Mass Arbitration: How Thousands of Individual Claims Became a Strategic Problem

Document type: Article Practice area: Litigation — Arbitration Jurisdiction: United States (federal) Last reviewed: 5 September 2026

For about twenty years, the defense-side arbitration strategy was straightforward and it worked.

Put an arbitration clause in the consumer terms. Add a class action waiver. The Supreme Court will enforce it — AT&T Mobility LLC v. Concepcion, 563 U.S. 333 (2011) held that the FAA preempts state rules conditioning enforceability on class availability; American Express Co. v. Italian Colors Restaurant, 570 U.S. 228 (2013) held that the economic irrationality of pursuing a small claim individually does not invalidate the waiver; Epic Systems Corp. v. Lewis, 584 U.S. 497 (2018) held the National Labor Relations Act does not override class waivers in employment agreements.

The theory was that most small claims would simply never be brought. A consumer with a $60 dispute will not hire a lawyer, and a lawyer will not take a $60 case one at a time. The class device existed precisely to aggregate such claims, and the waiver eliminated it.

Then somebody did the arithmetic differently.

If a company's arbitration clause requires individual arbitration, and if the company's own terms require it to pay the arbitrator's fees, then filing ten thousand individual arbitrations imposes a cost on the company that has nothing to do with the merits. At consumer filing fee schedules, the administrative and arbitrator fees can run one to two thousand dollars per case, paid predominantly by the business, before anyone examines a single claim. Ten thousand claims is ten to twenty million dollars in fees. The claims could all be worthless and the bill would be the same.

That is mass arbitration. It is not a loophole; it is the clause working exactly as written, applied by someone who read it more carefully than the drafter did.

How the mechanism works

The technique is simple to describe and operationally demanding to execute.

Step one — claimant acquisition. A firm advertises, typically online, to people who used a product or service. Sign-up is a web form. The economics support large-scale advertising because the per-claim recovery target is modest and the volume is enormous.

Step two — mass filing. The firm files individual arbitration demands, often in tranches of several hundred to several thousand, with the institution named in the company's own clause.

Step three — the fee event. The institution invoices the initial administrative and arbitrator fees. Under consumer arbitration protocols — which most institutions adopted specifically to make consumer arbitration fair — the business pays most of them. The invoice arrives before any assessment of the merits.

Step four — the choice. The company can pay the fees and arbitrate thousands of individual cases, each with its own arbitrator, schedule, and hearing. Or it can settle. Or it can refuse to pay, which triggers its own set of problems.

Step five — the leverage. The settlement value is driven not by the merits but by the process cost. A company facing $18 million in filing fees for claims worth perhaps $400 each has an obvious incentive to settle for less than the fee exposure regardless of liability.

The precision of the strategy is what makes it effective. It uses the company's own contract, the institution the company selected, the fee schedule the institution adopted for consumer fairness, and the individual-claims structure the company insisted on. Every element was chosen by the defense.

What happened when companies tried to get out

The first wave of responses produced a body of case law that practitioners should know, because it establishes the boundaries of what works.

Refusing to pay. Several companies declined to pay the institutional fees for mass filings. The institutions closed the cases for nonpayment, and claimants went to court to compel arbitration — under the companies' own clauses. Courts compelled. One district judge, addressing a delivery platform facing several thousand courier claims, observed with some relish that the company had wanted individual arbitration and now had it, and declined to relieve it of the consequences. The opinion is widely quoted for the proposition that a party that drafted an arbitration clause cannot complain when the other side uses it.

Amending the clause mid-stream. Companies added mass arbitration provisions — batching, bellwethers, staged filing — after mass filings began. Courts have generally held that the version in effect when the claim arose or when the claimant accepted governs, and that unilateral amendment applied retroactively to pending claims is ineffective and sometimes evidence of unconscionability.

Switching institutions. Some companies moved to newer providers offering mass-filing procedures at lower cost. The Ninth Circuit held one such arrangement unconscionable, finding that the provider's mass arbitration protocol — which permitted a small number of bellwether cases to bind thousands of non-participating claimants, gave claimants no meaningful role in arbitrator selection, and sharply limited discovery and review — was so one-sided that it could not be enforced, and further held that the delegation clause failed for the same reason.

Aggressive bellwether provisions. Clauses providing that a handful of test cases would be arbitrated first, with all other claims stayed indefinitely and no tolling or process guarantee, have been struck down as unconscionable. Courts objected that claimants outside the bellwether group had no path to a hearing on any timeline they could predict.

The lesson from all of it: countermeasures drafted to make mass arbitration impossible tend to fail. Countermeasures drafted to make mass arbitration orderly tend to survive. The difference is whether every claimant retains a real path to an individual adjudication.

The institutional response

Arbitration providers, which faced their own problems administering thousands of near-identical filings, restructured their rules.

The AAA's mass arbitration supplementary rules apply when a threshold number of similar demands are filed against the same party by the same or coordinated counsel. Their principal features:

A process arbitrator, appointed to resolve administrative disputes — whether filings meet the requirements, whether claims are properly within the rules, fee allocation questions, and the mechanics of the process — separate from the merits arbitrators.
A restructured fee schedule, with a per-case initiation fee substantially lower than the standard consumer schedule and a global filing structure that reduces the up-front cliff.
Filing requirements, including verification that each claimant has authorized the filing and that the demand contains individualized information.
Administrative consolidation of common procedural questions.

JAMS mass arbitration procedures take a comparable approach with different mechanics, including a process administrator, a modified fee structure, and provisions for coordinated scheduling.

What the rules changed. The pure fee-leverage play is weaker than it was. A per-case initiation fee in the low hundreds rather than the low thousands changes ten thousand claims from a twenty-million-dollar event to something more manageable.

What they did not change. The underlying arithmetic still favors volume. And the rules do not resolve the merits: a company facing ten thousand claims still has ten thousand claims.

The clause architecture that has emerged

A consensus design has formed among companies that have thought about this carefully. Its components are worth understanding both as drafting guidance and as a map of where the litigation will go.

Informal dispute resolution as a genuine condition precedent

The most effective single provision requires an individualized pre-arbitration process:

Written notice of the dispute, containing the claimant's name, contact information, account identifier, a description of the dispute, and the specific relief sought.
Signed by the individual claimant, not only by counsel. This requirement — that the claimant personally sign or verify — is the operative feature. It converts a web-form sign-up into an individualized act.
A negotiation period, typically thirty to sixty days, during which the company must actually engage.
A telephone conference with the individual claimant, at a mutually convenient time.
A tolling provision so the process cannot be used to run out limitations periods.

Why it works. It imposes a per-claimant cost on the claimant side proportional to the per-claim cost on the company side, and it does so through a process that is genuinely useful — many disputes settle at this stage. Courts have generally enforced these provisions where the company actually performs its side.

Why it fails when it fails. A company that treats informal resolution as a paperwork gate — auto-responding to notices, declining to schedule conferences, refusing to engage — will find the provision read against it. The condition must be real in both directions.

Batching, not blocking

Provisions grouping claims into batches of, say, fifty or one hundred, with each batch assigned to a single arbitrator and proceeding on a defined schedule, have generally been enforced when they satisfy three conditions:

Every claim gets adjudicated. No claimant is left permanently stayed.
The schedule is defined. Batches proceed at stated intervals, not at the company's discretion.
Limitations are tolled for claims awaiting their batch.

What courts strike down is the version that stays the ninth batch through the twentieth indefinitely while the first eight proceed, with no timetable and no tolling.

Bellwethers with an off-ramp

Bellwether structures survive when they include a mediation or global-resolution step after the test cases, and a defined path forward if resolution fails. They fail when the bellwether results bind non-participating claimants, or when the non-bellwether claims have no path at all.

Small-claims carve-outs

Many clauses permit either party to bring an individual claim in small claims court instead of arbitration. This is genuinely pro-consumer, it is cheap, and it reduces the pool of claims that generate arbitration fees. It also improves the clause's unconscionability posture materially.

Fee-shifting for meritless filings

Provisions permitting an arbitrator to award fees against a party that files a claim found to be frivolous or brought for an improper purpose are enforceable if mutual and tied to a standard resembling Rule 11. They are aimed at the subset of mass filings that include claimants who never used the product.

A meaningful opt-out

A thirty-day opt-out with a simple mechanism, honored in practice, remains one of the strongest answers to a procedural unconscionability argument. Opt-out rates are consistently tiny, and the provision costs almost nothing.

And what not to do

Do not draft a provision that makes arbitration practically unavailable to any claimant.
Do not amend the clause after mass filings begin and apply it retroactively.
Do not select a provider whose mass procedure gives claimants no role in arbitrator selection or no individual adjudication.
Do not impose costs on claimants that exceed what a court would charge. Green Tree Financial Corp.-Alabama v. Randolph, 531 U.S. 79 (2000) placed the burden on the party resisting arbitration to show prohibitive costs — but a clause that plainly imposes them is inviting the showing.

A worked example

Corrigan Fitness Holdings operates a subscription app with 3.1 million users. In March, its outside counsel receives notice that a plaintiffs' firm has filed 6,400 individual arbitration demands with the AAA alleging that Corrigan's auto-renewal disclosures violated several state statutes. Average claimed damages: about $210.

The fee exposure. Under the AAA's mass arbitration supplementary rules, the per-case initiation fees are far below the standard consumer schedule — but 6,400 cases still generate a substantial six-figure initial invoice, and if the cases proceed individually, arbitrator compensation across 6,400 matters would be an eight-figure number.

The merits exposure. Even assuming full liability, 6,400 × $210 is about $1.34 million. Statutory damages in two states could push it higher, perhaps to $4 million.

The problem, stated plainly: the process could cost several times the maximum merits exposure.

What Corrigan's counsel, Rosalind Achterberg-Nwosu, actually does.

Week one — verify the filings. She obtains the demand list and runs it against Corrigan's user database. Findings: 4,850 are matched active or former subscribers. 900 cannot be matched to any account. 410 match accounts that were never charged a renewal fee. 240 are duplicates.

This matters enormously, and it is the first step in every mass arbitration. The process arbitrator under the AAA rules can address filings that do not meet the requirements, and 1,550 unmatched, uncharged, or duplicate claims is a real argument, made with data.

Week two — check the clause. Corrigan's terms, last revised in 2023, contain: individual arbitration, a class waiver with a severance clause, an express delegation provision, and — this is the good news — a thirty-day informal dispute resolution requirement with an individualized written notice requirement signed by the claimant.

The claimants had not complied. The firm had sent a single letter listing 6,400 names. Corrigan's clause required individual notices with account details and the specific relief sought.

Week three — the process arbitrator. Rosalind raises two issues: compliance with the IDR condition precedent, and the 1,550 filings that do not correspond to charged accounts. She does not argue that mass arbitration is improper as such, because that argument loses and it damages her on everything else.

The outcome of that phase. The process arbitrator holds that the IDR requirement is a condition precedent the claimants must satisfy, and stays the filings for sixty days to permit compliance. It also directs that claimants provide account identifiers.

What that produced. Of 6,400, approximately 4,100 completed individualized notices. The rest fell away — some were duplicates, some could not identify an account, some claimants did not respond to their own counsel.

Week eight through twenty — actual resolution. Corrigan reviews the 4,100 and finds that about 1,700 involve a specific disclosure defect in an interface version live for eleven weeks in 2023. Those claims have genuine merit. The remaining 2,400 involve interface versions Corrigan believes were compliant.

The settlement. Corrigan offers full refunds plus a modest statutory increment to the 1,700, resolving them at about $780,000. It arbitrates a batch of the remaining claims as bellwethers, wins most, and resolves the rest at nuisance value.

Total cost: roughly $2.4 million including fees and defense costs, against an initial exposure that looked like it could exceed $15 million.

Three things made the difference.

One — the clause had a real IDR provision, drafted before the filings. It could not have been added afterward.

Two — Corrigan verified the filings against its own data in week one. A quarter of the claims did not survive contact with the account records.

Three — Corrigan took the meritorious claims seriously. The 1,700 claims with a real disclosure defect were real claims, and paying them promptly cost less than litigating them and removed the moral force from the rest of the campaign.

What this looks like from the plaintiff side

It is worth stating the other view honestly, because it explains why the practice is durable.

The claims are frequently real. Auto-renewal defects, junk fees, mislabeled products, and data practices that violate state statutes affect millions of people in amounts too small to litigate individually. When the class device is unavailable by contract, mass arbitration is the only aggregation mechanism left.

The system was designed by the defense. Companies chose individual arbitration, chose the institution, and accepted the consumer fee protocols. A claimant using that structure is doing what the contract says.

The fee protocols exist for a reason. Consumer arbitration protocols shifted fees to businesses precisely because arbitration would otherwise be unaffordable for individuals. A company arguing that those fees are unfair at scale is arguing against the condition on which its clause was upheld.

And courts have noticed the asymmetry. The most quoted judicial observations in this area are directed at companies that wanted individual arbitration until they got it.

Where the practice is genuinely criticized — including by courts and by some plaintiffs' lawyers — is the subset of campaigns built on unverified claimant lists. Filings on behalf of people who never used the product, who did not authorize the filing, or who appear twice are a real problem, and the institutional rules requiring claimant verification respond to it directly.

The employment version

Mass arbitration in employment is structurally similar and legally distinct in three ways.

The claims are larger. Wage and hour claims — unpaid overtime, meal and rest breaks, off-the-clock work, misclassification — commonly run into thousands or tens of thousands of dollars per worker. The fee-leverage dynamic still operates, but the merits exposure is real in a way it often is not in consumer cases.

The class waiver is on firmer footing but the collective mechanism is different. Epic Systems settled that the NLRA does not bar class waivers in employment arbitration agreements. But the Fair Labor Standards Act collective action operates by opt-in, which means a mass arbitration campaign and an FLSA collective are functionally similar in a way that a mass arbitration and a Rule 23 class are not.

PAGA and its analogues complicate everything. California's Private Attorneys General Act authorizes an aggrieved employee to sue on behalf of the state for civil penalties. Viking River Cruises, Inc. v. Moriana, 596 U.S. 639 (2022) held that the FAA preempts a state rule barring division of PAGA actions into individual and non-individual claims, so an employer may compel arbitration of the individual PAGA claim. The Court's suggestion that the non-individual claims would then be dismissed for lack of standing was subsequently rejected as a matter of California law by the California Supreme Court, which held that a plaintiff retains standing to pursue the representative claims in court. The practical result is a bifurcated proceeding, and it is a live area.

The Ending Forced Arbitration Act removes sexual assault and harassment claims from predispute arbitration at the claimant's election, which means an employment mass arbitration campaign will sometimes have a court-side component the employer cannot compel.

The drafting consequences for employment agreements:

The IDR provision must be realistic for an individual worker — no requirement that a claimant retain counsel or produce documents they do not have.
Fee allocation must satisfy the applicable state standards, which in several states are more demanding than the federal baseline.
Batching must account for the possibility that individual claims are large enough to warrant individual treatment.
The clause must address what happens to PAGA or analogous representative claims.

Does the arbitration clause still make sense?

This is the question that should be asked and usually is not.

The original case for consumer arbitration rested on eliminating class exposure. That case is weaker now. Mass arbitration reproduces aggregate exposure through a different channel, and it does so on a fee structure the company pays.

Run the comparison honestly:

	Class action in court	Mass arbitration
Aggregation	Rule 23 certification required	No certification; volume is the mechanism
Gatekeeping	Certification, dispositive motions	Very little; fees precede scrutiny
Who pays the process	Each side its own; court is free	Business pays most institutional fees
Merits screening	Judge, early	Rarely, and late
Precedent	Published, binding	None
Appeal	Available	Effectively none
Publicity	Public docket	Usually confidential
Settlement structure	Class-wide release, court-approved	Claim-by-claim, no global peace
Global peace	Achievable	Very hard

The last row is the one most often overlooked. A certified class settlement, approved by a court, binds absent class members and ends the exposure. A mass arbitration settlement binds only the claimants who signed. The next firm can file another six thousand demands next quarter. For a company with a systemic practice affecting millions of users, class settlement may be the only route to finality.

Some companies have concluded that arbitration no longer serves them and removed the clause from consumer terms. That is a defensible business decision, and it should be made deliberately rather than by inertia.

Others retain arbitration but redesign it — real informal resolution, small claims carve-outs, orderly batching, verification requirements, and a serious claims-handling function. That is the mainstream approach and it works when the company actually operates the process rather than merely papering it.

Where the law is heading

Four developments are worth watching.

Unconscionability doctrine is doing more work. The decisions striking down aggressive mass arbitration countermeasures apply ordinary state contract law, which § 2 of the FAA preserves. As long as the challenges are grounded in generally applicable unconscionability principles rather than arbitration-specific hostility, they survive preemption. Expect more of them.

Delegation clauses are not a shield here. Several courts have held that where the arbitration procedure itself is unconscionable, the delegation provision embedded in it fails too, because the same defects infect it. A company relying on delegation to keep a court from reviewing a mass arbitration protocol should not assume it will work.

Institutional rules will keep evolving. The AAA and JAMS rewrote their mass arbitration procedures under pressure from both sides, and further revision is likely as the volume continues.

Legislative attention is intermittent but real. Proposals to restrict predispute consumer and employment arbitration recur in Congress and in state legislatures. The Ending Forced Arbitration Act showed that targeted federal legislation is achievable when a specific category of claim attracts consensus.

And one structural point. Morgan v. Sundance, Inc., 596 U.S. 411 (2022) articulated the principle that will govern most of this: the federal policy "is about treating arbitration contracts like all others, not about fostering arbitration." Courts applying that principle will neither rescue companies from clauses they drafted nor invent special rules to defeat them. Both sides get ordinary contract law, applied to a contract the company wrote.

The economics, in detail

Understanding the numbers is what separates a considered response from a panicked one. Here is the arithmetic, laid out.

The variables:

N — the number of filed demands.
F — the per-case fee borne by the business at the initiation stage.
A — arbitrator compensation per case if the case proceeds to a hearing.
D — internal and outside defense cost per case.
V — average merits value per meritorious claim.
p — the proportion of demands that are verifiable and meritorious.

Total process cost ≈ N × F + (cases proceeding) × (A + D). Total merits exposure ≈ N × p × V.

The pathology of mass arbitration is that the first number can exceed the second by an order of magnitude when V is small. That is why the strategy exists.

Which levers actually move the numbers:

Reducing N. This is where verification work pays. Duplicates, unmatched claimants, claimants who were never charged, and claimants who do not complete an individualized notice all reduce N — often by twenty to forty percent, sometimes more. This is a data exercise, and it is the highest-return activity in the first month.

Reducing F. Largely determined by the institution's mass arbitration fee schedule and by the clause. Selecting an institution with a defensible mass procedure is a drafting decision made years in advance.

Reducing the number of cases that reach a hearing. Batching, bellwethers, and early resolution of meritorious claims all work here. Every claim resolved before an arbitrator is appointed saves A + D.

Increasing p's visibility. Knowing which claims are real allows a company to pay them and defend the rest. A company that cannot distinguish settles the whole population at an average that overpays the meritless and underpays the meritorious.

What does not move the numbers: arguing that mass arbitration is improper. It is not, and the argument costs credibility with the process arbitrator and the court.

Practical governance: who owns this

Mass arbitration exposure sits at the intersection of legal, product, data, and customer support, and companies that handle it well have assigned ownership before anything is filed.

Legal owns the clause, the institutional relationship, and the response playbook.

Data and analytics own the ability to run a claimant list against account records within days. This capability — a documented query that takes a list of names, emails, and phone numbers and returns account status, charge history, and interface version — is the single most valuable operational asset in a mass arbitration response. Build it before you need it.

Product and engineering own the formation-evidence package: which terms version was in effect, what the interface displayed, and the acceptance logs. This is the same package needed for any motion to compel.

Customer support owns the informal dispute resolution process. If the clause requires a thirty-day negotiation period and a telephone conference, someone must actually conduct them, at scale, with records. A company whose IDR provision is enforced but whose support function cannot perform it has created an obligation it will breach.

Finance owns the reserve. Mass arbitration exposure is lumpy and it arrives without warning.

And someone senior should own the question of whether the clause is still worth having. That review should happen annually, with the class-versus-mass comparison table in front of the decisionmaker.

The short version

Mass arbitration is the predictable consequence of a strategy that worked too well. Companies removed the class device and left individual arbitration as the only channel, then discovered that the channel could carry volume.

The countermeasures that work make the process orderly. Individualized notice with a real negotiation period, batching with defined schedules and tolling, verification requirements, small claims carve-outs, meaningful opt-outs, and mutual fee-shifting for frivolous filings.

The countermeasures that fail make the process unavailable. Indefinite stays, binding bellwethers, retroactive amendments, and providers whose protocols deny individual adjudication.

And the most valuable response is not a clause at all. It is the ability to run a claimant list against the company's own records in week one, to identify which claims are real, and to pay those promptly. The companies that handle mass arbitration well are the ones that treat it as a claims problem with a legal component, rather than a legal problem to be papered around.

How the first ninety days actually go

For a company receiving its first mass filing, the sequence is predictable enough to plan.

Days 1–5. Notice arrives from the institution, typically as a bulk filing notification with a claimant list. Obtain the list in a usable format immediately. Notify insurers — several policies respond to mass arbitration and notice provisions are strict. Assemble the response team: legal, data, product, support, finance.

Days 5–20. Run the claimant list against account records. Produce four buckets: verified with charges, verified without charges, unmatched, and duplicates. This is the single most important work product of the period, and it will be used in every subsequent negotiation and filing.

Days 10–25. Read the clause in the version applicable to each claimant. In a company with several terms versions live over the claim period, different claimants may be governed by different clauses. Map claimants to clause versions.

Days 15–30. Assess condition-precedent compliance. Did claimants provide individualized notice? Did they sign? Did they identify accounts and specify relief? Did the company perform its side of any IDR obligation? This last question is asked honestly, because a court will ask it.

Days 20–40. Engage the process arbitrator or process administrator under the applicable mass procedure. Raise administrative issues with data: unverified filings, condition-precedent noncompliance, fee allocation, and the proposed sequencing. Do not argue that mass arbitration is impermissible.

Days 30–60. Merits triage on the verified population. Identify the subset with genuine exposure — a defect in a specific interface version, a specific billing period, a specific state's requirements. This is claims work, not legal work, and it should be done by people who know the product.

Days 45–90. Begin resolving the meritorious subset. Negotiate the process for the rest: batching schedule, bellwether selection, mediation timing. Set a reserve.

Throughout. Preserve everything. A litigation hold covering the relevant interface versions, billing records, support tickets, and internal communications should issue in week one, and it should be broad enough to cover the class action that may follow if the arbitration clause is held unenforceable as to any group.

Frequently asked questions

Can we just refuse to pay the arbitration fees? It has been tried. Institutions close the cases for nonpayment, claimants move to compel in court under the company's own clause, and courts compel. Some clauses and some states also treat nonpayment as a material breach that waives the right to arbitrate, letting claimants proceed in court — which is worse. Refusal is not a strategy.

Can we amend our terms to add a mass arbitration protocol? Prospectively, yes, with proper notice and acceptance. Retroactively as to pending or already-accrued claims, generally no — and doing it after filings begin is evidence of unconscionability.

Are batching provisions enforceable? Generally yes, if every claim is adjudicated on a defined schedule with limitations tolled. Generally no, if claims outside the first batches are stayed indefinitely.

Can a bellwether result bind non-participating claimants? No. That is the provision the Ninth Circuit found unconscionable, and it is the clearest line in this area.

What about a small claims carve-out? Include it. It is pro-consumer, it reduces the fee-generating pool, and it strengthens the clause against unconscionability challenges.

Does a delegation clause protect our mass arbitration protocol from judicial review? Not reliably. Where the protocol itself is unconscionable, courts have held that the delegation provision fails with it.

How quickly can we identify which filings are real? If the data capability exists, days. If it does not, weeks — and the weeks are expensive, because the process arbitrator's schedule does not wait.

Should we settle globally or claim by claim? Neither, usually. Segment: pay the verified meritorious claims promptly, contest the unverified, and bellwether the genuinely disputed. A global number set before segmentation overpays.

Does a mass arbitration settlement give us peace? Only with the claimants who signed. There is no analogue to a certified class release. This is the strongest argument for keeping a path to class settlement available.

Should we keep the arbitration clause at all? Ask the question annually with the comparison table in hand. For some businesses the answer is now no. For most it remains yes, provided the clause is designed for orderly mass handling and the company can actually operate the process it promises.

What is the single most useful thing we can do before anything is filed? Build the query that matches a claimant list against account records, charge history, and interface version. It is a data engineering task, it takes a few weeks, and it is worth more than any clause revision.

What plaintiffs' counsel should get right

The durability of mass arbitration depends on the quality of the filings, and the campaigns that have drawn judicial criticism share the same defects.

Verify claimants before filing. Institutional rules now require confirmation that each claimant authorized the demand and that the demand contains individualized information. A list assembled from a web form without verification produces duplicates, people who never used the product, and people who did not know their name was submitted. Every one of those is ammunition for the respondent and a problem with the process arbitrator.

Satisfy the conditions precedent. If the clause requires individualized written notice signed by the claimant, send individualized written notices signed by the claimants. A bulk letter listing six thousand names does not comply, and the resulting stay costs months.

Keep the individualized information individual. Demands that are identical except for a name invite an argument that the claims were not individually assessed. Account numbers, dates, amounts, and the specific transaction at issue take more work and they make the filings much harder to dismiss.

Be prepared to try cases. A campaign whose only value is fee leverage is vulnerable once the fee structure changes, and the institutional rules changed it. Campaigns built on claims that can actually be won are durable.

And engage the informal process. Where the clause requires a negotiation period and the company performs, many claims resolve there — faster and at higher net recovery than an arbitration would produce.

A closing observation

There is a tidy irony in all of this, and it is worth stating because it points at the underlying lesson.

The consumer arbitration protocols that shift fees to businesses were not imposed on anyone. They were adopted by the arbitration providers, with industry participation, to answer the criticism that arbitration was unaffordable for individuals — and courts upholding class waivers relied on those protocols as evidence that individual arbitration remained a viable channel for small claims. Concepcion itself emphasized the consumer-friendly features of the AT&T clause at issue, including provisions that made individual arbitration cheap and quick for the consumer.

Mass arbitration takes that reasoning seriously. If individual arbitration is a viable channel for small claims — and companies argued, successfully, that it is — then individual arbitrations will be filed. Many of them.

The lesson for drafters is not that arbitration clauses are traps. It is that a dispute resolution provision is an operating system, not a shield, and it will be run by people whose interests are opposed to yours. The right question at the drafting stage is not "what does this prevent" but "what does this cause, if someone reads it carefully and has a reason to use it at scale."

Companies that ask that question end up with clauses that handle volume gracefully, claims functions that can distinguish real disputes from noise, and dispute resolution provisions they are willing to have enforced against them. Companies that do not end up with an invoice.