FORMER OFFICIALS OF THE U.S. SECURITIES AND EXCHANGE COMMISSION AND LAW PROFESSORS; THE SECURITIES INDUSTRY AND FINANCIAL MARKETS ASSOCIATION, Amici Supporting Appellants.

Appeal from the United States District Court for the Eastern District of Virginia, at Alexandria. Leonie M. Brinkema, District Judge. (1:24-cv-00151-LMB-LRV)

Argued: May 7, 2026 Decided: July 20, 2026

Before RICHARDSON, QUATTLEBAUM, and RUSHING, Circuit Judges.

Page 2Reversed and remanded by published opinion. Judge Quattlebaum wrote the opinion, in which Judge Richardson and Judge Rushing joined.

ARGUED: Jeffrey Bryan Wall, GIBSON, DUNN & CRUTCHER LLP, Washington, D.C., for Appellants. Deepak Gupta, GUPTA WESSLER LLP, Washington, D.C., for Appellees. ON BRIEF: Benjamin L. Hatch, MCGUIREWOODS LLP, Washington, D.C.; Judson O. Littleton, Washington, D.C., Richard C. Pepperman II, Jacob E. Cohen, SULLIVAN & CROMWELL LLP, New York, New York, for Appellants. Carol C. Villegas, Christine M. Fox, Jake Edward Bissell-Linsk, LABATON KELLER SUCHAROW LLP, New York, New York; Gregory Beck, GUPTA WESSLER LLP, Washington, D.C.; Chad Johnson, Noam Mandel, Jonathan Zweig, Desiree Cummings, New York, New York, Douglas Wilens, ROBBINS GELLER RUDMAN & DOWD LLP, Boca Raton, Florida, for Appellees. Todd G. Cosenza, Brady Sullivan, Amanda M. Payne, WILLKIE FARR & GALLAGHER LLP, New York, New York, for Amici Former Officials of the United States Securities and Exchange Commission and Law Professors. Kevin M. Carroll, THE SECURITIES INDUSTRY AND FINANCIAL MARKETS ASSOCIATION, Washington, D.C.; Benjamin W. Snyder, Vladimir J. Semendyai, PAUL HASTINGS LLP, Washington, D.C., for Amicus The Securities Industry and Financial Markets Association.

Page 3QUATTLEBAUM, Circuit Judge:

This appeal involves the requirements for class certification under Rule 23 of the Federal Rules of Civil Procedure. Certain shareholders of aerospace manufacturer The Boeing Company seek to hold that company and several of its former officers liable for violating Sections 10(b) and 20(a) of the Securities Exchange Act. In the plaintiffs’ view, Boeing repeatedly misrepresented its emphasis on safety in the wake of two deadly airplane crashes. These false and misleading statements, they contend, artificially inflated or maintained artificial inflation in Boeing's stock price. The plaintiffs allege that a subsequent in-flight safety incident and disclosures revealed the truth—that Boeing wasn’t actually prioritizing safety. As a result, the artificial inflation dissipated, and the plaintiffs claim they suffered billions in losses. They seek to recover these losses as a class.

For that to happen, the plaintiffs must first satisfy Rule 23. That rule operates as a procedural safeguard to balance the benefits of proceeding as a class—the primary one being the efficient resolution of numerous disputes that share common questions—with the cost of subjecting defendants to numerous individual claims of plaintiffs who may not have prosecuted claims on their own. Several Supreme Court decisions explain what plaintiffs must do to show class certification is appropriate and how district courts are to determine whether class certification is, in fact, merited under Rule 23.

One of those decisions, Comcast Corp. v. Behrend , 569 U.S. 27 (2013), takes center stage in this appeal. Comcast concerns class damages. It requires plaintiffs, at the classcertification stage, to put forth evidentiary proof of a damages methodology showing how damages are calculable on a class-wide basis in a manner that is consistent with their theory Page 4of liability and that is not speculative. Comcast also requires district courts to conduct a rigorous analysis to determine whether plaintiffs have satisfied this burden. Here, the district court certified a class after concluding that the plaintiffs satisfied Comcast and Rule 23. But the plaintiffs did not provide a damages methodology consistent with Comcast 's commands. And the district court did not conduct the rigorous analysis Comcast requires. So, the district court erred in certifying the class. We reverse and remand.

To frame the issues, we start with some basics on class actions before turning to plaintiffs' factual allegations.¹ Then, we outline the procedural history of this case, describing both the motion-to-dismiss and class-certification proceedings below.

The way litigation traditionally works in the United States, a person or entity that claims to have been wronged must file a complaint, serve it on the alleged wrongdoer and then prosecute the case. Class actions, though, are different. Authorized under Rule 23, 2 a Page 5class action is a “lawsuit in which the court authorizes a single person or a small group of people to represent the interests of a larger group.” Class Action , Black’s Law Dictionary (12th ed. 2024). In other words, a party does not have to file and prosecute their case to recover damages. If it is similar enough to other claims, a representative plaintiff can file the suit and prosecute it on behalf of the group.

Class actions involve 'a seemingly endless tug of pros and cons.' Stafford v. Bojangles’ Rests., Inc. , 123 F.4th 671, 678 (4th Cir. 2024). “On the one hand, class actions offer the promise of resolving many similar suits at a single time.” Id. On the other hand, proceeding as a class implicates due process concerns—“absent class members’ grievances are litigated not directly by the class members themselves but by class representatives,” and “[d]efendants in those actions can be sued by a class of hundreds or thousands of people Page 6who are relieved of their normal responsibility to file and prosecute their own case.' Sharp Farms v. Speaks , 917 F.3d 276, 305 (4th Cir. 2019) (Quattlebaum, J., concurring). In addition, “class-action lawsuits can ratchet liability to potentially ruinous levels and force companies to settle or bet the store.” Stafford , 123 F.4th at 678.

Rule 23 seeks to balance these interests. It “is a structural safeguard in our legal system developed to provide the benefits of class actions while simultaneously avoiding the risks associated with them.' Speaks , 917 F.3d at 305 (Quattlebaum, J., concurring); see also Stafford , 123 F.4th at 678. As a result, plaintiffs wishing to pursue a class action 'must comply” with Rule 23. Stafford , 123 F.4th at 678. When they do, the district court may certify a class, allowing the class action to proceed. See Fed. R. Civ. P. 23(c)(1)(A).

There are two categories of Rule 23 requirements plaintiffs must satisfy-those under Rule 23(a) and those under Rule 23(b). See Fed. R. Civ. P. 23(a)–(b). Rule 23(a) sets out four “[p]rerequisites” that every class action must meet—numerosity, commonality, typicality and adequacy of representation. Fed. R. Civ. P. 23(a)(1)–(4).³ Rule 23(b) adds to these requirements, providing that a class action may only proceed if it falls into one of three different categories—(1) proceeding without a class would create inconsistencies or variance, be dispositive of non-class members’ interests or impair or impede these interests; (2) injunctive or declaratory relief is appropriate on a class-wide basis; or (3) common Page 7questions of law and fact predominate over individual questions and class adjudication is superior to individual adjudication. Fed. R. Civ. P. 23(b)(1)–(3).

This case concerns whether class certification was proper under Rule 23(b)(3). This subsection allows a class to maintain an action if “the court finds that the questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.' Fed. R. Civ. P. 23(b)(3). In this way, Rule 23(b)(3) seeks “to cover cases in which a class action would achieve economies of time, effort, and expense, and promote uniformity of decision as to persons similarly situated, without sacrificing procedural fairness or bringing about other undesirable results.' Amchem Prods., Inc. v. Windsor , 521 U.S. 591, 615 (1997) (citation modified). But a party proceeding under this subsection can’t just say so; it must actually satisfy its requirements.

Boeing manufactures aircraft for civilian and military uses.⁴ And it's really only one of two commercial airplane builders. It and Airbus SE manufacture 99% of large Page 8commercial airplanes. So, when Airbus announced the next generation of its A320 plane, the A320neo, in December 2010, Boeing felt pressure to keep up. To do so, Boeing decided to update its best-selling 737 model. It first delivered its new airplane—the 737 MAX—to customers in May 2017.

In October 2018, Lion Air Flight JT 610, flying a Boeing 737 MAX, crashed shortly after takeoff, resulting in the loss of all on board. In March 2019, the same thing happened to Ethiopian Airlines Flight ET 302, flying another 737 MAX. That month, the Federal Aviation Administration grounded all 737 MAX airplanes. Investigations into these events uncovered a software issue. A defective sensor reported the planes’ noses were too high, which activated the planes’ Maneuvering Characteristics Augmentation System and drove their noses down.

Congress got involved. The United States House of Representatives’ Committee on Transportation and Infrastructure determined, among other things, that costs, scheduling and other pressures at Boeing undermined the safety of the new aircraft and led to safety compromises.

So did the Department of Justice and the Securities and Exchange Commission. On January 7, 2021, the DOJ and Boeing entered a three-year deferred prosecution agreement, which included a $2.5 billion settlement, to resolve a charge that Boeing conspired to defraud the FAA. And the SEC and Boeing later agreed to a $200 million settlement over Page 9allegations Boeing misled investors about the safety of its 737 MAX aircraft after the crashes.

During this time, Airbus gained competitive ground. Its market share in the commercial aircraft market increased from 45.3% in 2018 to 62.5% in 2019. And Boeing’s revenues temporarily dipped at around that same time—from $101 billion in 2018 to $58.1 billion in 2020—before rebounding somewhat to $77.8 billion in 2023.

Trying to turn things around, Boeing committed to improving the safety of its airplanes. For instance, in April 2019, it announced a board committee to examine safety and quality control policies. That September, Boeing adopted the committee’s recommendations in full, including the creation of a board committee on aerospace safety, a product and services safety organization and a design requirements program. Then, in January 2020, Boeing highlighted revisions to its codes of conduct, emphasizing that it expected its employees to report concerns and that it would not retaliate against those who did.

Another announced change involved a practice called traveled work. Traveled work happens when an airplane needs to move down the assembly line so as to not hold up production but work on the plane at its current station on the line isn’t complete. To prevent delays, the plane moves down the line anyway, and incomplete work is completed, or repairs are made, later on. Boeing said it was eliminating traveled work because it negatively affected productivity and quality.

Between January 7, 2021 and January 8, 2024, Boeing’s officers “repeatedly” made statements about safety. See, e.g. , J.A. 81. These are the allegedly false and misleading Page 10statements that form the basis of this lawsuit. According to the plaintiffs, there are 40 relevant statements that span three years in time.⁵ We won't list all of them here, but to provide a flavor for the various comments, here are a few examples—as categorized by the plaintiffs:

• · March 5, 2021 - 'Safety is, simply put, our highest priority. We are deeply committed to strengthening our safety processes through continuous improvement, learning and innovation.” J.A. 83.⁶
• · November 2, 2022 – “Safety dominates Boeing.” J.A. 84.

• · March 3, 2023 – “We are working together to foster a culture of continuous improvement and enhance performance by creating an environment where employees are comfortable identifying gaps, seeking help and speaking up without fear of retaliation.” J.A. 88.

• · July 28, 2021 – “Turning to our efforts to drive stability, with every action we are driving toward engineering excellence, production system stability, and firsttime quality and delivery predictability while holding ourselves accountable to the highest standards.” J.A. 89.
• · April 27, 2022 – “On the supply side we are carefully managing supply chain constraints and working through issues as they arise to ensure the stability of our production system.” J.A. 90.

Page 114. Statements about Regulatory Compliance

• · January 31, 2022 - 'We also are fully cooperating with U.S. government investigations related to the accidents and the 737 MAX, including an ongoing investigation by the Securities and Exchange Commission.” J.A. 92.

According to the plaintiffs, these statements, and others like them, weren’t true. To explain why, they point to several whistleblowers. One whistleblower recounted how Boeing continued traveled work after the Lion Air and Ethiopian Airlines accidents, even though it said multiple times that it planned to eliminate that practice. Others described additional problems, including manufacturing shortcuts, unmet industry standards, airplane misalignment, improper quality control, work performed by unlicensed technicians, unimplemented safety improvements and continued retaliation against reporting employees.

As the plaintiffs see it, the truth began to emerge on the night of Friday, January 5, 2024, when Alaska Airlines Flight 1282-flying a 737 MAX-made an emergency landing after a door plug in the plane’s fuselage detached minutes following takeoff. On the next day of trading, January 8, Boeing’s stock price dropped 8%.

The loss of the door plug was caused in part by traveled work. Boeing had identified bad parts when the airplane’s fuselage arrived at Boeing’s manufacturing plant. Yet Boeing didn’t fix the issue when first noticed, so the 737 MAX went to the next station. When Page 12Boeing made the necessary repairs 19 days later, workers did not replace four bolts on the door plug they opened. This was the door plug that blew off midflight.⁷

The plaintiffs’ allegations give rise to two claims under the Securities Exchange Act. The first claim alleges the defendants violated Section 10(b) and its accompanying regulation, Rule 10b–5. And the second claim asserts violations of Section 20(a) against the individual defendants.

Section 10(b) and Rule 10b–5 “prohibit material misrepresentations and omissions in connection with the sale of securities.” Goldman Sachs Grp., Inc. v. Ark. Tchr. Ret. Sys. , 594 U.S. 113, 117–18 (2021).⁸ 'Relatedly, section 20(a) is the vehicle for imposing liability Page 13on control persons,' like officers. Singer v. Reali , 883 F.3d 425, 438 (4th Cir. 2018). “The liability of a control person under section 20(a) is derivative of—and dependent upon— liability of a controlled person under section 10(b).' Id. In other words, the plaintiffs' success rises and falls with their Section 10(b) claim. Section 10(b) claims have six elements—“(1) a material misrepresentation or omission by the defendant; (2) scienter; (3) a connection between the misrepresentation or omission and the purchase or sale of a security; (4) reliance upon the misrepresentation or omission; (5) economic loss; and (6) loss causation.” Id. (quoting Stoneridge Inv. Partners, LLC v. Sci.-Atlanta, Inc. , 552 U.S. 148, 157 (2008)).

Here, the plaintiffs asserted their Section 10(b) claim in the form of a class action. They proposed a class period running from September 30, 2019-the date of the first allegedly false and misleading statement—until May 14, 2024—the date of the last alleged exposure. The plaintiffs identified dozens of purportedly material, false and misleading statements made by Boeing and its officers with the intent to deceive the investing public over three and a half years. They alleged they relied on all these statements when purchasing Boeing stock.⁹ And they claimed these statements 'artificially inflated and/or Page 14artificially maintained” Boeing’s share price, causing them to lose money when the share price “declined as artificial inflation was removed from [the] share price” “[a]s truthful information was revealed to the market and/or as the previously undisclosed, understated, or misrepresented risks materialized.” J.A. 370.

Boeing moved to dismiss the amended complaint. According to Boeing, each allegedly false and misleading statement-which it said fell into one or more of six categories—wasn’t actionable. Memorandum in Support at 8–19, In re The Boeing Co. Sec. Litig. , No. 1:24-cv-00151-LMB-LRV (E.D. Va. June 21, 2024), ECF No. 51. Boeing contended three categories of statements—(1) general statements about safety and quality; (2) aspirational statements in Boeing’s codes of conduct; and (3) statements concerning opinions, beliefs and future intentions—lacked materiality because they were too general and subjective. Id. at 10-15. The other three categories concerned statements about Boeing’s production and delivery, statements about disclosure of risk factors in SEC filings and statements the plaintiffs said limited the FAA's recertification process for the 737 MAX to the Maneuvering Characteristics Augmentation System when it really involved not only that system but also additional airplane features. Id. at 15-19. In Boeing's view, Page 15these statements weren’t actionable because it hadn’t said anything false or misleading about its production and delivery, the plaintiffs hadn’t explained why its SEC disclosures were false and misleading with particularity and its statements never limited the FAA recertification process, among other reasons. Id.

In response, the plaintiffs stressed that while aspirational puffery and opinion often don’t meet the materiality requirement of a Section 10(b) claim, some courts have found that such statements can in contexts like the airline industry, where safety is paramount, or when made to reassure investors. And they emphasized repetition may make otherwise immaterial statements material. Then, the plaintiffs addressed each of Boeing's six categories of purportedly immaterial statements, insisting each statement in those categories was material.

The district court addressed Boeing’s motion during a hearing. It denied the motion without argument after stating only that it thought the “extensive amount of detail” in the amended complaint 'adequately hit[] all the necessary requirements to let the case go forward,” including “that there ha[d] been material misstatements made to the investing public.” J.A. 519.¹⁰

Having survived Boeing’s motion to dismiss, the plaintiffs moved to certify a class. In their motion, they described their class claims as based on a single course of conduct by Page 16Boeing—the making of false and misleading statements which caused Boeing’s stock to trade at artificially high levels. Memorandum in Support at 15, In re The Boeing Co. Sec. Litig. (E.D. Va. Dec. 13, 2024), ECF No. 104. The plaintiffs said Boeing's 'many misstatements artificially inflated and artificially maintained the price of Boeing’s stock.” Id. at 2; see also id. at 9 (noting common questions included 'whether, and to what extent, the market price of Boeing common stock was artificially inflated or maintained during the Class Period'). According to the plaintiffs, this inflation dissipated when the Alaska Airlines incident and other corrective events revealed the truth. Id. at 2.

On the issue of damages, the plaintiffs pointed to Comcast as setting the proper standard. Id. at 27. They argued they satisfied Comcast by invoking “the ‘out of pocket’ methodology” provided by an economist, Chad Coffman. Id. 11 The plaintiffs contended '[t]his approach matche[d] [their] theory of liability.' Id .

Coffman offered his damages methodology in a report filed with the plaintiffs' motion for class certification. Its damages discussion was brief—just a few pages. Coffman said the out-of-pocket methodology was the 'standard and well-accepted method for calculating class-wide damages in cases under Section 10(b) of the Exchange Act.” J.A. 589. Under this approach, each class member’s “damages are equal to the artificial inflation per share at the time of purchase minus the artificial inflation per share at the time of salePage 17 (or if the share is not sold prior to the full revelation of the fraud, just the artificial inflation at the time of purchase).” J.A. 589–90.

Coffman described the out-of-pocket calculation as requiring two types of inputs. One involved each class member’s purchase and sale histories. The other was artificial inflation per share per day, which he planned to calculate via “a detailed loss causation analysis.” J.A. 590. But he said any method for doing so would be common to the class.

Still, Coffman summarized how he might go about determining artificial inflation. He explained he might start with 'an event study that measures price reactions to disclosures that revealed the relevant truth, such as the price reaction to events such as the Alaska Airlines Incident that exposed systemic and pervasive manufacturing safety and quality deficiencies at Boeing.' J.A. 591. This approach, if followed, would require “disaggregating the price impact of corrective disclosures from confounding information”—in other words, excluding any decrease in the stock price not related to the alleged misstatements. J.A. 591. And he provided a non-exhaustive list of potential valuation approaches that he might use to achieve this.

Coffman also noted his analysis would need to show how artificial inflation changed over time. He gave three possible ways to do that. He could use “‘constant dollar inflation,’ which implies that the artificial inflation was the same dollar amount during the class period.” J.A. 592. Or he could use “‘constant percentage inflation,’ which implies the price was inflated by a consistent percentage in the absence of additional disclosures.” J.A. 592. Or a third “approach may be necessary, such as where the artificial inflation has evolved based upon the nature and timing of specific misstatements or the inflation varied on a Page 18daily basis as a result of information contained in internal documents obtained in discovery.” J.A. 592.

Boeing opposed class certification. Memorandum in Opposition, In re The Boeing Co. Sec. Litig. (E.D. Va. Jan. 21, 2025), ECF No. 115. It contended principally that Coffman’s proposed damages methodology did not meet Comcast 's requirements. Id. at 10–24. As Boeing saw it, Coffman had proposed no methodology at all—he had merely identified possible techniques that he might use to calculate damages without committing to an approach. Id. at 11-12. Nor did Coffman provide any detail about how he would account for the complexity of the case—involving numerous allegedly false and misleading statements, of varying kinds, made over the course of years in a changing environment. Id. at 14–19.

Boeing also argued Coffman’s report didn’t demonstrate how his proposed damages methodology was consistent with the plaintiffs' liability theory. Id. at 19–21. According to Boeing, to overcome the materiality problems with their alleged misstatements, the plaintiffs advanced a materiality-by-repetition theory-the materiality of the allegedly false and misleading statements crept in over time and, with it, so did artificial inflation. Id. at 4, 19–21. But Boeing argued that using a constant measure of inflation, like constantdollar inflation or constant-percentage inflation, is inconsistent with a liability theory premised on the evolution of inflation over time. See id. at 19-21. Plus, Boeing pointed out Page 19that Coffman provided no details on a third approach to measuring inflation which might attempt to accommodate that evolution. Id. at 21.¹²

In reply, the plaintiffs argued their “Out-of-Pocket Method is ‘well accepted’ and perfectly fits the theory of liability: the fraud artificially inflated Boeing’s stock price, and the damages methodology measures losses due to that inflation.” Reply at 1, In re The Boeing Co. Sec. Litig. (E.D. Va. Feb. 20, 2025), ECF No. 129 (emphasis in original); see also id. at 3 ('This method is perfectly consistent with Plaintiffs' liability theory: Defendants’ fraud inflated Boeing’s stock price, causing injury to investors as the truth was revealed and inflation was removed.”). They said they didn’t need to provide additional detail about determining inflation because “the only individualized inquiry necessary to calculate damages is the trivial and routine act of processing trading data from Class members.” Id. at 4 (emphasis in original). Accordingly, there was no predominance problem and any questions concerning the feasibility of measuring damages should be deferred for merits consideration down the road. Id. at 5.

With the plaintiffs' reply, Coffman submitted a rebuttal report. This report was longer than his initial report. He maintained his 'opinion that the Out-of-Pocket methodology can be reliably applied on a class-wide basis and consistently with Lead Plaintiffs’ theory of liability.” J.A. 1436–37. And he reiterated the high-level approach he proposed in his initial report. Beyond that, as to disaggregating the impact of any confounding information on Boeing’s share-price declines, Coffman maintained he could Page 20disaggregate such information if necessary but emphasized 'the task of identifying the specific analysis that will be used to parse out confounding information is typically conducted at a later stage of litigation.” J.A. 1440–41. Without committing to either, he offered “two potential approaches” to disaggregation, both of which “could” disaggregate confounding information from inflation dissipation if “conducted reasonably and properly.” J.A. 1441, 1442. Coffman also backed off his suggestion that he might use a constant approach to measuring artificial inflation. He emphasized that “[n]owhere d[id] [his Initial] Report claim that [he] intend[ed] to strictly use a constant dollar or constant percentage approach to back cast artificial inflation in this case.” J.A. 1444. And without saying how, he insisted he could calculate variable inflation if he determined inflation indeed varied during the class period. Finally, Coffman characterized the plaintiffs as having asserted an inflation maintenance, or price maintenance, theory of liability—that each alleged misstatement concealed the same underlying truth and, therefore, maintained Boeing’s stock price at an inflated level.

The district court agreed with the plaintiffs. It found Coffman's 'out-of-pocket methodology for calculating per-share inflation . . . 'widely accepted as the traditional measure of damages for Rule 10b-5 actions.’” J.A. 1695 (quoting In re NII Holdings, Inc. Sec. Litig. , 311 F.R.D. 401, 413–14 (E.D. Va. 2015)). And it determined the methodology 'fit[] plaintiffs' theory of liability: that investors were damaged by purchasing Boeing stock at inflated prices due to defendants’ fraud.” J.A. 1695. The district court also noted “the vast majority of courts, including [in the Eastern District of Virginia], have interpreted Rule 23(b)(3) and Comcast to permit the standard out-of-pocket method for calculating Page 21damages to be used in securities class actions like this one.” J.A. 1695 (citation modified). It further explained that “the caselaw in this Circuit does not require plaintiffs to conduct detailed damages modeling to meet Rule 23's predominance requirement at the class certification stage.” J.A. 1695. The district court then certified the class.

Less than one week after the district court certified the class, and less than a month after his rebuttal report, Coffman submitted a merits report. Coffman's analysis started from the premise 'that Defendants represented that Boeing was prioritizing safety and quality of its aircraft above production speed and short-term cash flows” while in reality “Boeing downplayed[] or failed to address numerous safety and quality concerns and failed to implement critical safety measures.” J.A. 1807–08. This “maintain[ed]” Boeing’s stock price “at artificially inflated levels, relative to the prices one would expect, had Boeing disclosed the truth at the beginning of (or during)” the class period. J.A. 1808.

Coffman identified six events-starting with the January 8, 2024 decline in Boeing's share price based on reports about the Alaska Airlines incident-which “reveal[ed] at least a portion of the relevant truth concealed by alleged misrepresentations and omissions.” J.A. 1808–09. Coffman then conducted an event study, “controlling for market and industry effects,” to determine the events that caused Boeing’s stock price to fall. J.A. 1811. He said Boeing’s price drops were “reasonable measure[s] of the amount of artificial inflation that was dissipated,' J.A. 1811-in other words, there was no confounding information that contributed to the decline in the share price—because he “carefully evaluated whether there was any other news outside the realization by market Page 22participants that there were broader manufacturing safety issues at Boeing that would draw regulatory scrutiny and hinder the ability of Boeing to maintain and grow its production of commercial aircraft and found none,” J.A. 1854.

Then, employing a constant-percentage-inflation approach, he calculated this purported artificial inflation. Coffman found constant-percentage inflation applicable because Boeing’s alleged misrepresentations “fundamentally concealed the same underlying truth” while Boeing’s stock price changed over time. J.A. 1894. According to Coffman, each of the numerous alleged misstatements over the three-year class period had the exact same effect—they maintained a price for Boeing stock that was at least 9.58% too high.¹³

Meanwhile, Boeing timely petitioned us for review of the district court's classcertification order under Rule 23(f). We granted the petition.¹⁴

Page 23In challenging the district court's certification decision, Boeing argues Comcast required the plaintiffs to put forth a class-wide, case-specific damages methodology consistent with their theory of liability. And Boeing says Coffman’s initial and rebuttal reports failed to do this. The plaintiffs disagree. They contend Coffman’s reports satisfy Comcast ’s requirements. To tee up our discussion of the parties’ positions, we’ll begin with Comcast itself, summarizing the decision and explaining what instructions it gives to plaintiffs and district courts.

Comcast applied Rule 23(b)(3) to a class of Comcast subscribers who alleged the company violated antitrust laws. 569 U.S. at 29. Comcast, a telecommunications provider, acquired a competitor’s infrastructure in the Philadelphia area in exchange for Comcast’s existing systems in Palm Beach and Los Angeles. Id. at 29-30. As a result, Comcast's Philadelphia market share increased from approximately 25% to 70%. Id. at 30. The plaintiffs alleged that both Comcast's transactions and its monopoly or attempted monopoly around Philadelphia violated the Sherman Act. Id. The plaintiffs eventually sought class certification. Id. In their certification motion, the plaintiffs proposed four theories of antitrust impact, also deemed theories of liability. See id. at 31, 37. The district court allowed only one of the four theories to proceed. Id. at 31. Nevertheless, the district court certified the class, even though the plaintiffs’ expert’s damages methodology “did not isolate damages resulting from any one theory of antitrust impact.' Id. at 32.

Page 24Comcast appealed. Id. It argued the district court improperly certified the class because the methodology “failed to attribute damages resulting from [] the only theory of injury remaining in the case.” Id. But the Third Circuit rejected this argument. Id. In its view, Comcast was focused too early on the merits; the plaintiffs did not yet have to tie their viable theory to a precise damages calculation. Id. The Supreme Court granted certiorari to consider whether “certification was improper because [the plaintiffs] had failed to establish that damages could be measured on a classwide basis.” Id. at 32 n.4.

The Supreme Court started by describing the level of analysis required by district courts in considering class-certification motions. Id. at 33–34. It noted Rule 23 “does not set forth a mere pleading standard.' Id. at 33 (quoting Wal-Mart Stores, Inc. v. Dukes , 564 U.S. 338, 350 (2011)). Rather, plaintiffs must satisfy at least one provision of Rule 23(b) with “evidentiary proof.” Id.

The Court also emphasized that, in addressing class certification, the district court must undertake a “rigorous analysis,” id. (quoting Dukes , 564 U.S. at 351), also described as a “close look,” id. at 34 (quoting Amchem , 521 U.S. at 615). This review might often overlap with the merits of the plaintiffs’ claims. Id. at 33-34. As a result, courts cannot discount 'arguments against [a] damages model that b[ear] on the propriety of class certification[] simply because those arguments would also be pertinent to the merits discussion.” Id. at 34.

Then, the Court explained the requirements for damages methodologies in class actions. A party seeking certification must put forward a methodology that demonstrates damages are measurable on a class-wide basis and are “consistent with its liability case, Page 25particularly with respect to the alleged anticompetitive effect of the violation.” Id. at 35 (citation modified). And while plaintiffs need not present exact measurements, their methodology must allow 'a just and reasonable inference' of damages. Id. (citation omitted). It cannot be 'speculative.' Id. (citation omitted). Otherwise, 'at the classcertification stage any method of measurement [would be] acceptable so long as it can be applied classwide, no matter how arbitrary the measurements may be.' Id. at 36 (emphasis in original). “Such a proposition would reduce Rule 23(b)(3)’s predominance requirement to a nullity.” Id.

Applying these principles, the Court found the plaintiffs' damages methodology flawed. Id. at 36–38. The plaintiffs’ methodology “assumed the validity of all four theories of antitrust impact.” Id. at 36. It did not “attribute damages” solely to the plaintiffs’ only viable theory. See id. at 37. Accordingly, it improperly “identifie[d] damages that [we]re not the result of the wrong.” Id. 15

Page 26Putting this all together, Comcast provides guidance for plaintiffs and district courts.

Plaintiffs must meet several requirements. First, they must put forth a damages “methodology.” Comcast , 569 U.S. at 34. Methodologies require more than a legal description of what damages are generically-they must explain how damages will be measured in a specific case. 'If the model does not even attempt to do that, it cannot possibly establish that damages are susceptible of measurement across the entire class for purposes of Rule 23(b)(3).” Id. at 35. A menu of options that the party will decide on later is not a damages methodology. It is, therefore, insufficient for plaintiffs to give examples of what they might, or could, do to measure damages—they have to tell the district court what their actual methodology is and how it resolves the predominance inquiry.¹⁶ Cf.

Page 27Spurlock v. Wexford Health Sources, Inc. , 175 F.4th 232, 249 (4th Cir. 2026) (finding damages methodology satisfied Comcast when 'expert report model[ed] economic damages based on the number of days a class member was denied adequate medical treatment”).

Second, the damages methodology must demonstrate how damages can be measured on a class-wide basis. Comcast , 569 U.S. at 34. This doesn't mean the methodology can't include any individualized inputs like data points concerning when class members bought and sold stock at corresponding price points. Cf. Tyson Foods, Inc. v. Bouaphakeo , 577 U.S. 442, 453 (2016) (recognizing “[w]hen ‘one or more of the central issues in the action are common to the class and can be said to predominate, the action may be considered proper under Rule 23(b)(3) even though other important matters will have to be tried separately, such as damages or some affirmative defenses peculiar to some Page 28individual class members'' (quoting 7AA Wright & Miller's Federal Practice and Procedure § 1778 (3d ed. 2005))). But the methodology must explain how damages will be measured so that the district court can determine if class-wide issues predominate over individual ones.

Third, “any model supporting a plaintiff’s damages case must be consistent with its liability case.' Comcast , 569 U.S. at 35 (citation modified). Demonstrating consistency requires a comparison. So, plaintiffs must first identify both comparators, stating their theory or theories of liability to compare against their identified damages methodology.¹⁷ Then, they must show that the methodology and the liability theory are sufficiently consistent. A damages methodology is inconsistent with a theory of liability, for example, if it measures damages attributable to theories of liability that are no longer part of a case. Id. at 37.

Fourth, the plaintiffs' damages methodology must allow 'a just and reasonable inference” of damages. See id. at 35 (citation omitted). It can’t be “speculative.” See id. (citation omitted). Though “[c]alculations need not be exact,” id. , plaintiffs can’t propose “ any method of measurement . . . so long as it can be applied classwide, no matter how Page 29arbitrary the measurements may be,' id. at 36 (emphasis in original). Otherwise, Rule 23(b)(3)'s predominance requirement becomes 'a nullity.' Id.

Fifth, plaintiffs must put forth “evidentiary proof” that their damages methodology satisfies all of the above requirements. Id. at 33. Rule 23 “does not set forth a mere pleading standard.' Id. (quoting Dukes , 564 U.S. at 350). It requires evidence. And plaintiffs are not relieved of these obligations if certification issues “overlap with the merits of the plaintiff’s underlying claim,” Id. at 33-34 (quoting Dukes , 564 U.S. at 351).

As for district courts, Comcast requires that they 'conduct a rigorous analysis to determine whether' the plaintiffs' damages methodology satisfies Rule 23 in the ways described above. See 569 U.S. at 35 (citation modified). As the Court explained, district courts must “take a close look at whether common questions predominate over individual ones.” Id. at 34 (citation modified).

Neither the plaintiffs’ class-certification materials nor the district court’s certification order meet the standards set by Comcast . To show why, we’ll start with the plaintiffs’ motion for class certification and Coffman’s initial report. Then, we turn to the plaintiffs' reply brief and Coffman's rebuttal report. We conclude by addressing the plaintiffs’ counterarguments and whether the district court properly conducted a rigorous analysis before certifying a class.

Page 30In their motion for class certification, the plaintiffs pointed to “the ‘out-of-pocket’ methodology' provided by Coffman. Memorandum in Support at 27, In re The Boeing Co. , ECF No. 104. Coffman, in turn, defined the out-of-pocket approach—“damages are equal to the artificial inflation per share at the time of purchase minus the artificial inflation per share at the time of sale (or if the share is not sold prior to the full revelation of the fraud, just the artificial inflation at the time of purchase).” J.A. 589–90.

That is not enough. Rather than offering a methodology, the plaintiffs provided only a legal description of damages. For a securities fraud case, that’s effectively saying that they are seeking compensatory damages. The out-of-pocket description broadly tells us what securities-fraud damages are—the difference between the price paid or received and the price that would have existed absent the fraud. But it doesn’t tell us how to determine the artificial inflation embedded in Boeing’s stock price on any day of the class period. Comcast demands that missing step.

True, Coffman explained that damages calculations will involve two inputs, (1) individualized data about the dates and prices of purchases and sales and (2) artificial inflation per share per day. But that description is so general that it would apply in any securities fraud case. And beyond that, Coffman never committed to any methodology. He said he could perform an event study to measure price reaction. He remained noncommittal on if he would do that and, if so, how he would do that. He seemed to acknowledge that he would need to disaggregate confounding information—which, again, means culling out declines in share price not caused by the alleged misrepresentations. But he didn’t say how Page 31he would do that. Instead, he provided a non-exhaustive list of valuation techniques. He didn’t choose one for this case or explain how any would work. Finally, Coffman gave three possible mechanisms for measuring artificial inflation-constant-dollar inflation, constant-percentage inflation and an unidentified third way. Again, he didn’t select from these approaches.

Rather than providing a methodology, Coffman offers a series of 'maybe[s],' “perhaps[es]” and “what ifs.” Speerly v. Gen. Motors, LLC , 143 F.4th 306, 324 (6th Cir. 2025) (en banc). That falls short of what Comcast requires. Comcast doesn’t require a party seeking class certification to actually calculate damages. But it requires more than just stating the legal description of damages-plaintiffs must put forth a methodology that explains how the plaintiff will measure damages in line with that legal description. Without a methodology, the district court has no way to rigorously evaluate whether damages are measurable on a class-wide basis, whether the damages methodology is consistent with the plaintiff’s liability theory or whether the proposed methodology for measuring damages is reasonable.

In addition to the lack of a damages methodology, the plaintiffs didn’t identify their theory of liability by the class-certification stage. True, they asserted Boeing’s allegedly false and misleading statements caused its stock to trade at artificially high levels, which caused a loss when the price fell after the truth emerged. Memorandum in Support at 15, In re The Boeing Co. , ECF No. 104. But that language would describe virtually all securities-fraud cases. The plaintiffs never committed to, let alone explained, how that happened. Such a generic description of securities fraud is not enough to satisfy Comcast .

Page 32Cf. Stafford , 123 F.4th at 675 ('We hold that the district court abused its discretion in certifying this class action because it employed an inappropriately high level of generality when (1) identifying the policies which allegedly unify prospective class members’ wideranging claims, and (2) creating overly broad class definitions.”).

In their amended complaint, the plaintiffs alleged generally that the 75 purported misrepresentations inflated the share price for Boeing stock or maintained it at an artificially high level. In other words, they offered alternative theories of liabilityinflation or maintenance. At the motion-to-dismiss stage, they suggested that even if no one statement caused the price to be inflated, or caused the existing inflated price to be maintained, collectively they did. This introduced alternatives to their two alternatives. By the time of Coffman’s rebuttal report, at least Coffman seemed to be settling on a theory that the alleged misrepresentations maintained artificially high prices. And in their briefs to us, that seems to be their theory. But among their alternative theories, the plaintiffs never committed before the district court.

We understand that litigants sometimes don’t want to tip their hand. But at the classcertification stage, the plaintiffs had to. That's because the district court needs that information to compare the plaintiffs’ damages methodology to their liability theory. To state the obvious, to conduct a comparison, you need to provide the necessary information about both comparators. Here, the plaintiffs provided neither.

Coffman’s rebuttal report provided a longer damages discussion than the few pages offered in his initial report. But it still didn’t provide the information Comcast requires.¹⁸ In that report, Coffman noted that assessing artificial inflation “typical[ly]” involves two steps. J.A. 1438. Step one is “assess[ing] the decline in artificial inflation corresponding to a corrective event,” which “typically” requires an event study “analyz[ing] Boeing’s stock price in relation to market and industry indices.” J.A. 1438. It also includes consideration of confounding events and whether “artificial inflation le[ft] the stock price, given Lead Plaintiffs’ theory of liability.” J.A. 1438.

At step two, Coffman reiterated that he could employ one of several mechanisms for measuring artificial inflation. As before, these were constant-dollar inflation, constantpercentage inflation or “another approach[,] when there is reason to believe that the amount of artificial inflation varied over time.' J.A. 1439. But still refusing to commit to any specific methodology at either step, he remarked that 'assessing the correct approach requires a case-specific loss causation analysis.” J.A. 1439.

Coffman then responded to two criticisms of his approach lodged by Boeing—that he had not measured the impact of information about other contributing factors on the share-price decline or artificial inflation itself. Starting with confounding information, Coffman “agree[d] generally with the proposition that disaggregation of confounding non- Page 34corrective information is required to reliably calculate artificial inflation.” J.A. 1440. But he disagreed that he needed to do so at the class-certification stage because he believed any problems would present class-wide. Nevertheless, “[p]urely by illustration,” he offered two ways in which he “could” tease out the effect of confounding information. J.A. 1441. As for measuring artificial inflation, Coffman emphasized that “[n]owhere” did he indicate he “intend[ed] to strictly use a constant dollar or constant percentage” inflation measurement. J.A. 1444. He pointed out that, if “true economic artificial inflation was not constant over the Class Period,' then the out-of-pocket measurement could still perform its job. J.A. 1445.

Like his initial report, Coffman's rebuttal report fails to describe his proposed methodology in sufficient detail. His substantive discussion merely described examples of what he might or could use—a discussion that would largely apply to every securities-fraud action. He never identified the actual approach that he would use or how he specifically planned to calculate damages based on that approach. A list of possible methodologies, described in general terms, isn’t a methodology, even if it is carried out in more pages.

The plaintiffs make two counterarguments. First, they argue that if any potential problems with their damages theories affect the entire class, they do not have to provide the methodology Comcast demands. That puts the cart before the horse. Comcast requires plaintiffs to put forth an actual methodology so that district courts can readily determine if the specific methodology is a class-wide methodology, is consistent with the liability theory and is not speculative. Plaintiffs can’t get around Comcast ’s frontloading mandate Page 35by emphasizing a backend conclusion. Cf. In re Marriott Int’l, Inc. , 78 F.4th 677, 687 (4th Cir. 2023) (noting Supreme Court requires rigorous analysis before certification). Without a sufficient methodology, district courts have no way of knowing whether damages calculations might overwhelmingly involve individual questions and inputs, even if it seems likely they won’t.¹⁹

A Fifth Circuit decision involving this same expert witness illustrates the importance of proof prior to certification. Ludlow v. BP, P.L.C. , 800 F.3d 674 (5th Cir. 2015), involved an explosion which caused the BP oil spill in the Gulf of Mexico. Id. at 677. Before the district court, the plaintiffs ultimately “argue[d] that Defendants’ process safety misstatements were a proximate cause of all their post-explosion investment losses because Plaintiffs were deprived of the opportunity to avoid the increased risk by divesting prior to the explosion.' In re BP P.L.C. Sec. Litig. , No. 4:10–md–2185, 2014 WL 2112823, at *11 (S.D. Tex. May 20, 2014). As a result, the plaintiffs sought damages for not only “the inflated price [they] paid” but also “the materialization of a risk [they] may have been willing to take.” See Ludlow , 800 F.3d at 690. In affirming the district court’s refusal to certify this class, the Fifth Circuit explained why this theory created predominance Page 36problems—“where the economic loss depends on the posture of the plaintiff vis-à-vis risk tolerance, that loss causation, and thus damage, cannot be presumed nor can it be found class-wide.” Id. at 691. Plaintiffs could not 'be compensated for the materialization of a risk [they] may have been willing to take,” absent “any mechanism separating [relevant] classes of plaintiffs.' Id. at 690.

Here, perhaps trying to avoid that same fate, the plaintiffs have not committed to an approach for disaggregating damages. Without doing so, it’s impossible to know whether their methodology might likewise cause individualized damages questions to predominate over common questions. That’s the trouble with putting the cart before the horse—it leaves us, and district courts, to speculate as to how the case will eventually turn out. Comcast requires more certainty before the court allows a class action to proceed.

Second, the plaintiffs contend our reading of Comcast conflicts with that in other circuits, creating a circuit split. Their best case is the Second Circuit's decision in Waggoner v. Barclays PLC , 875 F.3d 79 (2d Cir. 2017). It may be true that Waggoner interprets Comcast more narrowly than we do. But even so, it is distinguishable from this case. The plaintiffs in Waggoner pressed a single, committed theory-that specific misstatements about Barclays’ alternate trading system maintained a preexisting inflation that dissipated when one corrective disclosure revealed the truth. Id. at 88. The fit between the misstatements and the disclosure was tight, the theory was fixed and the court found that the expert’s constant-dollar model matched it. See id. at 88, 90, 105-06.

This case differs on every axis that mattered there. Here, the plaintiffs pled inflation and maintenance in the alternative and committed to neither before the district court. And Page 37Coffman committed to no methodology at all. Where Waggoner had a fixed theory and a matching methodology to test, the district court here had a moving target and a menu. Comcast requires something to compare. Waggoner had it, and this record does not.

In sum, neither of the plaintiffs’ counterarguments are persuasive. They failed to satisfy Comcast 's requirements.

Because the plaintiffs did not meet their burden under Comcast , they necessarily prevented the district court from conducting the required rigorous analysis of their damages methodology. Nevertheless, by approving the plaintiffs' insufficient proof, the district court misapplied Rule 23(b)(3).

The district court found that the out-of-pocket legal measure of damages qualified as a damages methodology under Comcast . As we have explained, it doesn't. And by accepting an inadequate damages methodology, the district court was unable to properly carry out its obligations under Rule 23(b)(3).

The district court also characterized the plaintiffs' theory of liability at 'an inappropriately high level of generality.” See Stafford , 123 F.4th at 675. According to the district court, the plaintiffs’ theory was “that investors were damaged by purchasing Boeing stock at inflated prices due to defendants’ fraud.” J.A. 1695. But that isn’t a liability theory. It simply restates the generic securities-fraud principles that misstatements inflate and the truth deflates. That describes every Section 10(b) case, which is precisely the problem. Comcast 's consistency inquiry compares two things, the damages methodology and the liability theory. A comparison needs two defined terms. A generic description that fits Page 38every case defined nothing for this one, so it left the district court with nothing to test the methodology against.

Having failed to properly identify the two comparators—the damages methodology and the legal liability theory—the district court’s certification order necessarily failed to perform a rigorous consistency comparison. Of course, as we have explained, the plaintiff did not give the district court the information needed to do this. The plaintiffs failed to provide an adequate methodology and failed to identify their liability theory beyond the most general of levels. But Comcast requires the district court to conduct a rigorous analysis that includes this comparison. If the party seeking class certification doesn't provide the necessary information about both comparators, the district court cannot certify the class.

Also, the plaintiffs’ failure to provide the required damages methodology prevented the district court from addressing Comcast 's other requirements. Without such a methodology, it could not, and thus did not, properly consider whether the plaintiffs' alleged damages were measurable on a class-wide basis or whether the plaintiffs’ purported methodology was reasonable as opposed to speculative.

The district court explained that the Fourth Circuit “does not require plaintiffs to conduct detailed damages modeling to meet Rule 23’s predominance requirement.” J.A. 1695. It is true that “[c]alculations need not be exact.” Comcast , 569 U.S. at 35. But there must be enough information in the plaintiffs’ filings for the district court to identify the plaintiffs’ methodology, specify their theory or theories of liability and explain whether the Page 39former is consistent with the latter and whether the methodology is reasonable. As already explained, the plaintiffs failed to provide such information here.

The district court also noted that its approach was similar to those of several other district courts within our circuit and was consistent with Waggoner . That may be true. But to the extent there was any question previously, we make clear today that Comcast requires a rigorous analysis into Rule 23’s requirements. Under Comcast , certification orders are not like participation trophies that are handed out to everyone on the tee ball team. To the contrary, district courts may only certify a class after rigorously analyzing whether the plaintiffs have complied with Comcast ’s commands.

Finally, the plaintiffs suggest Coffman's merits report-submitted after class certification—cures any Comcast deficiencies. To the extent the plaintiffs are asserting a harmless-error argument, there are three problems.

First, when asked at oral argument, the plaintiffs indicated they hadn't made a harmless-error argument. See Oral Argument at 38:25-:52 ('I think we haven't made the argument that you’re describing . . . .”). Second, we are suspicious that the harmless-error principle applies here, as “[t]he Supreme Court has emphasized the ‘rigorous analysis’ [] must be performed before a class is certified under Rule 23.' Marriott , 78 F.4th at 687 (emphasis in original) (quoting Dukes , 564 U.S. at 351). And it seems equally inappropriate to let a class-certification decision stand based on a report which the district court didn’t even consider in certifying a class in the first place. That’s particularly true here when the report was completed and filed such a short period after the rebuttal report and the class- Page 40certification decision. Concerningly, the plaintiffs' damages reports go from minimal discussion and no commitment in the initial report, to lengthier discussion but still no commitment in their rebuttal report, to, less than a week after the district court certified the class, a merits report that—with the time for scrutiny having passed—finally makes some commitments. If allowed, a plaintiff could bob and weave pre-certification and not commit to a firm methodology until he has the leverage of class certification to use as a sword in litigation.

Third, even if the plaintiffs had made a harmless-error argument and we could address it in this context, the district court’s error in certifying the class wasn’t harmless. It's true that 'an error by the district court does not automatically result in reversal.' Figueroa v. Butterball, LLC , 164 F.4th 312, 319 (4th Cir. 2026). And it’s also true that Rule 61 requires courts to “disregard all errors and defects” in an order “that do not affect any party's substantial rights.' Fed. R. Civ. P. 61. But we must be 'satisfied with fair assurance, after pondering all that happened without stripping the erroneous action from the whole, that the judgment was not substantially swayed by the errors.” Design Gaps, Inc. v. Distinctive Design & Constr. LLC , 162 F.4th 452, 473 (4th Cir. 2025) (internal quotation marks omitted) (quoting Shears v. Ethicon, Inc. , 109 F.4th 235, 241 (4th Cir. 2024)).

Here, we aren’t satisfied that the incorrect certification decision wasn’t swayed by the errors. That’s because even Coffman’s merits report isn’t enough for the plaintiffs to achieve class certification. In that report, Coffman finally committed to his methodology of determining artificial inflation. He opined that a constant-percentage-inflation measure Page 41of artificial inflation was appropriate and assessed the percentage to 9.58%, considering only the Alaska Airlines incident. According to Coffman, '[e]ven though there are numerous alleged misstatements, . . . they all fundamentally concealed the same underlying truth that Boeing was taking dangerous shortcuts and prioritizing short term cash flow . . . over correcting safety and quality control deficiencies and maintaining appropriate legal compliance.” J.A. 1894.

That, however, is inconsistent with the plaintiffs’ theory of liability. Recall that the plaintiffs alleged 40 different misstatements over a three-year period. The misstatements suppressed a variety of different types of information. For instance, some concerned safety and quality in manufacturing. Others were about Boeing's culture and encouraged its employees to report safety concerns. Others, still, concerned the production of safe airplanes at a stable and growing rate. And finally, some were about regulatory compliance.

It isn’t consistent with these allegations to assert a damages methodology based on the conclusion that these varying statements—of different kinds, at different points in time, with different industry and macroeconomic conditions and with evolving investors' perceptions of the riskiness of a Boeing investment—would maintain the alleged fraud by an exact percentage that never changes. Cf. Stafford , 123 F.4th at 680 (noting generalized allegations are often insufficient to achieve class certification). Conclusorily claiming that each statement had the exact same effect on the market is speculation, not a showing of consistency.

The plaintiffs may seek to base their claim on the large number of statements over a three-year period. But having done so, they are stuck with those allegations when Page 42applying Rule 23(b)(3). As the saying goes, they made their bed and now have to lie in it. Lying in it reveals that they cannot show their liability theory is consistent with the damages methodology Coffman provides in his merits report.

To strike the appropriate balance between certifying a lawsuit as a class action and restraining it to individual parties when certification is not warranted, plaintiffs must comply with the requirements of Rule 23, as interpreted by the Supreme Court in decisions like Comcast . But the damages methodology the plaintiffs attempted to offer here fell short of those requirements. And the district court didn’t perform the rigorous analysis Comcast requires. So, the district court's class-certification order is reversed, and the case is remanded for further proceedings consistent with this opinion.

REVERSED AND REMANDED