Audit Stripe Radar fees and cap renewal screens

The fee changed shape in October

You open Stripe after a quiet month and see a Radar line you did not budget for. As of October 21, 2026, Stripe Radar applies a per-transaction fee to every supported payment method rather than limiting it to cards. Beginning that date, Stripe bills Radar screening for every transaction in a subscription, not only the initial payment.

Before October 21, 2026, billable screening covered only cards, ACH Direct Debit, and SEPA Direct Debit, while digital wallets, buy-now-pay-later, non-ACH/SEPA bank debits, and stablecoins were not billed. For Stripe Billing subscriptions, screening and its fee are always applied to the first payment and to the first payment made after a payment-method change. Custom rules that were originally written for card payments begin evaluating other payment methods on October 21, 2026. Users covered by the transition notice move to new pricing on January 22, 2027.

The renewal cap is the cash lever

Stripe Billing renewals are where the fee can hide. Stripe's Dashboard lets a merchant limit Radar screening to a chosen number of successful renewal cycles, and renewal cycles that are skipped are not billed.

A long-running subscription is not a blank check for permanent screening. Once a customer has paid without issue, the marginal fraud risk may be lower than the marginal fee. The cap keeps protection where it is cheap and stops paying for screens that mostly confirm a payment that is already working.

Capping is a policy. It should match the point where the fee stops buying useful protection. For a bootstrapped subscription, that point is often earlier than the habit of screening everything. Choose between a fee that earns its place and a fee that simply runs.

Start with a short cycle count and watch a renewal batch. When the fee line is small, extend the cap only when the data says the extra screen is worth it. When the line is already large, shorten the cap before the next batch. Changing the setting is cheap, which makes it a good early control.

The dispute math still matters

Do not treat the cap as a reason to stop screening. Some payment processors add a separate per-dispute fee to chargeback exposure. A cardholder has up to 120 days after posting to dispute a charge, and some cases allow up to 540 days. Subscription businesses see dispute rates around 1% to 2%, compared with 0.5% to 0.7% for one-time e-commerce.

The labor cost is the hidden burn. Dispute-response work runs 2 to 3 hours or 2 to 5 hours per case, equivalent to about $82 of internal labor per chargeback. Mastercard's 2025 State of Chargebacks counts 261 million chargebacks globally in 2025, projects 324 million by 2028, and shows 46% of chargebacks as absorbed by merchants without contest. In Baseline's closed disputes, charges that collected on the first try won 86% (137 of 160), while charges that needed retries won 34% (10 of 29).

Disputes and fraud are different, but both hit cash. A charge can be legitimate and still become a dispute. The fee audit should sit next to the dispute log, not inside it.

The audit you can finish before lunch

Keep the audit short because the scope is narrow. You are auditing the fee line, not all fraud risk. Each item below is a check you can run before the next renewal.

• Export screened attempts by payment method and cycle.
• Compare volume before and after the October 21, 2026 change by method.
• Multiply screened attempts by the contracted rate.
• Set the renewal-cycle cap in Stripe's Dashboard.
• Review card-only rules for non-card methods.
• Recheck the fee line after a renewal batch.

The audit is a loop, not a one-time export. It starts with raw volume, isolates the October change, turns counts into cash, turns policy into a setting, catches non-card gaps, and confirms the line moved after a renewal batch.

Leave the cap where it is when the post-change fee line is small. Lower the cycle count when renewals push the line up.