into account: (i) the costs, risks, and delay of trial and appeal; (ii) the effectiveness of any proposed method of distributing relief, including the method of processing class-member claims; (iii) the terms of any proposed award of attorney's fees, including timing of payment; and (iv) any agreement required to be identified under Rule 23(e)(3); (D) the proposal treats class members equitably relative to each other.

Factor (C)(ii) deserves attention: courts now scrutinize claims rates and distribution mechanics. A settlement with a headline value of $40 million and a projected 2 percent claims rate is a $800,000 settlement, and judges say so.

Attorney's fees

Rule 23(h) governs. Two methods:

  • Percentage of the fund, typically 25 percent as a benchmark in common fund cases, adjusted up or down.
  • Lodestar (hours times reasonable rate), with a multiplier, more common in fee-shifting statutory cases and in megafund cases where a percentage would be excessive.

Courts frequently cross-check one method against the other. Clear sailing agreements (the defendant agrees not to oppose a fee request) and kicker or reverter clauses (unawarded fees revert to the defendant rather than the class) draw scrutiny as signs of self-dealing.

Cy pres distribution of unclaimed funds to charities is permitted in most circuits but disfavored where direct distribution is feasible, and it has attracted criticism from several Justices.

Objectors

Rule 23(e)(5), as amended in 2018, requires that objections state whether they apply to the objector, a subset, or the whole class, and the grounds. Critically, payments to objectors or their counsel in exchange for withdrawing an objection or dismissing an appeal now require court approval, which was aimed squarely at professional objectors who extracted payments to go away.

A class member who objects and is overruled may appeal without intervening. Devlin v. Scardelletti, 536 U.S. 1 (2002).

Settlement classes

Certification for settlement purposes only is permitted, and the manageability concerns of Rule 23(b)(3)(D) drop out because there will be no trial. Amchem, 521 U.S. at 620. But every other requirement, including predominance and adequacy, applies with full force, and Amchem and Ortiz are the reminders that courts should not certify a class for settlement that could never be certified for litigation.

The Ninth Circuit's en banc decision in the Hyundai and Kia fuel economy litigation, 926 F.3d 539 (9th Cir. 2019), addressed nationwide settlement classes and choice of law, holding that a court may accord more weight to the parties' consensus in the settlement context but must still conduct the predominance analysis.

Part VI: Interlocutory review under Rule 23(f)

Rule 23(f) permits a court of appeals, in its discretion, to permit an appeal from an order granting or denying class certification, if a petition is filed within 14 days after the order.

Two Supreme Court decisions define the boundaries:

Why 23(f) matters so much: for defendants, a certification order often ends the case, because the aggregate exposure forces settlement regardless of the merits (the "in terrorem" or "blackmail settlement" concern that Judge Posner and Judge Friendly both described). For plaintiffs, a denial usually ends the case, because individual claims are not worth pursuing. Rule 23(f) is the only realistic route to review before that pressure resolves the dispute.

Part VII: Defense strategy

Before certification

  1. Arbitration and class waivers. This is the single most effective defense in consumer and employment cases. AT&T Mobility LLC v. Concepcion, 563 U.S. 333 (2011), held the Federal Arbitration Act preempts state rules conditioning arbitration on the availability of classwide procedures. Epic Systems Corp. v. Lewis, 584 U.S. 497 (2018), upheld class waivers in employment arbitration agreements against a National Labor Relations Act challenge. If your client's terms include a properly formed arbitration clause with a class waiver, move to compel first. Formation is the battleground; see Website Terms of Service and Online Contract Formation.
  2. Standing. After TransUnion, attack the named plaintiff's concrete injury and develop evidence about uninjured absent members.
  3. Personal jurisdiction. Bristol-Myers Squibb Co. v. Superior Court, 582 U.S. 255 (2017), created serious problems for nationwide classes in forums where the defendant is not at home. The circuits have split over whether it applies to absent class members in federal class actions. See Personal Jurisdiction Over Online and Foreign Defendants.
  4. Rule 12. Move to dismiss on the merits, and consider a motion to strike class allegations where the defect is apparent from the pleading (rarely granted, but occasionally worth it).
  5. Merits discovery on the representative. Unique defenses defeat typicality. Look for releases, arbitration agreements, limitations problems, and credibility issues.

At certification

  1. Element-by-element predominance analysis. Show precisely which elements require individualized proof and why they dominate.
  2. Attack the damages model under Comcast: does it match the surviving liability theory?
  3. Attack the plaintiff's expert under Rule 702. Courts generally must resolve Daubert challenges to expert evidence offered in support of certification. See Expert Witnesses After the 2023 Amendment to Rule 702.
  4. Ascertainability, where the circuit recognizes an administrative feasibility requirement.
  5. Choice of law for multistate classes: submit a state-by-state analysis of material variations.
  6. Manageability: a concrete trial plan critique is more persuasive than an abstract complaint.
  7. Uninjured members: quantify them.

Plaintiff-side strategy

  1. Define the class narrowly and objectively. Overbroad definitions are the most common self-inflicted wound. Exclude uninjured members expressly.
  2. Pick the right representative. Someone without unique defenses, who understands the case, and who will hold up in deposition.
  3. Build the common-proof story. Identify the uniform policy, the standard form, the common representation, or the common defect.
  4. Present a damages model tied to the theory and have the expert prepared for a Rule 702 challenge at certification.
  5. Consider issue certification under Rule 23(c)(4), which permits certifying particular issues even where the whole claim cannot be certified. This is underused, and several circuits have endorsed it.
  6. Consider subclasses under Rule 23(c)(5) for choice of law or conflicting interests.
  7. File early and preserve tolling. American Pipe & Construction Co. v. Utah, 414 U.S. 538 (1974), tolls the limitations period for absent members during the pendency of a class action, but China Agritech, Inc. v. Resh, 584 U.S. 732 (2018), holds that tolling does not permit a successive class action after certification is denied. Absent members may file individually; they may not restart the class clock.

Part VI-A: Notice, claims administration, and what class members actually get

Two operational subjects decide whether a class action delivers anything, and both are now scrutinized at approval.

Notice. Rule 23(c)(2)(B) requires "the best notice that is practicable under the circumstances, including individual notice to all members who can be identified through reasonable effort." The 2018 amendments confirmed that notice may be by "United States mail, electronic means, or other appropriate means." In practice that has meant email and targeted digital campaigns supplementing or replacing mail, and courts now expect a declaration from the administrator describing reach rates and methodology.

The notice itself must state, in plain, easily understood language: the nature of the action; the class definition; the claims, issues, and defenses; that a member may appear through counsel; that the court will exclude members who request it; the time and manner for requesting exclusion; and the binding effect of a class judgment.

Claims administration. Rule 23(e)(2)(C)(ii) directs courts to consider "the effectiveness of any proposed method of distributing relief to the class, including the method of processing class-member claims." That single clause has changed settlement design. Courts now routinely ask:

  • What is the projected claims rate, and on what comparable settlements is the projection based?
  • Can members be paid directly from the defendant's records without filing a claim? Direct payment produces participation rates an order of magnitude higher than claims-made structures, and where the defendant has the payment credential on file, courts increasingly expect it.
  • How burdensome is the claim form? Requiring receipts for a $12 product purchased four years ago is a design choice with a predictable result.
  • What happens to residual funds? Pro rata redistribution to claimants is preferred over cy pres where feasible.
  • Is the fee award tied to funds actually distributed rather than to a theoretical maximum? Several courts now structure fees to be paid in tranches as distribution is verified.

The credibility point for practitioners on both sides: a settlement presented as "$40 million in benefits" that distributes $900,000 will draw objections, adverse commentary in the approval order, and sometimes denial. Settlements that report honestly, including a candid projected claims rate, are approved faster and survive appeal better. See Motions to Dismiss Under Rule 12 for the standing issues that determine who belongs in the class in the first place.

A worked example

Harbor Financial App, Inc. (fictional) charged users a "network fee" of $1.99 per transfer, disclosed in a footnote of its terms but not on the transfer confirmation screen. A plaintiff sues on behalf of "all United States users charged a network fee," alleging violations of state consumer protection statutes and unjust enrichment. The class is roughly 2.3 million users; aggregate fees are $37 million.

Harbor's first move: arbitration. Harbor's onboarding flow used a checkbox with the terms hyperlinked next to it, and the terms contain an arbitration clause with a class waiver. Whether the clause is enforceable depends on whether the interface gave reasonable notice and whether the user manifested assent. Harbor should move to compel arbitration immediately. If the clause holds, the case ends. If it does not, Harbor has learned something important about its interface.

If arbitration fails, standing. The named plaintiff paid the fee, so she has concrete injury. But Harbor should develop evidence about users who received promotional fee waivers, users who were shown a different screen after a 2023 redesign, and users who saw the fee disclosed on the confirmation page in the iOS build. Each is a subgroup that may be uninjured or differently situated.

Predominance. Harbor's strongest argument is the interface variation: at least four different disclosure flows existed over the class period across two platforms. If the disclosure differed materially, the "was it adequately disclosed" question is not common. Harbor should build a version-by-version chronology with screenshots, which is exactly the kind of concrete evidence that defeats predominance.

Choice of law. State consumer protection statutes differ on reliance, scienter, damages, and notice prerequisites. A nationwide class on those claims is vulnerable. The plaintiff's better play is a California class under the UCL and CLRA, or statewide subclasses in the ten largest states.

Damages. Common, mechanical, and computable from Harbor's own records. This favors the plaintiff, and Comcast is not a problem because the model (fees charged) matches the theory (fees were not adequately disclosed).

Likely outcome. If arbitration fails, certification of a narrower class limited to the users who saw a particular disclosure flow, in a subset of states, followed by settlement. The negotiation will turn on the claims rate: a fund with direct payments to identified accounts (Harbor has every user's payment credential) will be scrutinized far more favorably under Rule 23(e)(2)(C)(ii) than a claims-made structure.

Checklists

Plaintiff's certification checklist

  • Class definition: objective, ascertainable, not fail-safe, excludes uninjured members.
  • Representative free of unique defenses; deposition-ready.
  • Rule 23(a) elements supported by evidence, not argument.
  • Common contention identified with a common answer (Dukes).
  • Element-by-element predominance showing.
  • Damages model tied to the liability theory (Comcast), expert prepared for Rule 702.
  • Choice of law analysis; subclasses if needed.
  • Trial plan addressing manageability.
  • Rule 23(g) showing on counsel adequacy.
  • Consider Rule 23(c)(4) issue certification as a fallback.

Defendant's checklist

  • Arbitration clause and class waiver: formation evidence assembled.
  • Standing attack on named plaintiff and evidence on uninjured members.
  • Personal jurisdiction analysis for absent members.
  • Unique defenses against the representative.
  • Variation evidence (product versions, disclosures, contracts, policies over time).
  • Rule 702 challenge to the certification expert.
  • State-by-state law variation chart.
  • Concrete critique of the trial plan.
  • Rule 23(f) petition calendared for 14 days after any certification order.
  • CAFA removal analysis, if filed in state court.

Frequently asked questions

How many people do I need for a class? No fixed number. Forty or more usually satisfies numerosity; fewer may suffice with geographic dispersion or other joinder obstacles.

Do all class members have to be injured? Every member must have Article III standing to recover damages under TransUnion. Whether a class containing some uninjured members may be certified at all remains contested, and the Supreme Court has not resolved it.

Can a company avoid class actions with an arbitration clause? Often, yes. Concepcion and Epic Systems uphold class waivers in arbitration agreements. The fight is usually about contract formation, not enforceability. Note the practical countermeasure of mass individual arbitration, which has made some companies rethink these clauses.

What is the difference between (b)(2) and (b)(3)? (b)(2) is for classwide injunctive or declaratory relief, is mandatory, and does not permit individualized damages. (b)(3) is for damages, requires predominance and superiority, and requires notice and an opt-out right.

How long does a class action take? Certification alone commonly takes 18 months to three years, with expert discovery and briefing. Add appeal and, if not settled, trial.

What do class members actually receive? It varies enormously. Direct payments to identified accounts produce high effective rates. Claims-made settlements in consumer cases often see single-digit participation. Courts now examine this closely under Rule 23(e)(2)(C)(ii).

Can I opt out? From a (b)(3) class, yes, during the period stated in the notice. From (b)(1) and (b)(2) classes, no. Opting out preserves your individual claim, which matters if your damages are substantial.

What is a cy pres award? A distribution of unclaimed settlement funds to a charity whose work relates to the class's interests. Permitted in most circuits but disfavored where direct distribution is feasible, and increasingly criticized.

Are objectors legitimate? Some are, and the amended Rule 23(e)(5) is designed to keep the legitimate ones while stopping payments to professional objectors who file boilerplate objections and sell their appeals.

Closing thought

The honest assessment of class actions is that they are simultaneously the only practical remedy for widely distributed small harms and a mechanism whose settlement pressure can be untethered from the merits. Rule 23's certification requirements exist to keep those two realities in balance, and the Supreme Court's decisions over the last fifteen years have moved steadily toward requiring plaintiffs to show, with evidence, that a class trial would actually work.

For defendants, that means the case is usually decided at certification, and the evidence that decides it, the version histories, the policy variations, the disclosure flows, the uninjured subgroups, must be assembled early and presented concretely.

For plaintiffs, it means narrowing is strength. The most common reason good class cases fail is a class definition drafted for maximum exposure rather than for maximum coherence. A tight class with a clean common answer certifies. A sprawling one does not, and the sprawl usually cannot be fixed after the motion is briefed.

Related articles

This article is provided for general informational purposes and does not constitute legal advice. Class action doctrine varies significantly among circuits and continues to develop. Consult qualified class action counsel about any particular matter.