TCPA Class Certification: Using Consent to Defeat Predominance

In a 2018 TCPA fax case, the Ninth Circuit said it would “not consider the consent defenses that McKesson might advance or for which it has presented no evidence.” In my view, consent collected the same way from everyone tends to help the plaintiff. Consent that varies by lead source or by date can defeat certification, but only if the company has evidence of the variation, whether records, declarations, or testimony. I think a company that can show those differences has more defenses than the motion suggests.

What does a plaintiff have to prove to certify a TCPA class?

A TCPA damages class is sought under Rule 23(b)(3) of the Federal Rules of Civil Procedure, which requires the court to find that “the questions of law or fact common to class members predominate over any questions affecting only individual members,” according to the rule. That’s the predominance requirement.

The plaintiff carries the burden, and it isn’t a pleading burden. Sitting en banc in Olean Wholesale Grocery Cooperative v. Bumble Bee Foods, the Ninth Circuit held that plaintiffs must prove the facts showing each Rule 23 requirement “by a preponderance of the evidence.”

Consent cuts the other way. Prior express consent means the person called agreed, before the call, to be contacted at that number, and under Van Patten the call has to relate to the same subject matter as the context in which the person gave the number. The Ninth Circuit held in Van Patten v. Vertical Fitness Group that “prior express consent is an affirmative defense, not an element of a TCPA claim,” so the company has to prove it. In True Health Chiropractic v. McKesson, the court explained how those two burdens meet. The plaintiff still has to show predominance, “But the burden of proving consent strongly affects the analysis.”

When does consent that varies by lead source defeat predominance?

When the variation changes the answer to the consent question and the company can show it.

True Health shows both sides. McKesson argued that recipients of its fax advertisements had consented in different ways. For one group, the asserted consent came from product registrations and software license agreements, and the court found that, “So far as the record shows, there is little or no variation in the product registrations and the EULAs.” Predominance was satisfied for that group. For a group of 55 fax numbers, McKesson’s consent defense rested on “individual communications and personal relationships between McKesson representatives and their customers,” and the court affirmed the denial of certification for those numbers. It sent a third group, built on registration checkboxes and consent forms, among other things, back to the district court. Our post on defending junk fax claims covers what counts as permission for a fax.

Calling consent individualized is a slogan, and I don’t think slogans win certification fights. What works is a map: how many numbers came from the company’s website form, how many from each lead vendor, how many from in-store sign-ups or sales calls, and which version of the consent language each person saw. Every mechanism is a potential subclass. In True Health, the group that failed was the one whose consent rested on individual conversations, proven through a sales rep’s declaration and a former rep’s testimony. Vendor-sourced consent works the same way only if the company has evidence of what the vendor collected, such as the vendor’s records or testimony. If it can’t, the court may treat the defense as speculation and ignore it.

Time matters too. If the consent language changed in the middle of the class period, or a vendor relationship started or ended, the defense can argue that the class splits along those dates. For marketing calls to cell phones that use an autodialer (equipment with the capacity to store or produce numbers using a random or sequential number generator, and to dial them) or a prerecorded voice, FCC rules call for prior express written consent, a signed written agreement that meets specific disclosure requirements, under 47 C.F.R. § 64.1200(a)(2). On February 25, 2026, the Fifth Circuit held in Bradford v. Sovereign Pest Control of TX that “The statute provides no basis for concluding that telemarketing calls require prior express written consent but not oral consent,” as the earlier post on oral consent explains, but that ruling doesn’t bind California courts. Where the rule applies, a form that met it one year and not the next can split the class.

Typicality under Rule 23(a) can matter too. In Abboud v. Circle K Stores, a federal court in Arizona denied certification of a Do Not Call text class in September 2025, partly because the plaintiff denied having any business relationship with the company while its evidence suggested many class members had one. (The National Do Not Call Registry is the federal list of numbers whose owners have opted out of telemarketing.)

Does the company have to prove its consent defense at certification?

It has to show real evidence, not possibilities. True Health declined to consider consent defenses unsupported by evidence, refusing to let “speculation and surmise” decide a certification ruling.

A Sixth Circuit order from December 2025 shows what weak evidence looks like. In the Humana order, the court declined to hear Humana’s appeal from an order certifying a prerecorded-call class. Humana argued that its “wrong number” call notes didn’t prove lack of consent and offered declarations from thirteen people who said they had consented. The court wrote that “Humana’s argument that its own record-keeping is deficient or misleading is unavailing,” and that, “without knowing how many declarations it sought,” the thirteen declarations, “representing 0.05% of the potential class members,” didn’t raise predominance concerns. Still, the review was deferential, and the court acknowledged that Humana “does raise valid concerns” that consent issues could overwhelm common ones.

If consent really varies, I’d prove it before the certification motion, with a sample large enough to mean something and a record of how many people were contacted and how many responded.

Can a wrong-number or reassigned-number class be certified?

They can be. The Sixth Circuit left a certified class of this kind in place in Humana, and the Abboud order describes two earlier Arizona federal decisions, which that plaintiff cited as supporting certification, involving classes made up only of non-customers. In my view, the Ninth Circuit’s liability rule makes these claims dangerous, so the defense fight is often over who’s in the class.

The exposure comes from N.L. v. Credit One Bank, an individual case tried to a jury over autodialed debt-collection calls. The Ninth Circuit approved a jury instruction that consent from the “intended recipient of the call” wasn’t enough, and held, based on the statute’s “called party” wording, that the instruction “complied with the statute.” N.L. was itself a reassigned-number case: the customer’s number had been reassigned to the plaintiff’s mother, and the customer’s consent didn’t cover calls to the new holder.

The FCC does offer a narrow safe harbor. Under 47 C.F.R. § 64.1200(m), a caller that had consent from a number’s earlier subscriber, checked the Reassigned Numbers Database, and got an erroneous answer that the number hadn’t been disconnected isn’t liable for the call, but the caller is “bearing the burden of proof and persuasion” on that point. It doesn’t help a company that never queried the database.

On certification, I think the outcome usually depends on whether anyone can reliably sort non-customers from customers. In Davis v. Capital One, the Fourth Circuit affirmed the exclusion of the plaintiff’s expert and the denial of certification of a class of people who weren’t current customers and received prerecorded calls. Capital One’s expert tested the methodology on a sample and, according to the opinion, found that of the 666 numbers flagged as potential class members, “over 75 percent were in fact Capital One customers.” The decision is unpublished and doesn’t bind even the Fourth Circuit. The class ruling rested on the Fourth Circuit’s ascertainability rule, which the Ninth Circuit doesn’t follow.

Abboud is closer to home. It involved Do Not Call claims, where a call to someone with an established business relationship (for example, a purchase within the prior 18 months, as defined in § 64.1200(f)(5)) isn’t a telephone solicitation. The plaintiff proposed a subclass of people who never gave the company their numbers, some identified through “wrong number” replies. The court held that whether each one had a business relationship from in-store purchases “cannot be resolved via a common inquiry or a representative body of evidence.” That defense applies to Do Not Call claims. It won’t help with autodialed or prerecorded calls to cell phones, because the exceptions in § 227(b)(1)(A) don’t include a business relationship.

In my view, a company’s own data decides wrong-number cases, so I’d want a defense expert testing the plaintiff’s matching method on real data before the certification hearing. Our post on challenging a TCPA data expert covers how to test that methodology.

Does Briseno mean a class gets certified even if members can’t be identified?

No, though the Abboud plaintiff read Briseno too broadly. In Briseno v. ConAgra Foods, the Ninth Circuit held that “A separate administrative feasibility prerequisite to class certification is not compatible with the language of Rule 23.” So a defendant in a California federal court can’t win just by arguing that the plaintiff has no easy way to identify class members, the way Capital One did in the Fourth Circuit.

Briseno didn’t remove the identification problem from the case. It moved it. The court said that a separate feasibility rule “would render that manageability criterion largely superfluous,” meaning the difficulties of managing the class still count under Rule 23(b)(3). And individual defenses still count under predominance. When the Abboud plaintiff argued that the company’s business-relationship defense could be handled later in a claims process, the court responded: “This argument overreads Briseno.”

In the Ninth Circuit, don’t argue ascertainability as a freestanding requirement. Argue that consent or the called party’s identity can’t be resolved with common proof, and that the plaintiff’s method for sorting class members is unreliable under the evidence rules. Briseno also leaned on claim affidavits after a liability finding. The Supreme Court has since held in TransUnion LLC v. Ramirez that “Every class member must have Article III standing in order to recover individual damages,” so I wouldn’t treat the affidavit stage as a free pass for absent members.

What happens to class members who weren’t injured?

They can’t recover damages in federal court, and in the Ninth Circuit their presence is a predominance question. The earlier post on TCPA standing covers TransUnion and the standing rules in detail.

Olean rejected the argument that Rule 23 bars certifying a class with more than a de minimis number of uninjured members, but it said a court must consider whether uninjured members make the class definition “fatally overbroad.” Quoting the Seventh Circuit, the en banc court wrote that such a class “is defined too broadly to permit certification.” Olean also noted, citing earlier Ninth Circuit law, that a court may not create a fail-safe class, one defined to include only people the defendant injured. Plaintiffs often define TCPA classes by objective facts instead, as in Davis and Abboud, such as calls to non-customers or to numbers on the Registry. Those definitions tend to sweep in people who consented or had a business relationship, and that’s the opening. Abboud applied the Olean overbreadth rule to deny certification of a class that included millions of people who had given the company their numbers.

How the main certification defenses line up

The records a defendant needs before the certification motion

1. Put a litigation hold on call and text logs, dialer and texting-platform data, and every consent record for the numbers in the proposed class, including the version of the consent language each person saw and the lead source it came from.
2. Pull opt-out and internal do-not-call records with dates, along with wrong-number dispositions and number-change notes, so revocations and reassignments can be sorted by person and date.
3. Pull the National Do Not Call Registry scrub records, the version dates used, and any Reassigned Numbers Database queries and responses.
4. Collect lead-vendor and data-broker contracts, the consent records each vendor can actually produce, call scripts, and any recorded or prerecorded message files.
5. Talk to a lawyer before responding to the demand or complaint, contacting the named plaintiff or class members, or offering anything to settle.

The firm’s page on TCPA litigation defense explains how it approaches class certification in TCPA cases.

Questions businesses ask about TCPA class certification

If the court certifies a class, is that the end of the consent defense?

No. Rule 23(c)(1)(C) provides that “An order that grants or denies class certification may be altered or amended before final judgment.” A company that develops better consent evidence after certification can move to decertify or to narrow the class. In my view, a decertification motion needs new evidence, not the same argument made again.

Can we appeal an order certifying a TCPA class right away?

Only with permission. Rule 23(f) lets a court of appeals permit an appeal from a certification order, and a short filing deadline applies. The Sixth Circuit wrote that “the Rule 23(f) appeal is never to be routine.” Humana’s petition failed mainly because it hadn’t shown an abuse of discretion, and it offered no financial data showing the certification would force a settlement.

Does a wrong-number note in our system hurt or help us?

In my view, it mostly helps the plaintiff. It’s evidence the person never consented and a way to find class members. Humana argued that its notes could mean many things and lost, and the court cited a decision saying courts “do not look favorably upon the argument that records a defendant treats as accurate for business purposes are not accurate enough to define a class.” If your notes really are unreliable, I’d prove it with a meaningful sample, not a handful of declarations.

Do we need an expert to oppose class certification?

In wrong-number cases, I’d usually say yes. Plaintiffs rely on experts to match phone numbers to subscribers, and the Humana order noted that courts have disagreed about whether one expert’s “reverse-append” method is reliable. In Davis, a defense expert’s test showing the method flagged mostly existing customers was one of several reasons the court upheld excluding the plaintiff’s expert, along with her failure to fully test her own method.

Opposing class certification in a TCPA case

If your company is facing a motion to certify a TCPA class, the consent records and lead-source data you hold often decide whether the class holds together. Rosenberger + Kawabata represents businesses defending TCPA lawsuits and class actions through its business litigation practice, including class claims over calls to wrong or reassigned numbers. Contact Rosenberger + Kawabata online for a free and confidential consultation, or call (310) 894-6921.

Sources

The class action rule is Federal Rule of Civil Procedure 23. The statute is the Telephone Consumer Protection Act at 47 U.S.C. § 227. The FCC’s TCPA rules, including the written-consent rule, the established business relationship definition, and the Reassigned Numbers Database safe harbor, are at 47 C.F.R. § 64.1200 (as in effect March 18, 2026).

The Ninth Circuit decisions are Olean Wholesale Grocery Cooperative, Inc. v. Bumble Bee Foods LLC, 31 F.4th 651 (9th Cir. 2022) (en banc); True Health Chiropractic, Inc. v. McKesson Corp., 896 F.3d 923 (9th Cir. 2018); Van Patten v. Vertical Fitness Group, LLC, 847 F.3d 1037 (9th Cir. 2017); Briseno v. ConAgra Foods, Inc., 844 F.3d 1121 (9th Cir. 2017); and N.L. v. Credit One Bank, N.A., 960 F.3d 1164 (9th Cir. 2020) (abrogated on the autodialer issue by Facebook, Inc. v. Duguid, 592 U.S. 395 (2021)).

The Supreme Court decision is TransUnion LLC v. Ramirez, 594 U.S. 413 (2021).

The other decisions are Bradford v. Sovereign Pest Control of TX, Inc., No. 24-20379 (5th Cir. Feb. 25, 2026); In re Humana, Inc., No. 25-0502 (6th Cir. Dec. 30, 2025); Davis v. Capital One N.A., No. 24-1507 (4th Cir. Aug. 25, 2025, amended Aug. 26, 2025) (unpublished); and Abboud v. Circle K Stores Inc., No. CV-23-01683-PHX-DWL (D. Ariz. Sept. 30, 2025).

This post is general information about the Telephone Consumer Protection Act and related laws as of March 18, 2026. It is not legal advice, and reading it does not create an attorney-client relationship. Courts and the FCC change the rules in this area often, and the outcome of any claim depends on its facts. Past results do not guarantee future outcomes.