Certification and the Trial Plan

Two vehicles, one underlying question

A meal-and-rest case is ordinarily pleaded as both a class action for the unpaid premiums and a PAGA action for civil penalties, and the two are governed by different procedures. The class action proceeds under Code of Civil Procedure section 382 and must be certified before it can be tried on a class basis. The PAGA action need not be certified at all — the California Supreme Court has held that a representative action under PAGA is not a class action and is not subject to class-certification requirements. Yet both vehicles converge on the same problem: whether the employer's liability to a group of employees can be established without depriving the employer of its right to contest liability as to each of them. Select a vehicle to compare how each handles that problem.

Fig. 1. The two vehicles compared. Certification and the motion to strike for unmanageability exist on the class side and not the PAGA side — but the due-process constraint on representative proof governs both.

The certification burden and the standard of review

Certification is the plaintiff's burden, carried by a preponderance of the evidence, and the trial court does not take the pleading's word that common questions predominate. It conducts a rigorous analysis of the actual showing, examining the evidence the parties submit and the way the case would in fact be tried. The plaintiff must establish each element of the community of interest, not merely assert it; a class is certified because the record demonstrates that the elements are met, and denied because it does not.

Fig. 2. The elements of the community of interest. In a meal-and-rest case the contest is almost always predominance; numerosity, typicality, and adequacy are usually conceded or readily met.

The standard of review on appeal shapes how a defendant frames its opposition. A certification ruling is reviewed for abuse of discretion, and a reviewing court will not disturb an order supported by substantial evidence and resting on correct legal criteria. That deference has a limit that is itself a defense strategy: where the trial court relied on improper criteria or an erroneous legal assumption, the order is reversible without deference — which is the ground on which the Court of Appeal in Brinker first reversed, and the reason a defendant litigates certification around the governing legal standard for the claim rather than only the weight of the competing declarations. Whether the court applied the correct meal-and-rest rules — provide rather than ensure, the timing requirements, the major-fraction rest standard — is reviewable as a question of law, and an order certifying on a mistaken view of those rules does not survive.

Class definition and ascertainability

Before predominance is reached, the class must be defined, and in California the ascertainability requirement is — after Noel v. Thrifty Payless, Inc. (2019) 7 Cal.5th 955 — a modest threshold rather than a merits gate. A class is ascertainable if it is defined by objective characteristics and common transactional facts, so that membership can be determined by reference to those criteria. Noel rejected the stricter view that would have required the plaintiff to prove, at certification, a means of identifying and notifying individual members; concerns about locating and notifying members, the Court held, belong to the manageability and superiority inquiries, not to ascertainability. For a workforce-wide meal-and-rest class the requirement is rarely a serious obstacle, because the class is the employer's non-exempt employees over the period, identifiable from the employer's own payroll and time records.

Two definitional defenses survive Noel and are worth pressing notwithstanding its permissive thrust. The first is the fail-safe class — one defined by the elements of the claim itself, so that membership turns on whether the person has a valid claim, as in a proposed class of employees “who were denied compliant meal periods.” Such a definition is improper, because it would bind only those who ultimately prevail and leave everyone else free to relitigate; the answer is to insist that the class be defined by objective job and time criteria rather than by the legal conclusion in dispute. The second is overbreadth — a class that sweeps in employees who could have no claim, such as those governed by a different policy, covered by a valid waiver, or exempt under a collective bargaining agreement — which Noel and Brinker permit the court to address by narrowing the definition. Beyond those, a class-definition attack is better routed through predominance, manageability, and superiority than through ascertainability, which Noel deliberately made difficult to wield as a standalone bar.

Typicality, adequacy, and the vulnerable representative

The community of interest also requires representatives whose claims are typical of the class and who will adequately protect it, and these elements give a defendant an avenue that does not depend on winning the predominance contest. Typicality fails where the named plaintiff is subject to a defense, or possesses facts, peculiar to that plaintiff and not shared by the class: a representative whose own claim would be consumed by an individual dispute is advancing a personal claim rather than the class's. A plaintiff who signed a valid meal-period waiver, who was covered by a collective bargaining agreement's meal exemption, or whose missed periods are documented as voluntary is atypical of a class whose theory is that compliant periods were never provided — and the defense develops those plaintiff-specific facts precisely to sever the representative from the class he proposes to lead.

Adequacy requires that the representative and class counsel be free of disqualifying conflicts and able to prosecute the action vigorously, a requirement a defendant invokes where the named plaintiff's interests diverge from the class's or where the plaintiff's limited knowledge of the case undermines the representation. The attack on the representative carries a second edge in the current landscape: where an arbitration program reaches the plaintiff, compelling the individual claim tests the plaintiff's status directly, and a determination that the plaintiff suffered no violation removes the representative from the class on the class side just as it defeats standing on the PAGA side — the sequence developed in Arbitration, Standing & the Headless-PAGA Split. Typicality and adequacy are therefore not afterthoughts to predominance; they are an independent route to denying or narrowing certification by contesting the person the plaintiff has chosen to carry the class.

Community of interest: the predominance inquiry

The decisive element is predominance, and Brinker Restaurant Corp. v. Superior Court (2012) 53 Cal.4th 1004 fixed how a court approaches it. Certification asks whether the plaintiff's theory of liability is, by its nature, susceptible to common proof — not whether the theory will ultimately prevail. The merits are examined only to the extent they are enmeshed with the question whether common or individual issues predominate; a court does not resolve the merits at certification, but it must understand them well enough to determine how the case will be tried. A court evaluating predominance must determine whether the elements necessary to establish liability are susceptible to common proof, or, if not, whether there are ways to manage effectively the proof of any elements that require individualized evidence.

Ayala v. Antelope Valley Newspapers, Inc. (2014) 59 Cal.4th 522 sharpened the focus in a way that governs the meal-and-rest analysis. The certification question, Ayala held, is one step removed from the merits: it is not how the employer's policy or practice operated as to each employee, but whether the employer's policy or right is sufficiently uniform to permit a classwide answer — whether there is a common way to show the employer treated the class the same. Variation in how a uniform policy played out for individual employees does not defeat certification if the policy itself is common; what defeats it is the absence of any common policy or practice to adjudicate. The certification fight in a meal-and-rest case is therefore a fight over characterization: whether the record discloses a uniform policy or practice whose lawfulness is one common question, or only a set of individual circumstances with no common thread.

Brinker's three subclasses: certification is theory-dependent

Brinker is not only the source of the governing merits rule; it is a worked demonstration that certification turns on the theory rather than the claim type. The same plaintiffs sought to certify three subclasses arising from the same workforce, and the Court reached three different results — affirming one, remanding one, and reversing one — for reasons that map precisely onto whether a common policy or a common means of proof existed.

Fig. 3. Brinker's differentiated certification. Rest periods certified on a uniform unlawful policy; meal periods remanded because certification rested on a wrong legal standard; off-the-clock decertified for want of common proof. The lesson is that the certifiability of a meal-and-rest claim is decided by its theory, not its label.

The practical instruction for a defendant is to disaggregate. A plaintiff who pleads “meal-and-rest violations” as an undifferentiated whole obscures that the rest theory may rest on a uniform written policy while the meal theory depends on individualized provision and the off-the-clock theory has no common proof at all. Separating the theories at the certification stage allows each to be tested on its own footing — and frequently leaves the plaintiff with one certifiable theory rather than a certified case sweeping in claims that cannot be commonly proven.

Policy versus practice: the three forms of the common theory

The uniform-theory showing takes three forms, and they are not equally easy to certify or to defeat. The first is a written policy unlawful on its face — a handbook that grants a twenty-minute meal, or authorizes rest only once per full four hours as in Brinker. Its illegality is a pure question of law common to everyone it governs, and it certifies readily; the defense is rarely about predominance and almost always about whether the written policy was actually in force or had been superseded. The second is a facially lawful policy alleged to be unlawful as applied — a compliant handbook undercut by a contrary practice — where the plaintiff must prove the gap between the policy and the conduct, and the defendant answers that the lawful policy was in fact followed.

The third, and the one most contested in food service, is a uniform unlawful practice with no written analog: chronic understaffing during service rushes, an automatic meal deduction applied without confirming a duty-free period, or documented pressure to forgo breaks. Here the plaintiff must establish that a practice existed and was common — not merely that some employees missed some breaks — and the defendant's response is to show that no uniform practice existed: that the written policy was lawful and followed, that managers ran breaks differently across locations and shifts, and that the reasons any given employee missed a break varied. Whether the plaintiff has proven a common practice, as opposed to assembling anecdotes that vary employee by employee, is frequently the entire certification contest, and it is won on the texture of the operational record rather than on the abstract availability of class treatment. The auto-deduction and recording practices that most often supply — or defeat — the practice theory are treated in Records as Presumptive Proof.

The merits rule that drives predominance

What makes meal-and-rest certification distinctive is that the governing merits rule cuts toward individualization. Brinker held that an employer's obligation is to provide a compliant meal period — to relieve the employee of duty, relinquish control, and permit an uninterrupted thirty minutes — but not to ensure that the period is taken or that no work is performed. Because the employer need only provide and not police, the defendant's argument at certification is that whether any given employee took the provided break, and why, is an individual question: one employee skipped lunch to leave early, another to earn a premium, a third was genuinely impeded. If liability requires that employee-by-employee inquiry, common questions do not predominate and the class should not be certified.

The plaintiff answers on two fronts, and both convert individual questions back into common ones. The first is the uniform unlawful policy or practice just described: where one exists, its lawfulness is a single common question that predominates regardless of individual variation. The second is the records-based presumption. Justice Werdegar's concurrence in Brinker observed that when an employer's records show no meal period was taken, a rebuttable presumption arises that none was provided — a presumption the Supreme Court later adopted as the holding of Donohue v. AMN Services (2021) 11 Cal.5th 58. The presumption makes the employer's own time records common proof of violation, shifting to the employer the burden of coming forward with evidence that compliant periods were in fact provided.

Fig. 4. The predominance battleground. The waiver and provision questions underlying it are developed in Meal-Period Waivers and “Provide” versus “Ensure.”

One distinction disciplines both sides of that contest and is frequently decisive. Individualized damages do not defeat certification: that class members are owed different premium amounts, or that some are owed nothing, is a calculation problem, not a predominance problem, and California courts have repeatedly refused to deny certification on that ground alone. Individualized liability is another matter: where the fact of a violation — whether a compliant period was provided to a given employee — cannot be established without an individual inquiry, predominance fails. The defendant's task at certification is therefore to characterize the contested issue as one of liability rather than damages, showing that the provision question itself varies employee by employee; the plaintiff's task is the reverse, locating liability in a common policy and confining the variation to the size of each recovery.

Brinker applied: the meal-and-rest certification cases

The liability–damages distinction is not abstract; it is the holding of the post-Brinker decisions that actually govern meal-and-rest certification. In Faulkinbury v. Boyd & Associates, Inc. (2013) 216 Cal.App.4th 220, a security-guard employer required all guards to sign on-duty meal-period agreements and maintained no compliant rest policy. Reconsidering its own earlier decision in light of Brinker, the Court of Appeal reversed the denial of certification, holding that the lawfulness of the uniform on-duty-meal policy and of the absence of a rest policy could be determined on a classwide basis. The court expressly corrected the error it had made before Brinker: it had conflated damages with liability, reasoning that the employer would be liable only on an individualized showing that each guard actually missed a break. Brinker made clear that liability turns on whether the policy is lawful — a common question — while whether a given employee was able to take the break “goes to damages,” and “the fact that individual employees may have different damages does not require denial of the class certification motion.”

Bradley v. Networkers International, LLC (2012) 211 Cal.App.4th 1129 reached the same result on the same reasoning, holding that the legality of an employer's lack of a meal-and-rest policy could be resolved classwide even though which employees missed breaks, and how many, varied — that variation being a matter of damages, not predominance. Together with Lubin, these decisions fix the operative rule for the food-service defendant: a uniform policy, or the uniform absence of one, presents a common question that supports certification, and the individualized character of who missed which break, and what each is owed, will not by itself defeat it. The defense that prevails is not the observation that damages vary, but the showing that liability itself — the provision question — cannot be answered without an individual inquiry, because no uniform policy or practice controls it.

One point of divergence between California and federal law is worth marking, because a defendant will occasionally invoke the federal rule. Comcast Corp. v. Behrend (2013) 569 U.S. 27 held, in the federal Rule 23 setting, that a damages model must measure damages attributable to the plaintiff's liability theory on a classwide basis, and that certification can fail where damages are not susceptible to such a common model. California has not adopted that requirement; under Sav-On, Brinker, Faulkinbury, and Noel, the need to prove damages individually does not defeat certification so long as liability is common. A meal-and-rest defendant litigating in state court should not expect Comcast to carry the damages-variation point, and should locate its predominance argument in the liability inquiry, where California law makes the contest live.

Representative proof and the Duran limit

Once a group is before the court — whether a certified class or the aggrieved-employee group in a PAGA action — the plaintiff must still prove violations across it, and the temptation is to do so by sampling and extrapolation. Duran v. U.S. Bank National Assn. (2014) 59 Cal.4th 1 is the controlling limit, and it is the defendant's principal instrument. The trial court there had certified a class of 260, heard testimony from a sample of roughly twenty, barred the employer from introducing evidence about anyone outside the sample, and extrapolated the result to the whole class, producing a verdict near fifteen million dollars. The Supreme Court reversed on two independent grounds. The plan was a “trial by formula” — the phrase drawn from Wal-Mart Stores, Inc. v. Dukes (2011) 564 U.S. 338 — that violated the employer's due-process right to present its affirmative defenses; and the statistical model was unsound, the sample too small and unrepresentative and its margin of error intolerable. Sampling is not forbidden, but a representative-evidence plan must be statistically reliable and must preserve the defendant's ability to litigate its defenses as to the class.

The reliability requirement has real content, and it is where a defendant's statistician earns the engagement. A defensible sample is randomly drawn, large enough to support the inference, and representative of the class; its margin of error is quantified and controlled rather than assumed away. Because the opinion that an extrapolation is reliable is expert opinion, it is subject to the trial court's gatekeeping under Sargon Enterprises, Inc. v. University of Southern California (2012) 55 Cal.4th 747, which directs the court to exclude expert testimony that is speculative or unsupported by the material on which the expert relies. A sampling plan whose methodology cannot withstand that scrutiny does not become admissible because it is convenient. Federal law runs in parallel: Tyson Foods, Inc. v. Bouaphakeo (2016) 577 U.S. 442 cabined Dukes, permitting representative evidence where each member of the group could have relied on the sample to prove an individual claim — but not to supply proof the individual could not have offered. Duran sets the conditions any such plan must satisfy, and they are the checklist a defendant uses to test, and frequently to defeat, a proposed trial plan.

Fig. 5. The Duran conditions for representative proof.

The first condition deserves separate emphasis, because the permissible use of statistics depends on what they are asked to do. The scrutiny is far stricter when a sample is used to establish liability than when it computes aggregate damages after liability is shown — select each to see why.

Fig. 6. The liability–damages distinction in representative proof. The defendant's objection is strongest where the plaintiff would use a sample to prove the fact of violation across the group; it is weaker, though not absent, where liability is common and only the aggregate amount is extrapolated.

The trial plan as a certification requirement

Duran did more than police a completed trial; it instructed trial courts to confront manageability before certifying. A court must consider, at the certification stage, whether the case can in fact be tried — whether a workable trial plan exists that proves the class claims through common evidence while preserving the defendant's right to litigate its defenses. A plaintiff who cannot articulate such a plan has not shown that common issues are manageable, and certification should be denied; a court that certifies without confronting how the case will be tried has not conducted the rigorous analysis the law requires.

For the defense this converts the trial plan from a post-certification problem into a certification-stage demand. The opposition asks the court to require the plaintiff to specify, before any class is certified, how each violation will be proven across the class and how the employer's defenses will be litigated — and then tests that plan against Duran's reliability and due-process conditions. A plan that depends on extrapolating liability from an unrepresentative sample, or that has no answer for the employer's evidence that particular employees were provided breaks or signed valid waivers, is a plan that fails at the threshold. The same demand carries into the PAGA setting through a different door: Estrada forecloses striking the claim, but the court's manageability tools and the due-process floor still require the plaintiff to show a workable plan of proof for the aggrieved group, which is the subject of the next section.

PAGA without certification: manageability after Estrada

The PAGA action reaches the same group without the certification gate, which once led defendants to ask courts to strike PAGA claims as unmanageable in the way an uncertifiable class is dismissed. That route is now closed. Estrada v. Royalty Carpet Mills, Inc. (2024) 15 Cal.5th 582 held that a court has no inherent authority to strike a PAGA claim on manageability grounds, rejecting the contrary holding of Wesson v. Staples the Office Superstore, LLC (2021) 68 Cal.App.5th 746 and aligning with Woodworth v. Loma Linda University Medical Center (2023) 93 Cal.App.5th 1038, in which review was granted and held. The Court reasoned that PAGA is not a class action and that grafting a manageability-to-dismiss requirement onto it would import a class-action limitation the Legislature did not adopt. A meal-and-rest defendant cannot, therefore, defeat the representative penalty claim by demonstrating that trying it would be unwieldy.

What survives that holding is the line of containment that actually matters, and Estrada was careful to preserve it. The Court stated that PAGA's public purpose is subordinate to constitutional due process, and that trial courts retain their ordinary case-management tools — to be used, not to dismiss, but to keep the proceeding within due-process bounds. Those tools are substantial: a court may limit the evidence and the scope of the claim, require the plaintiff to present a workable plan of proof, define and narrow the aggrieved group, phase the proceeding, and admit representative testimony only under the reliability and fairness conditions Duran imposes. The 2024 reform reinforced the point by confirming the court's authority to limit the evidence presented at trial and the scope of a claim so that it can be effectively tried. The decisive synthesis is this: although the manageability-to-dismiss requirement does not apply to PAGA, the Duran due-process floor does — a PAGA representative trial may no more establish violations by an unreliable formula that denies the employer its defenses than a class trial may. The available course in a large PAGA matter is thus to constrain the representative claim rather than to strike it, and the penalty caps and cure pathways that compress what survives are treated in Civil Penalties under PAGA.

Fig. 7. The tools that survive Estrada. None dismisses the claim; each narrows it or disciplines its proof, and each is exercised within the due-process floor Duran sets — the same floor that governs a class trial.

The aggrieved-group problem

The PAGA group is bounded in a way the 2024 reform sharpened, and the boundary is a defense the class device has no analog for. Standing now requires the plaintiff to have personally suffered each of the Labor Code violations alleged, so the representative claim can reach only the violations the named plaintiff actually experienced. A plaintiff who suffered meal-period violations but no rest-period violations cannot pursue rest-period penalties on behalf of others; the personal-experience requirement caps the claim at the plaintiff's own injuries. The aggrieved group is then the set of employees who suffered those same violations during the limitations period, and defining it is itself contested — because the broader the group the plaintiff asserts, the more the proof must be both manageable under Estrada and reliable under Duran across the whole of it.

A defendant works that boundary from two directions. The first is to contest the named plaintiff's own experience, since a plaintiff who did not suffer a given violation cannot represent it — and where an arbitration program exists, the individual claim can be compelled and the plaintiff's status tested directly, the sequence developed in Arbitration, Standing & the Headless-PAGA Split. The second is to hold the plaintiff to proof of the defined violations across the defined group, resisting an expansion of the group beyond what reliable, manageable evidence can support. The group is not whatever the complaint asserts; it is what the plaintiff can establish, employee by employee or through admissible common proof, within the due-process limits that govern the trial.

The Donohue presumption at trial

The records-based presumption is not confined to the certification stage; it operates at trial, and it shapes the order of proof for the entire group. Under Donohue v. AMN Services (2021) 11 Cal.5th 58, time records showing that a meal period was short, late, or missed raise a rebuttable presumption that a compliant period was not provided — and the employer bears the burden of rebutting it. In a representative meal-and-rest trial the plaintiff can therefore establish a prima facie case from the employer's own records, which means the employer's evidence, not the plaintiff's, frequently determines the outcome.

That allocation makes the rebuttal evidence the center of the defense, and it is the same evidence that resists certification. The presumption is overcome by a showing that compliant periods were in fact provided notwithstanding the records — that the employee was relieved and chose to work or to shorten the break, that a valid waiver applied to the shift, or that the record reflects a recording practice rather than a missed period. Valid meal-period waivers, contemporaneous attestations, and a recording system that captures the reason a period was not taken are what supply that rebuttal, and they do double duty: deployed at trial they overcome the presumption, and deployed at certification they fragment the predominance the plaintiff needs. The presumption and the rounding and recording rules around it are developed in Records as Presumptive Proof; the waivers that rebut it, in Meal-Period Waivers.

Settlement and judicial approval

Most meal-and-rest matters resolve by settlement, and both vehicles are subject to judicial approval the defendant should understand before negotiating. A class settlement requires the court to act as guardian of the absent class members: under Kullar v. Foot Locker Retail, Inc. (2008) 168 Cal.App.4th 116, it must have before it enough information about the strength of the claims and their realistic value to assess whether the settlement is fair, adequate, and reasonable. The assessment runs through the familiar factors — the strength of the plaintiffs' case, the risk and expense of continued litigation, the stage of the proceedings and the discovery completed, the experience and views of counsel, the presence of a governmental participant, and the reaction of the class (Dunk v. Ford Motor Co. (1996) 48 Cal.App.4th 1794) — with the strength of the case balanced against the amount offered as the most important factor. Counsel's fee is scrutinized as well: a percentage-of-the-common-fund award is permissible, but the court may apply a lodestar cross-check to confirm its reasonableness (Laffitte v. Robert Half Internat. Inc. (2016) 1 Cal.5th 480), and incentive awards to the named plaintiffs receive their own review.

The PAGA component carries its own approval regime, and it does not run on the same track. Labor Code section 2699(l)(2) requires the superior court to review and approve any PAGA settlement, and the proposed settlement must be served on the Labor and Workforce Development Agency, which may object. Moniz v. Adecco USA, Inc. (2021) 72 Cal.App.5th 56 supplies the standard: the court evaluates whether the settlement is fair, reasonable, and adequate in view of PAGA's purposes — to remediate present labor-law violations, deter future ones, and maximize enforcement — and it must be able to assess the fairness of the allocation between the penalty recovery and the balance of the settlement. The Supreme Court in Turrieta v. Lyft, Inc. (2024) 16 Cal.5th 664 disapproved Moniz on one point only, holding that a nonparty PAGA plaintiff lacks standing to intervene in or challenge another plaintiff's settlement; Moniz's articulation of the approval standard otherwise governs.

Three structural features follow for the defendant. The penalty recovery is distributed thirty-five percent to aggrieved employees and sixty-five percent to the state, and aggrieved employees may not opt out of the PAGA settlement — so a combined class-and-PAGA settlement is frequently structured with two separate payments and two releases, the class release subject to opt-out and the PAGA release not (Amaro v. Anaheim Management (2021) 69 Cal.App.5th 521). The court will scrutinize the amount allocated to the PAGA claim to confirm it is fair to those it binds, which constrains a defendant's preference to load consideration onto the opt-out-bearing class release. And because Turrieta forecloses a competing PAGA plaintiff from disrupting an approved settlement, a defendant facing parallel representative actions retains a measure of control over which matter it resolves and on what terms — though the court's independent approval obligation, and the LWDA's right to be heard, remain the checks the defendant must satisfy.

Decertification — and its limits

Certification is not the end of the class inquiry. A class certified on the theory that a uniform policy controls can be decertified when the evidence developed in discovery shows that liability in fact depends on individual circumstances, and Duran teaches that a class which cannot be tried without depriving the employer of its defenses should not survive to verdict. The defendant's mid-case tools are real: a motion to decertify when individualized issues emerge, a motion to bifurcate liability from damages, and a challenge to any trial plan that would extrapolate liability from a sample without controlling its error or preserving the employer's defenses.

But the decertification argument has a limit a careful defendant respects, and Lubin v. The Wackenhut Corp. (2016) 5 Cal.App.5th 926 marks it. There the trial court decertified a class of security officers on the view that Wal-Mart v. Dukes barred the use of statistical sampling; the Court of Appeal reversed. Decertification, it held, may not rest on an overreading of Dukes — which Tyson Foods had since confined — nor on the frequency with which violations occurred, which is a damages question, nor on a demand that the plaintiff “conclusively establish” a uniform policy. The certification inquiry, on a decertification motion as on the original one, asks whether the theory of recovery is amenable to common treatment, not whether the plaintiff will win it. The lesson for the defense is that a decertification motion succeeds when it shows the absence of a common policy or means of proof, or a trial plan that cannot preserve the employer's defenses — and fails when it merely recharacterizes individualized damages as a predominance defect or stretches a federal commonality case past what California law has adopted. The mirror risk falls on the plaintiff: a class certified too broadly, or a PAGA group defined too expansively, can collapse at trial when the proof required to establish each violation proves unmanageable within due process.

The defensive posture

The procedural defense assembles from the pleadings forward rather than at the certification hearing, and it has a definite order. The first task is to disaggregate the theories and contest the uniform-policy characterization on which both certification and the PAGA group depend — to establish, through the written policy and the operational record, that the employer provided compliant periods and that the reasons employees did not take them vary across people, shifts, and locations. The second is to build the individualized record that does that work: valid waivers for the shifts that carry them, attestations that capture provision and the reason for any missed period, and a recording system that distinguishes a missed period from a recording artifact. The third is to demand a trial plan at the certification stage and to test it against Duran — insisting that the plaintiff specify how each violation will be proven across the group and how the employer's defenses will be litigated, and meeting any liability extrapolation with the gatekeeping challenge Sargon supplies, so that an unreliable formula is excluded rather than belatedly protested.

Where the representative claim cannot be defeated outright, the same materials constrain it. Estrada forecloses a motion to strike the PAGA claim, but the court's authority to limit evidence and scope — reinforced by the 2024 reform — together with the personal-experience boundary on the aggrieved group, lets a defendant hold the plaintiff to reliable proof and narrow the group to those for whom that proof exists. Combined with the penalty caps, the cure pathways, the scienter defenses to the derivative claims, and the arbitration sequence treated elsewhere in this section, the trial-plan discipline is what converts a representative meal-and-rest matter from an undifferentiated aggregate exposure into a claim the size of what the plaintiff can actually prove. That, rather than a single dispositive ruling, is how these matters are won and contained.