The Charge Nobody Remembers Agreeing To: Subscription Auto-Renewal and Free-Trial Disputes in 2026

There is a category of dispute that sits uncomfortably between fraud and buyer's remorse, and it now accounts for a substantial share of the dispute volume most issuers handle. The customer sees a charge they do not recognise. They did authorise it, sixteen months ago, at the point of taking a free trial, through a checkout flow that mentioned auto-renewal in terms that were legally sufficient and cognitively invisible. They do not remember. From the issuer's side the transaction is a properly authenticated recurring payment from a legitimate merchant. From the customer's side it is money leaving their account for something they are not using and did not knowingly agree to keep paying for.

Both descriptions are accurate, which is why this category is so operationally expensive. It cannot be resolved by determining whether the transaction was fraudulent, because it was not. It cannot be resolved by determining whether the customer consented, because in a formal sense they did. The dispute is about whether the consent was meaningful, whether the cancellation route was usable, and whether the merchant did what the rules increasingly require. None of those questions can be answered from transaction data.

Through 2026 the position has shifted meaningfully. Consumer protection regimes in several major markets have moved from disclosure-based requirements towards outcome-based ones: reminders before renewal, cancellation that must be as easy as sign-up, and explicit consent for the transition from trial to paid. Card scheme rules have moved in parallel. The effect is that a category which used to be adjudicated on documentation is increasingly adjudicated on merchant behaviour, and a large number of merchants have not yet noticed.

Regulatory and Market Context

The regulatory direction is consistent across jurisdictions even where the instruments differ. The common elements are pre-renewal notification with sufficient lead time to act, cancellation that is available through the same channel used to subscribe and requires no materially greater effort, clear and prominent disclosure of the full price and renewal cadence before the commitment is made, and (for trials) affirmative consent to the transition to a paid subscription rather than reliance on inertia. Regimes vary in their specifics and in their enforcement vigour, but a merchant designed to the strictest of them will pass the others, and a merchant designed to the weakest will fail several.

Enforcement has also changed character. Historically, subscription practices attracted consumer-protection attention as a conduct matter with reputational consequences and modest financial ones. The more recent pattern involves substantial penalties, mandated redress to affected customers, and (significantly for acquirers) undertakings that change how the merchant may present and bill subscriptions going forward. That converts a conduct issue into a commercial-model issue, and a merchant whose acquisition economics depend on unremembered renewals has a business risk rather than a compliance one.

The scheme rules operate on a separate track and are the mechanism issuers actually use. Both major networks require pre-notification for recurring charges in defined circumstances, require confirmation of the transition from trial to full billing, mandate an accessible cancellation method, and require descriptors that identify the merchant recognisably. Reason codes for cancelled recurring transactions and for services not provided as described are the operative routes, and compelling evidence requirements have tightened in ways that put the burden on the merchant to demonstrate ongoing consent and notification, not merely initial authorisation. Where the merchant cannot evidence the reminder or the accessible cancellation route, the outcome is increasingly predictable.

Open banking and variable recurring payment rails introduce a parallel structure that is, at present, better designed. A mandate with a defined cap and an in-application revocation route addresses several of the failure modes described above by construction: the customer can see and cancel every mandate in one place, on infrastructure the payment provider controls. Adoption for retail subscriptions is growing, and firms should expect the comparison to be made explicitly by regulators and by customers.

What the Data Is Showing

TrustSphere's engagement data separates this dispute population into four groups with materially different economics and outcomes. The first and largest is genuine non-recognition: the charge is legitimate, the customer authorised it originally, and they simply do not recall. Merchant descriptor quality is the single dominant factor here, and it remains poor at a rate that surprises most acquirers: payment facilitator names, holding company names, unrelated brand names and non-obvious abbreviations, all technically compliant and practically useless. A material share of this group is resolvable by the issuer at first contact if the descriptor can be enriched into something the customer recognises, and unresolvable if it cannot.

The second is failed cancellation, and it is where the merchant behaviour question becomes decisive. The customer attempted to cancel and the attempt did not take effect: a retention flow that required a telephone call at restricted hours, a cancellation buried behind multiple confirmation steps, a request submitted and unacknowledged, or a cancellation processed after the renewal date because the merchant's notice period exceeded the customer's remaining term. In these cases the customer's evidence is typically weak and their account of events is typically accurate, which is a difficult combination for an adjudication process built on documents.

The third is the trial conversion, which produces the highest dispute rate per transaction of any recurring category. The pattern is well known: prominent free trial, low-friction sign-up, disclosure of the subsequent charge present but subordinated, no meaningful reminder before conversion, and a first charge that arrives at full price after the customer's attention has moved on. Where the merchant's own churn data shows a large proportion of trial converts who never use the service after conversion, that is not a marketing success and both regulators and acquirers are now reading it that way.

The fourth is the price-increase dispute, which has grown noticeably. A renewal at a materially higher price, notified by email to an address the customer no longer monitors, produces a charge the customer neither expects nor recognises even where they remember the subscription. Scheme rules and consumer regimes increasingly require affirmative treatment of material changes, and merchants relying on a notice-and-continue model are exposed.

Across all four the timing signature is consistent and useful: disputes concentrate sharply on renewals following a long dormant period, on first charges after trial conversion, and on the first renewal after a price change. Those are predictable events, known in advance to the merchant and, through recurring transaction data, inferable by the issuer.

Implications for Financial Institutions

The first implication is for issuers, and it concerns resolution before dispute. A large share of this population can be closed at first contact if the customer can be shown what the charge is: enriched merchant data, brand name, logo, the date the subscription started, the prior charge history, and a direct cancellation route. Issuers that have deployed enriched transaction data report substantial reductions in dispute volume in this category, and the economics are unusually favourable because the alternative is a case handled by a person. This is one of the few areas of dispute management where the customer-experience improvement and the cost reduction point in the same direction.

The second is subscription visibility as a product capability rather than a nice-to-have. A customer who can see every recurring commitment on their account in one place, with amounts, cadence, next charge date and a cancellation route, generates fewer disputes because they cancel deliberately rather than discovering by surprise. Where variable recurring payment rails are available, the mandate model provides this natively, and issuers should expect to be compared against it.

The third is for acquirers, and it is a portfolio risk point. Subscription and trial-based merchants require underwriting that treats the acquisition model as the risk: trial-to-paid conversion rates that are high but followed by immediate non-usage, cancellation routes that are materially harder than sign-up, descriptors that do not match the consumer-facing brand, and dispute ratios concentrated on renewal events. These are observable at onboarding and in monitoring, and they predict both dispute exposure and regulatory exposure, the latter increasingly carrying the greater cost, because an enforcement outcome that changes the merchant's billing model can remove the acquirer's revenue overnight.

The fourth is that dispute adjudication should be aligned to what the rules now require rather than to what they used to. Evidence of initial authorisation is no longer sufficient where the applicable rules require pre-renewal notification, accessible cancellation and affirmative trial conversion. Issuers should ensure their representment assessment explicitly tests for those elements, and acquirers should ensure merchants understand that a signed-up-once record will not carry a defence.

The fifth is a data point most institutions have and do not use. Recurring transaction identifiers make it possible to see a subscription's full history: when it started, what it has cost, whether the price has changed, and how long it has been dormant. That makes it possible to prompt the customer proactively (before the renewal that would have produced the dispute) with a simple notification of an upcoming charge on a long-dormant subscription. It is inexpensive, it is welcomed by customers, and it converts a dispute into a decision.

Conclusion

Subscription disputes are not a fraud problem and are not, mostly, a dishonesty problem. They are a consent-decay problem: a permission given once, at a moment of low attention, that continues to have financial effect long after the context that produced it has been forgotten. The regulatory response has correctly moved from asking whether consent was obtained to asking whether it was maintained (reminders, easy cancellation, affirmative conversion), and the scheme rules have followed.

For issuers the opportunity is to resolve most of this population before it becomes a case, through enriched merchant data, visible subscription inventories and proactive prompts ahead of predictable renewal events. For acquirers the exposure is that merchants whose economics depend on unremembered renewals now carry regulatory risk that can change their business model, not merely their dispute ratio. Both sides are better served by making the charge recognisable than by arguing afterwards about whether it was agreed.

Suggested Next Steps

• Deploy enriched merchant identification (brand name, logo, subscription start date and prior charge history) into digital banking and contact-centre workflows, and measure first-contact resolution as the primary metric for non-recognition disputes.

• Build a customer-facing subscription inventory showing all recurring commitments with amount, cadence, next charge date and a direct cancellation route, and benchmark it against the mandate visibility available on variable recurring payment rails.

• Use recurring transaction identifiers to prompt customers ahead of predictable dispute events: renewals after long dormancy, first charges after trial conversion, and first renewals after a price increase.

• Update acquiring underwriting and monitoring for subscription merchants to assess trial-conversion and post-conversion usage patterns, cancellation-path symmetry, descriptor recognisability and renewal-concentrated dispute ratios as indicators of both dispute and regulatory exposure.

Sources: Visa and Mastercard rules on recurring, instalment and trial-based transactions, including pre-notification, cancellation access and descriptor requirements; UK Digital Markets, Competition and Consumers Act subscription contract provisions; European Union consumer rights and unfair commercial practices frameworks; US Federal Trade Commission enforcement on negative-option and automatic renewal practices; Financial Conduct Authority Consumer Duty; Open Banking variable recurring payments standards; UK Finance card dispute reporting; TrustSphere Risk Index, April 2026.

TrustSphere Risk Index Vendor Spotlight: Ethoca

Ethoca scores 6.6 out of 10 in the TrustSphere RiskTech Index 2026, in the Disputes and Post-Transaction category. The capability profile is deliberately narrow: Enterprise Fraud Risk Management 8, Fraud Detection 7, Client Lifecycle Orchestration 6, Transaction Monitoring and Screening 5, with Behavioural Biometrics 2, Watchlist Screening 2 and Document Authentication 2. The composite sits above the index mean on the strength of assessed depth in its category rather than any breadth.

The proposition has two halves and buyers frequently evaluate only the first. The alert half distributes confirmed fraud and dispute notifications from issuers to merchants in near real time, allowing a refund before a chargeback is raised. The digital-receipt half is the one that matters for this typology: enriched merchant identification (brand name, logo, purchase detail, subscription context and a link to the merchant's own cancellation route) surfaced inside the customer's banking application at the point where they are looking at a charge they do not recognise.

The fit against subscription disputes is close to exact, because the dominant sub-population is non-recognition rather than dishonesty. A recurring charge from a payment facilitator descriptor is unidentifiable; the same charge presented with the consumer-facing brand, the subscription start date and the previous charge history is usually resolved on sight. Institutions that have deployed enrichment report the effect concentrated precisely where the theory predicts: in first-contact resolution of non-recognition cases, before any dispute record is created.

The second relevant property is that the alert channel gives merchants an opportunity to refund a genuine cancellation failure before the case enters the scheme process. In this typology a meaningful share of disputes arise from a cancellation that did not take effect, and the merchant is often willing to refund once it is visible; what is missing is a mechanism to tell them in time. That is a modest capability that removes a disproportionate amount of avoidable cost from both sides.

The limitations should be stated plainly. This is a network effect product and its value is a direct function of participation on both sides. Enrichment coverage depends on the merchant supplying the data; alert value depends on the merchant subscribing to receive it. Coverage in the specific merchant categories that generate the institution's dispute volume (streaming, software, gaming, fitness, digital media) is the only relevant coverage figure, and it should be tested against the institution's own top disputed merchants rather than accepted as a global number.

Second, enrichment resolves recognition and does not touch consent. Where the underlying issue is a trial that converted without meaningful notice, or a price increase notified to a dormant email address, showing the customer a clear brand logo does not make the charge legitimate; it merely makes the dispute better informed. Firms should not expect enrichment to reduce the disputes that arise from genuine merchant conduct problems, and should be careful that improved recognition metrics are not read as improved outcomes.

Third, integration into digital banking is a front-end programme, not a data feed, and it competes for the same scarce delivery capacity as every other application change. Institutions that treat it as a fraud-team procurement and not as a digital roadmap item tend to buy it and then not ship it.

The questions to press are: what enrichment coverage do you have across our top hundred disputed merchants by volume, specifically in subscription categories; what proportion of our recurring transaction population would return a logo and a full purchase detail; what is the measured first-contact resolution uplift at comparable institutions; can subscription context (start date, charge history, next expected charge) be surfaced and not only the current transaction; and what is the realistic front-end integration effort and reference timeline?

The verdict is that Ethoca's 6.6 fairly reflects a narrow product that addresses the largest single component of this dispute category directly and cheaply. It resolves the charge nobody recognises. It does not resolve the charge nobody meaningfully agreed to, and buyers should size the benefit against the right half of the population.

TrustSphere Risk Index Vendor Spotlight: Verifi

Verifi scores 6.4 out of 10 in the TrustSphere RiskTech Index 2026, in the Disputes and Post-Transaction category. The capability profile mirrors its category peer: Enterprise Fraud Risk Management 8, Fraud Detection 7, Client Lifecycle Orchestration 7, Transaction Monitoring and Screening 5, with the same expected floor across identity and screening capabilities. The index treats it as market-dependent, reflecting that value is determined by network participation within a given card ecosystem.

The proposition centres on pre-dispute resolution and case collaboration: order-detail exchange that allows an issuer to present merchant and purchase information during the customer conversation, and rapid-resolution mechanisms that allow a merchant to refund or provide evidence before a formal chargeback is created. The strategic idea is that most disputes are information failures rather than genuine disagreements, and that resolving them before the scheme process begins is cheaper for everyone including the customer.

The relevance to subscription disputes is in the structure of the evidence question. This category has moved from being adjudicated on initial authorisation to being adjudicated on ongoing merchant conduct: was a pre-renewal reminder sent, was cancellation as easy as sign-up, was trial conversion affirmatively consented. A collaboration mechanism that can carry that evidence into the issuer's decision at the pre-dispute stage is materially more useful than one that can only carry proof of the original sign-up, and buyers should test explicitly whether the evidence model accommodates the newer requirements or still assumes the older ones.

The second relevant property is deflection of the resolvable population. Where the customer's complaint is that they cancelled and were charged anyway, and the merchant's record shows a cancellation request that failed, the correct outcome is a refund within hours rather than a contested case over weeks. Pre-dispute collaboration is the only mechanism in the card estate that produces that outcome, and in this typology the resolvable share is unusually high because relatively few of these disputes involve any dishonesty on either side.

The limitations are the familiar ones for this category and they are not minor. Participation is everything: a merchant that is not enrolled cannot collaborate, and enrolment skews towards large merchants with mature dispute operations. Long-tail subscription merchants, which generate a disproportionate share of trial-conversion disputes because their acquisition models depend on them, are the least likely to participate, and they are precisely the population the institution most needs covered.

Second, there is a real risk of the mechanism being used to suppress rather than resolve. A merchant that refunds pre-dispute avoids a chargeback record, and a merchant that does so systematically while continuing the underlying practice keeps its dispute ratio clean and its conduct unchanged. Acquirers should monitor pre-dispute refund volumes as a risk indicator in their own right rather than treating a falling chargeback ratio as evidence of improvement, because in this category the two can move in opposite directions.

Third, operational integration determines whether any of this is realised. The mechanism only works if it sits inside the issuer's dispute workflow at the moment the customer is on the line, and if the response window is short enough to matter. Institutions that bolt it on as a separate portal for a specialist team capture a fraction of the available benefit.

The questions to press are: what is enrolment coverage across small and mid-sized subscription merchants specifically, as distinct from overall merchant count; can the evidence model carry pre-renewal notification, cancellation-path and trial-conversion consent records rather than only original authorisation; what is the median merchant response time and what proportion respond within the customer conversation window; how are systematic pre-dispute refunds without conduct change surfaced to acquirers; and how does the mechanism integrate into a first-line contact-centre workflow rather than a back-office queue?

The verdict is that Verifi's 6.4 reflects a useful mechanism whose value is set by participation and by workflow placement rather than by technology. In this dispute category it addresses the failed-cancellation population, which is the group where both parties usually agree once the facts are visible. Paired with Ethoca's enrichment covering the non-recognition population, the two address the two largest components. The third component, trial conversions that were never meaningfully consented to, is a merchant conduct problem, and no dispute tool resolves it. That one belongs to underwriting.

TrustSphere helps financial institutions design and deploy intelligent fraud and financial crime detection solutions. Visit www.trustsphere.ai