Checklist · Litigation

Opposing Class Certification: A Defense Checklist

Type
Checklist
Jurisdiction
National
Date
2024-08-03
Topic
Litigation

Class certification is the whole case. A certified class converts a modest dispute into bet-the-company exposure and creates settlement pressure that has almost nothing to do with the merits; a denial usually ends the litigation. This checklist walks the defense side through the sequence that actually decides certification motions — the threshold kills (arbitration and class waiver, Rule 12, standing), the attack on the class definition, the attack on the named representatives, and then the substantive fights over commonality under Wal-Mart Stores, Inc. v. Dukes, predominance, the damages model under Comcast Corp. v. Behrend, uninjured class members after TransUnion LLC v. Ramirez, expert exclusion under Daubert and Rule 702, and superiority and manageability. It shows how to build the certification record through discovery rather than argue it from the pleadings, how to write the individualized-issues catalog that judges actually rely on, and what to do in the fourteen days after you lose. Every phase is followed by an explanation of why it matters and the traps that swallow careless defendants.

Opposing Class Certification: A Defense Checklist

Litigation > Class Actions | Published 3 August 2024 - Updated 11 December 2025 | UpLaw - uplaw.us

Certification is not a procedural way station. It is the merits, decided early, by a judge, on an incomplete record, with the company's survival riding on the answer. Defend it that way.

Keywords: opposing class certification, Rule 23 defense, predominance, commonality, Dukes, Comcast, TransUnion, class waiver, ascertainability, fail-safe class, adequacy, Daubert, uninjured class members, trial plan, Rule 23(f)

How to use this checklist

Certification decides the economics of a class action. Once a class is certified, aggregate exposure jumps by three or four orders of magnitude, insurers start talking about tower exhaustion, and the board begins to price settlement against a verdict the company cannot survive. Deny certification, and the same case is a single-plaintiff dispute worth what one plaintiff's damages are worth — usually nothing.

So the defense of a class action is the defense of certification, and it begins the day the complaint arrives, not the day the motion is filed. The phases below run in rough chronological order. Work them in parallel where you can: the discovery you take in Phase 11 is what makes the arguments in Phases 4 through 10 land. For the doctrinal architecture underneath all of it, read Class Action Litigation: A Comprehensive Guide to Rule 23 alongside this checklist, and read your opponent's playbook in Moving for Class Certification: A Practical Checklist — the best certification oppositions are written by lawyers who know exactly what the plaintiff is trying to build.

Our running example. Alderwood Nutrition, Inc. sells a supplement called MetaboLift through retail and its own website. A putative nationwide class alleges the label claim "clinically shown to boost metabolism" is deceptive under state consumer-protection statutes. Plaintiff proposes a class of "all purchasers of MetaboLift." We will use Alderwood throughout.

Phase 0 — The threshold kills: arbitration, Rule 12, and standing

  • [ ] Pull every version of the customer agreement, terms of sale, EULA, or employment agreement in force during the class period; determine whether an arbitration clause with a class-action waiver applies to the named plaintiff.
  • [ ] Confirm the clause's formation record (clickwrap assent logs, signature pages, acknowledgment forms) — you will have to prove assent, not assert it.
  • [ ] Move to compel arbitration promptly; do not litigate the merits first and manufacture a waiver.
  • [ ] Check for a delegation clause assigning arbitrability to the arbitrator.
  • [ ] Screen the complaint for Rule 12(b)(6) defects under Bell Atlantic Corp. v. Twombly, 550 U.S. 544 (2007), and Ashcroft v. Iqbal, 556 U.S. 662 (2009) — a dismissal on the elements is cheaper than a certification fight.
  • [ ] Test the named plaintiff's own Article III standing: did she buy the product, see the representation, suffer a concrete injury?
  • [ ] Consider a motion to strike class allegations only where the defect is apparent on the face of the pleading (a facially fail-safe definition, a nationwide class under a statute with no extraterritorial reach).
  • [ ] Preserve every defense in the answer; see Answering a Complaint and Pleading Affirmative Defenses: A Practical Checklist.

Why it matters, and the traps

An enforceable class waiver ends the case before Rule 23 is ever reached. AT&T Mobility LLC v. Concepcion, 563 U.S. 333 (2011), held that the Federal Arbitration Act preempts state-law rules conditioning enforceability on the availability of classwide arbitration; American Express Co. v. Italian Colors Restaurant, 570 U.S. 228 (2013), rejected the argument that a waiver is unenforceable merely because individual arbitration is uneconomic; and Epic Systems Corp. v. Lewis, 584 U.S. 497 (2018), extended the point to employment agreements over an NLRA challenge. Stolt-Nielsen S.A. v. AnimalFeeds International Corp., 559 U.S. 662 (2010), and Lamps Plus, Inc. v. Varela, 139 S. Ct. 1407 (2019), confirm that class arbitration cannot be inferred from silence or ambiguity.

Two traps. First, waiver by litigation conduct — filing dispositive motions, taking merits discovery, and then remembering the arbitration clause is how defendants lose an otherwise winning motion. Move early. Second, the drafting trap: many clauses were written before the modern case law and do not clearly cover the claims, the parties, or the transaction. If you are on the drafting side of this problem, Drafting an Arbitration Clause: A Practical Checklist and Arbitration in Practice: A Comprehensive Guide to Alternative Dispute Resolution are the places to start.

One thing that does not work: buying off the named plaintiff with a Rule 68 offer of judgment. Campbell-Ewald Co. v. Gomez, 577 U.S. 153 (2016), holds that an unaccepted offer of complete relief does not moot the plaintiff's claim or the putative class action. Stop trying.

Phase 1 — Attack the class definition

  • [ ] Diagram the proposed definition element by element. Ask: who is in, who is out, and how would a stranger to the case tell?
  • [ ] Flag fail-safe definitions — those that define membership by reference to the merits ("all purchasers who were deceived by the label"). Membership cannot depend on winning.
  • [ ] Flag overbroad definitions that sweep in large numbers of persons who could not have been injured (purchasers who never saw the representation, who bought before it appeared, who received refunds).
  • [ ] Flag subjective criteria ("all persons who believed…") that no record can test.
  • [ ] Identify whether membership can be determined from any records at all — company data, retailer data, loyalty programs, receipts — or only from claimant say-so.
  • [ ] Look up your circuit's ascertainability rule. It genuinely differs.
  • [ ] Check whether the definition matches the claims actually pleaded and the class period actually supported.

Why it matters, and the traps

Rule 23 nowhere uses the word "ascertainability," but every court requires that a class be defined by objective criteria, and the circuits diverge sharply on how much more is required. The Third Circuit in Carrera v. Bayer Corp. (3d Cir. 2013) demanded a reliable, administratively feasible method of identifying members — a rule with real teeth for low-value retail products bought with cash. The Seventh Circuit in Mullins v. Direct Digital, LLC (7th Cir. 2015) and the Ninth Circuit in Briseno v. ConAgra Foods, Inc. (9th Cir. 2017) expressly rejected any freestanding administrative-feasibility requirement, treating manageability as a Rule 23(b)(3)(D) superiority factor instead. Do not brief the Third Circuit rule in the Ninth. Nothing tells a judge you are recycling a brief faster.

The fail-safe objection is the most under-used weapon in the defense arsenal. A class of "all purchasers who were deceived" is unworkable in both directions: if the defendant wins, there is no class and no res judicata, so absent members can sue again — which is precisely the unfairness Rule 23 exists to prevent.

Alderwood, applied. "All purchasers of MetaboLift" is overbroad on its face. Alderwood's data shows 31% of unit volume was sold through subscription auto-ship to repeat customers who bought before the challenged claim appeared on the label, and 12% of purchasers took the money-back guarantee and were refunded in full. Both groups are in the proposed class; neither can have suffered the injury alleged. That is not a quibble about definitions — it is the seed of the predominance and TransUnion arguments in Phases 4 and 7.

Phase 2 — Attack the representatives: typicality, adequacy, and unique defenses

  • [ ] Depose every named plaintiff. Do not stipulate. Do not shorten.
  • [ ] Establish exactly what each plaintiff saw, when, where, and why she bought — the reliance narrative is made or unmade here.
  • [ ] Probe unique defenses: arbitration agreement, release, statute of limitations, prior refund, resale, non-purchase, purchase for resale, unclean hands.
  • [ ] Probe credibility: prior inconsistent statements, litigation history, serial plaintiff status, bankruptcy schedules omitting the claim (a judicial-estoppel goldmine).
  • [ ] Probe knowledge and control: does the plaintiff know what the case is about, who the defendant is, what relief she seeks? A plaintiff who cannot describe her own claim is an inadequate representative.
  • [ ] Examine class counsel's adequacy under Rule 23(g) — resources, experience, conflicts, and any conflict between subgroups (past versus future claimants; refunded versus unrefunded purchasers).
  • [ ] Identify intra-class conflicts that make one representative unable to fairly represent the whole.

Why it matters, and the traps

Rule 23(a)(3) and (a)(4) are not formalities. Amchem Products, Inc. v. Windsor, 521 U.S. 591 (1997), refused certification of a sprawling asbestos settlement class in part because a single set of representatives could not fairly represent both the presently injured and the exposure-only future claimants — their interests in present compensation versus a preserved fund were structurally opposed. That is the template: find the subgroup whose interests the named plaintiff is trading away.

The unique-defense attack works because a representative "beset by unique defenses" will spend the trial litigating her own case, not the class's. If Alderwood's records show the lead plaintiff clicked through terms containing an arbitration clause in 2022, she is either arbitrating or she is atypical — and either answer helps.

The trap is replacement. Knocking out one named plaintiff usually just buys you an amended complaint with a cleaner one. So aim higher: use the deposition to establish facts that are structural — that reliance varied, that exposure varied, that refunds were available — rather than merely embarrassing. Plan the examination with Preparing to Take a Deposition: A Practical Checklist and The Federal Deposition Toolkit: From Notice to Trial Testimony.

Phase 3 — Defeat commonality under Dukes

  • [ ] Force the plaintiff to articulate the common question with precision, then ask what the common answer is.
  • [ ] Show that answering the question does not resolve any element of any claim for the class as a whole.
  • [ ] Where the theory rests on a policy or practice, prove the policy was discretionary, decentralized, or inconsistently applied.
  • [ ] Where the theory rests on a uniform representation, prove the representation was not uniform — labels changed, ad copy varied, channels differed.
  • [ ] Insist on the "rigorous analysis" the court owes: certification requires proof, not allegation.

Why it matters, and the traps

Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011), is the hinge of modern defense practice. Commonality, the Court held, requires more than a shared question; it requires that the claims "depend upon a common contention" whose determination "will resolve an issue that is central to the validity of each one of the claims in one stroke." Millions of women sharing the question "did Wal-Mart discriminate?" was not enough, because a company policy of delegated discretion to thousands of local managers produces no common answer. Dukes also confirms that a court must undertake a rigorous analysis that will frequently "overlap with the merits," building on General Telephone Co. of the Southwest v. Falcon, 457 U.S. 147 (1982).

The trap for defendants is arguing commonality when you should be arguing predominance. Commonality under Rule 23(a)(2) is a low bar in most circuits — one genuinely common answer clears it. Spend two pages there and thirty on Rule 23(b)(3). The exception is the Dukes-style decentralized-discretion case, where commonality is the whole ballgame.

Phase 4 — Defeat predominance: build the individualized-issues catalog

  • [ ] Chart every element of every claim, and for each element mark whether it can be proven with common evidence or requires member-by-member proof.
  • [ ] Do the same for every affirmative defense.
  • [ ] Identify choice-of-law problems: a nationwide class under fifty states' consumer-protection statutes multiplies the elements (some require reliance, some do not; some require pre-suit notice; some bar class actions for statutory damages outright). The plaintiff bears the burden of a workable law-application plan.
  • [ ] Quantify: how many mini-trials, how many hours, how many witnesses?
  • [ ] Draft the individualized-issues table as an exhibit. Judges use it.

Why it matters, and the traps

Amchem teaches that Rule 23(b)(3) predominance is "far more demanding" than Rule 23(a)(2) commonality. The question is not whether common issues exist but whether they predominate — whether the case can be tried as a class or would collapse into thousands of individual proceedings wearing a class-action costume.

State-law variation is the quietest killer of nationwide classes. Consumer-protection statutes are not fungible. Some require actual reliance; some create a presumption; some require a demand letter; some exclude class relief. The plaintiff who proposes a fifty-state class must show how a jury could apply that law — and a proposed grouping of states into "three subgroups" that ignores real doctrinal differences is not a plan, it is a wish.

The trap is generality. "Individual issues predominate" is not an argument. "Predominance fails because each class member must prove (a) exposure to one of nine label variants, (b) reliance under the twenty-two states requiring it, (c) that no refund was received, and (d) that suit was timely under limitations periods running from two to six years — none of which is provable from Alderwood's records" is an argument.

Phase 5 — Individualized reliance, causation, and injury

  • [ ] Establish the number of distinct representations, packages, and channels during the class period.
  • [ ] Establish that many purchasers never saw the challenged claim (point-of-sale placement, online listings, third-party retailers, repeat auto-ship purchases).
  • [ ] Establish the diversity of purchase motives with company survey data, reviews, and plaintiff testimony.
  • [ ] In securities cases, attack the Basic presumption directly with price-impact evidence.
  • [ ] Resist any "Trial by Formula" — the plaintiff cannot try a sample and extrapolate liability to unexamined members.

Why it matters, and the traps

Reliance is where consumer classes go to die. Where the governing state law requires individual reliance and the record shows that purchasers encountered the product through different labels, different ads, and different channels, no common evidence can establish that each member was deceived.

Alderwood, applied. Discovery produces: nine label variants over the four-year class period, of which three never carried the "clinically shown" claim; a marketing deck showing that 44% of unit sales flowed through third-party retail listings that did not reproduce the label copy; a customer-survey vendor's data showing "recommended by a friend" as the top purchase driver; and a lead plaintiff who testified she bought MetaboLift because her sister used it and only read the label "later, maybe." Under the consumer statutes of the twenty-two class-member states requiring actual reliance, there is no way to prove deception without asking each purchaser what she saw and why she bought. Predominance fails.

Two doctrinal guardrails. Dukes rejected "Trial by Formula" — the plaintiff may not try a sample set of claims and apply the average across the class to establish liability. But Tyson Foods, Inc. v. Bouaphakeo, 577 U.S. 442 (2016), held that representative statistical evidence is permissible in a class case where the same evidence would be admissible to prove an individual plaintiff's case (there, a time study in an FLSA donning-and-doffing suit where the employer kept no records). The distinction is whether the sample substitutes for absent proof of an element that varies, or supplies a fact that is genuinely uniform. Brief that line carefully; plaintiffs will read Tyson Foods for everything it is worth.

In securities cases, the fight has its own vocabulary. Basic Inc. v. Levinson, 485 U.S. 224 (1988), supplies a rebuttable fraud-on-the-market presumption of reliance; Halliburton Co. v. Erica P. John Fund, Inc., 573 U.S. 258 (2014), holds that a defendant may rebut it at the certification stage with evidence of no price impact; and Goldman Sachs Group, Inc. v. Arkansas Teacher Retirement System, 141 S. Ct. 1951 (2021), directs courts to consider the generic nature of the alleged misrepresentations in assessing price impact. See Securities Fraud Litigation: A Comprehensive Guide to Rule 10b-5 Private Actions for the full architecture.

Phase 6 — Individualized affirmative defenses and limitations

  • [ ] Map limitations periods across the class-member states and the accrual rules (discovery rule, fraudulent concealment, tolling).
  • [ ] Identify members who signed releases (refund conditions, settlement of prior disputes, employment separations).
  • [ ] Identify members bound by arbitration agreements even if the named plaintiff is not.
  • [ ] Identify members with assumption-of-risk, comparative-fault, or misuse defenses (common in products cases) — and confirm how each class-member state actually treats them.
  • [ ] Show that each defense requires member-specific evidence.

Why it matters, and the traps

Individualized defenses count toward the predominance calculus every bit as much as individualized elements. If half the class is time-barred unless each member proves she could not reasonably have discovered the claim, the court must run that inquiry one member at a time. If Alderwood's refund program required a signed release, every refunded purchaser presents a release defense.

A structural point that defense counsel underuse: the defenses themselves are state law, and they are not uniform. Rule 23 is federal, but the substantive claims and defenses in most consumer, tort, products-liability, and insurance-coverage class actions are creatures of state law, and the states genuinely diverge. There is no nationwide rule of tort or products liability, and any brief that asserts one will be corrected by the court. A few of the fault lines that matter at certification:

  • Comparative versus contributory fault. Most states apply some form of comparative fault (pure or modified, with different bar thresholds); a small handful retain contributory negligence as a complete bar. The defense's effect on a class member's recovery therefore depends on which state's law governs her claim.
  • Design-defect tests. Some states apply a consumer-expectations test, some a risk-utility test, some both or a hybrid; the Restatement (Third) of Torts: Products Liability § 2 approach is influential but is not the law everywhere. Which test applies changes what proof is common and what is individualized.
  • Failure-to-warn and learned-intermediary doctrine. Adopted in most states, but with meaningful variation in scope and exceptions.
  • Insurance-coverage classes. Coverage, bad faith, and the duty to defend are state-regulated and state-specific; a nationwide coverage class almost always founders on that variation.

That variation is not merely a caveat — it is an argument. If the defenses to a nationwide class are drawn from fifty different bodies of law, the individualized-issues catalog in Phase 4 has to be built state by state, and the plaintiff's law-application plan has to survive it. Verify each governing state's rule; do not extrapolate from the state you practice in.

Two traps. First, preservation: an affirmative defense not pleaded is an affirmative defense you cannot invoke at certification. Second, overreach: courts are alert to defendants who conjure defenses that the defendant plainly never intended to press against thousands of small-dollar members. Show that you actually will press them — that the defense is real, individualized, and outcome-determinative. In products cases the individualized-defense argument is unusually strong; see Defending a Products Liability Claim: A Practical Checklist.

Phase 7 — Uninjured class members after TransUnion

  • [ ] Quantify the uninjured population with hard data (refunds, non-exposure, satisfied users, resellers).
  • [ ] Argue that Article III requires every member of a damages class to have suffered concrete harm.
  • [ ] Argue that identifying the injured requires individualized inquiry — which is itself a predominance failure.
  • [ ] Check your circuit's rule on the tolerated percentage of uninjured members.

Why it matters, and the traps

TransUnion LLC v. Ramirez, 594 U.S. 413 (2021), held that "[e]very class member must have Article III standing in order to recover individual damages," and that a bare statutory violation without concrete harm will not do — building on Spokeo, Inc. v. Robins, 578 U.S. 330 (2016). The 6,332 class members whose inaccurate OFAC alerts were never disseminated to a third party lacked standing; only the 1,853 whose reports were disseminated could recover.

The doctrine cuts two ways for the defense, and the second way is the sharper one. Yes, uninjured members cannot recover. But more usefully: if determining who is injured requires a member-by-member inquiry, the class cannot be tried at all. The circuits have not converged on how many uninjured members are too many — the Ninth Circuit sitting en banc in Olean Wholesale Grocery Cooperative v. Bumble Bee Foods, LLC (9th Cir. 2022) tolerated a de minimis number while emphasizing that a district court cannot certify where a damages model sweeps in a substantial uninjured population, and the D.C. Circuit's In re Rail Freight Fuel Surcharge Antitrust Litigation line refused certification where the plaintiffs' regression predicted damages for thousands of shippers who suffered none. Know your circuit's number, and if there is none, argue the principle. Antitrust practitioners will recognize the fight; see Antitrust Law: A Comprehensive Guide to the Sherman, Clayton, and FTC Acts.

Phase 8 — Attack the damages model under Comcast

  • [ ] Identify the plaintiff's theory of liability with precision, then read the damages model against it.
  • [ ] Show the mismatch: the model measures harm from conduct the court did not (or will not) certify, or bundles multiple theories the model cannot disaggregate.
  • [ ] Show the model cannot distinguish injured from uninjured members.
  • [ ] Show the model requires member-specific inputs (what each purchaser would have paid, what each would have done) that no common evidence supplies.
  • [ ] Retain your own economist to run the counter-model and expose the assumptions.

Why it matters, and the traps

Comcast Corp. v. Behrend, 569 U.S. 27 (2013), reversed certification because the plaintiffs' regression model calculated damages from four theories of antitrust impact when only one had been certified. The model, the Court said, "identifie[d] damages that are not the result of the wrong" — and a model that cannot be tied to the certified liability theory cannot support predominance.

The Comcast argument is the most technically demanding piece of the opposition and the most rewarding. Conjoint analyses and price-premium regressions in consumer cases routinely fail on their own terms: they measure willingness-to-pay in a hypothetical world without modeling supply-side response, they assume every purchaser valued the claim identically, and they cannot say which purchasers would have bought anyway. In Alderwood, the plaintiff's conjoint survey yields a 14% price premium attributable to the "clinically shown" claim — but the model assumes universal exposure to a claim that 44% of purchasers never saw, and it cannot price the refunded purchasers out. That is Comcast squarely.

The trap is overreading Comcast. Most circuits hold that damages need not be perfectly common; individualized damages calculations alone rarely defeat predominance. The winning argument is never "damages vary." It is "the model does not measure the injury this theory of liability produces, and it cannot separate the injured from the uninjured." For the underlying valuation concepts, see The Litigation Damages Toolkit: Valuing and Proving Money Awards.

Phase 9 — Exclude the plaintiff's expert under Daubert and Rule 702

  • [ ] Move to exclude the certification expert on a full Daubert record — do not treat it as a weight-not-admissibility skirmish.
  • [ ] Attack methodology: was the survey population the right one, the control adequate, the regression specified correctly, the data reliable?
  • [ ] Attack fit: does the opinion actually answer the classwide question, or a different one?
  • [ ] Invoke the December 2023 amendment to Fed. R. Evid. 702, which restates that the proponent must show admissibility by a preponderance and that the opinion must reflect a reliable application of the methodology to the facts.
  • [ ] If the expert falls, say so plainly: without the model, predominance collapses.

Why it matters, and the traps

Certification usually rides on one expert. Take that expert out and the motion dies with him. The Supreme Court has not squarely resolved whether a full Daubert inquiry is required at certification, but Dukes pointedly doubted the district court's contrary view, and the Seventh Circuit in American Honda Motor Co. v. Allen (7th Cir. 2010) held that where an expert's report is critical to certification, the court must conduct a full Daubert analysis before ruling. Most courts now do. Brief it as settled; do not concede a lesser standard.

The trap is treating the Daubert motion as an afterthought filed with the opposition brief. It needs its own record — the expert's deposition, his prior testimony, his data files, his code. Work it with Challenging an Expert Under Daubert and Rule 702: A Practical Checklist and, for the evidentiary framework, The Federal Rules of Evidence: A Comprehensive Guide.

Phase 10 — Superiority, manageability, and the missing trial plan

  • [ ] Work the four Rule 23(b)(3) factors: members' interest in individual control, existing litigation, desirability of the forum, and manageability.
  • [ ] Demand a trial plan. Ask, in the brief and at argument: how would this actually be tried?
  • [ ] Show that the plan requires thousands of mini-trials, or that it violates the Seventh Amendment or the Rules Enabling Act by depriving the defendant of the right to present individualized defenses.
  • [ ] Argue that available alternatives — the company's refund program, a regulatory remedy, individual small-claims actions with fee-shifting — make a class inferior.
  • [ ] Resist Rule 23(c)(4) issue certification as an end-run around predominance.

Why it matters, and the traps

Manageability is where all the earlier arguments come home. A judge who has read your individualized-issues table and your Comcast attack, and who then asks plaintiff's counsel "how do we try this?" and hears an answer about "streamlined claims administration," has your answer.

The Rules Enabling Act point, 28 U.S.C. § 2072(b), deserves more airtime than it gets: Rule 23 cannot "abridge, enlarge or modify any substantive right," and a trial plan that prevents the defendant from litigating a valid individual defense against each claimant does exactly that. Dukes said as much in rejecting Trial by Formula.

Expect the plaintiff's fallback: certify liability issues under Rule 23(c)(4) and leave damages for individual proceedings. Answer it on two fronts — the issues proposed are not genuinely common, and a bifurcated structure risks a Seventh Amendment reexamination problem when a second jury must revisit facts the first one found.

Phase 11 — Build the record: discovery is the opposition

  • [ ] Serve targeted written discovery on the named plaintiffs on exposure, reliance, purchase history, refunds, and litigation history; see Serving Written Discovery Requests: A Practical Checklist.
  • [ ] Seek leave for discovery from absent class members where variation among them is the core defense — courts allow it where it is not unduly burdensome and goes to certification, not merits harassment.
  • [ ] Mine your own data: label change logs, channel-mix reports, refund and warranty data, customer surveys, review-site sentiment, ad-buy records.
  • [ ] Depose the plaintiff's expert before the opposition is due.
  • [ ] Preserve the sequencing fight: resist a certification schedule that denies you the discovery you need to oppose.
  • [ ] Move to compel when stonewalled — see Moving to Compel Discovery and Seeking Rule 37 Sanctions: A Practical Checklist.

Why it matters, and the traps

Certification is decided on evidence. Dukes demands "significant proof," and courts have grown impatient with defendants who oppose certification with adjectives. Every proposition in your brief should cite a deposition line, a business record, or an expert declaration.

The trap is your own data. Bifurcated certification discovery is a trap when the facts that defeat predominance live in your systems and your opponent has not asked for them — but it is also a trap when your data shows uniformity you have not looked at. Run the analysis internally before you commit to the individualized-variation theory. For the discovery architecture, see Federal Civil Discovery: Mastering the Tools, Rules, and Pitfalls and The Federal Discovery Toolkit: A Roadmap From Rule 26(f) to Sanctions.

Phase 12 — The Rule 23(b)(2) misuse argument

  • [ ] Ask whether the plaintiff seeks money. If so, (b)(2) is almost certainly the wrong vehicle.
  • [ ] Show the requested injunction is not "indivisible" — that it does not provide relief to each member in a single stroke.
  • [ ] Show the monetary relief is not incidental to the injunctive relief.
  • [ ] Note that a (b)(2) class has no notice or opt-out right — which is exactly why the Constitution will not tolerate individualized damages inside it.

Why it matters, and the traps

Dukes held that Rule 23(b)(2) "does not authorize class certification when each class member would be entitled to an individualized award of monetary damages," and that backpay was not incidental. Plaintiffs still try to smuggle restitution, disgorgement, and statutory damages into (b)(2) classes to escape predominance and notice. Call it what it is. And watch for the hybrid — a (b)(2) injunctive class plus a (b)(3) damages class — which requires you to run the full predominance analysis on the (b)(3) piece anyway.

Phase 13 — If you lose: Rule 23(f), decertification, and settlement posture

  • [ ] Calendar the 14-day deadline in Fed. R. Civ. P. 23(f) from entry of the certification order. It is not extendable by stipulation.
  • [ ] Frame the petition around the recognized grounds: a death-knell effect on the defendant, an unsettled and fundamental legal question, or a manifest error.
  • [ ] Consider a stay of district-court proceedings pending the petition.
  • [ ] Preserve decertification under Rule 23(c)(1)(C) — an order may be altered or amended before final judgment as the record develops.
  • [ ] Renew the attack at summary judgment; see Moving for Summary Judgment Under Rule 56: A Practical Checklist.
  • [ ] Re-price the case honestly and consider a classwide resolution — see Settling a Class Action: A Rule 23(e) Approval Checklist.

Why it matters, and the traps

Rule 23(f) exists because certification orders are effectively final as a practical matter — the point Coopers & Lybrand v. Livesay, 437 U.S. 463 (1978), refused to make through the "death knell" doctrine and Congress and the Rules Committee later addressed by rule. Review is discretionary; the petition must sell the appellate court on why this order deserves interlocutory attention.

The fatal trap: do not try to manufacture appellate jurisdiction by dismissing your own case. Microsoft Corp. v. Baker, 137 S. Ct. 1702 (2017), holds that plaintiffs cannot create a final judgment by voluntarily dismissing after a certification denial; the reasoning cuts against defense-side gamesmanship too. Take the 23(f) route, and take it in fourteen days. On the mechanics of appellate practice, see Taking a Civil Appeal: A Practical Checklist and Federal Civil Appeals: A Comprehensive Guide.

Common Mistakes to Avoid

  1. Arguing the merits instead of certification. "Our label was true" is a summary-judgment argument. The certification argument is "whether it was true cannot be tried in one stroke."
  2. Briefing the wrong circuit's ascertainability rule. Heightened administrative feasibility is the law in some circuits and expressly rejected in others. Check.
  3. Sleeping on the arbitration clause. Litigate for a year and you have waived the cheapest win available.
  4. Making an offer of judgment to pick off the named plaintiff. Campbell-Ewald closed that door.
  5. Opposing with argument, not evidence. Dukes requires significant proof. Adjectives lose.
  6. Overreading Comcast. "Damages vary" is not a predominance defense. Mismatch between model and liability theory is.
  7. Conceding the Daubert standard at certification. Insist on a full analysis where the expert is critical.
  8. Forgetting the affirmative defenses in the answer. Unpleaded defenses cannot power a predominance argument.
  9. Ignoring Rule 23(c)(4). If you do not brief issue certification, the plaintiff will get it as a consolation prize.
  10. Missing the 14-day Rule 23(f) window. There is no relief from it.
  11. Attacking only the named plaintiff. Knock out one and the plaintiff amends. Attack the structure.
  12. Discovering your own data too late. Certification is won in the company's systems long before it is won in the brief.
  13. Treating state substantive law as if it were national. Tort, products liability, and insurance law vary by state — sometimes decisively. Asserting a nationwide rule concedes credibility and forfeits your best predominance argument, which is precisely that the law is not uniform.
  14. Relying on stale doctrine or stale numbers. Chevron deference is no longer good law after Loper Bright (2024); HSR thresholds, tariff rates, and sanctions designations change constantly. Verify before you cite.

The Consolidated Defense Checklist

Threshold

  • [ ] Arbitration clause + class waiver identified, assent provable, motion to compel filed early
  • [ ] Rule 12 motion evaluated; class-strike motion considered where the defect is facial
  • [ ] Named plaintiff's Article III standing tested
  • [ ] All affirmative defenses pleaded

The class

  • [ ] Definition analyzed for fail-safe, overbreadth, subjectivity
  • [ ] Circuit's ascertainability rule identified and applied
  • [ ] Membership-identification method shown to be unworkable (or shown to require individual inquiry)

The representatives

  • [ ] Every named plaintiff deposed
  • [ ] Unique defenses, credibility problems, and knowledge gaps documented
  • [ ] Rule 23(g) adequacy of counsel and intra-class conflicts assessed

Rule 23(a) and (b)

  • [ ] Commonality attacked where the theory rests on decentralized discretion (Dukes)
  • [ ] Individualized-issues catalog built element by element, claim by claim, defense by defense
  • [ ] Choice-of-law variation charted for any multistate class
  • [ ] Reliance, causation, and injury shown to be individualized
  • [ ] Limitations and release defenses shown to require member-specific proof
  • [ ] Uninjured-member population quantified (TransUnion)
  • [ ] Damages model attacked for mismatch and for sweeping in the uninjured (Comcast)
  • [ ] Certification expert challenged under Daubert and Rule 702
  • [ ] Superiority and manageability briefed; trial plan demanded
  • [ ] Rule 23(b)(2) misuse and Rule 23(c)(4) end-run addressed

Record and aftermath

  • [ ] Certification discovery obtained, including from absent members where warranted
  • [ ] Plaintiff's expert deposed before opposition filed
  • [ ] Rule 23(f) petition calendared at 14 days; stay considered
  • [ ] Decertification under Rule 23(c)(1)(C) preserved; summary judgment planned

Primary Authority

Rules

Cases

Circuit law to check, not assume. Ascertainability doctrine diverges materially: the Third Circuit (Carrera v. Bayer Corp.) requires administrative feasibility, while the Seventh (Mullins v. Direct Digital, LLC) and Ninth (Briseno v. ConAgra Foods, Inc.) reject a freestanding requirement. The tolerance for uninjured class members likewise varies (compare the Ninth Circuit's en banc decision in Olean Wholesale Grocery Cooperative v. Bumble Bee Foods, LLC with the D.C. Circuit's In re Rail Freight Fuel Surcharge Antitrust Litigation line). Confirm your circuit's rule before you brief.

Secondary sources. Wright & Miller, Federal Practice and Procedure §§ 1751–1805 (class actions); Newberg and Rubenstein on Class Actions; Manual for Complex Litigation (Fourth) § 21 (class action procedures); Restatement (Third) of Torts: Products Liability (for the substantive-law variation discussed in Phase 6).

Jurisdiction, currency, and what to verify before you rely on it

Certification practice is federal, but almost everything you will be arguing about is not uniform. Before you brief, verify:

  1. Your circuit's law, not the reported cases you happen to remember. Ascertainability, the tolerated share of uninjured class members, whether a full Daubert analysis is required at certification, and the appetite for Rule 23(c)(4) issue certification all differ by circuit. The divergences described above are real and material.
  2. State substantive law, claim by claim and defense by defense. Tort, products liability, and insurance law are state law. There is no national rule of design defect, no national rule of comparative fault, no national rule of bad-faith coverage, and no national consumer-protection statute. Land use and zoning — which occasionally generates putative classes of neighboring owners — is likewise state and local law, governed by the enabling act, the local code, and state constitutional takings doctrine. Confirm each governing state's rule and never generalize from one.
  3. Administrative-law premises, if your class claim touches an agency's interpretation. Do not build an argument on Chevron deference: it was overruled by Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024). Courts now exercise independent judgment on questions of statutory meaning, giving an agency's view only Skidmore-style respect for its persuasiveness. Related current authority includes Corner Post, Inc. v. Board of Governors of the Federal Reserve System (2024) (APA limitations accrual) and SEC v. Jarkesy (2024) (Seventh Amendment jury right for civil penalties). If a brief or form you inherited leans on Chevron, it is out of date.
  4. Anything with a moving number. Thresholds, rates, and lists change constantly and should never be cited from memory or from a document of this vintage: Hart-Scott-Rodino reporting thresholds and filing-fee tiers are adjusted annually; tariff rates and trade-remedy duties change by executive and agency action; and OFAC sanctions designations (the SDN List — the subject matter of TransUnion itself) are updated continuously. If a class theory depends on any of them, pull the current official figure and confirm the effective date before you rely on it.
  5. The rules themselves. Fed. R. Evid. 702 was amended effective 1 December 2023, and Rule 23 has been amended repeatedly. Check the current text.

Related UpLaw Resources

Related Guides & Articles

Related Checklists

Related Toolkits

This checklist is general legal information from UpLaw (uplaw.us), not legal advice, and does not create an attorney-client relationship. Laws vary by state and circuit and change over time; verify current requirements and consult qualified counsel before acting.