Keeping chargeback ratios under control in high-risk e-commerce

For a high-risk merchant, the chargeback ratio is the number that matters most to the acquirer. It affects your reserve, your pricing, your ability to raise caps and, if it stays high for long enough, whether you keep the account at all. Understanding how the ratio works, and where disputes actually come from, is the first step to controlling it.

How chargeback ratios are measured

In broad terms, a chargeback ratio compares the number of disputes received in a period with the number of transactions processed. The details matter: networks and acquirers differ on whether they count by number or value, which month's transactions they compare against, and whether fraud reports are included. Ask your acquirer exactly how they calculate it for your account.

Card networks run monitoring programmes, such as Visa's VAMP and Mastercard's Excessive Chargeback Merchant (ECM) programme. Their thresholds and methodologies are revised from time to time, and programmes typically start flagging merchants around a 0.9% to 1% chargeback ratio, combined with a minimum number of disputes. Above that, merchants can face monthly fees, remediation plans and eventually termination. Acquirers usually set internal alarms below network thresholds, so the effective limit for your account is often lower than the published one. Always check current thresholds with your acquirer.

Where chargebacks really come from

Disputes fall into a few families, and each one needs a different fix:

• True fraud: a stolen card was used. This is addressed with fraud screening, velocity rules, 3DS2 and good device data.
• Friendly fraud: the cardholder made the purchase but disputes it, sometimes because they forgot, sometimes deliberately. Clear descriptors, confirmation emails and good evidence help.
• Unrecognised charges: the customer genuinely does not recognise the descriptor or the amount. This is common with subscriptions and with brands that sell under a different name from the company.
• Subscription disputes: the customer did not realise they signed up for recurring billing, or could not find how to cancel.
• Service disputes: the product did not arrive, arrived late or did not match the description.

In high-risk e-commerce, the non-fraud categories are often a large share of the total. That is good news: they are mostly within your control.

Prevention starts before the payment

Transparent offers

Make recurring terms, trial conversions and total prices impossible to miss. Display the recurring amount and frequency next to the payment button, not only in the terms. Send a confirmation email that restates them. For trials, send a reminder before the first full charge where network rules or local law require it, and consider doing so even when they do not.

Recognisable descriptors

The descriptor should match the brand name the customer saw at checkout, ideally with a support phone number or URL. If you operate several brands, use dynamic descriptors per brand where your acquirer allows it. Unrecognised charges are among the easiest disputes to eliminate.

Easy cancellation and refunds

A refund costs you the sale. A chargeback costs you the sale, a dispute fee, operational time and a hit on your ratio. A customer who cannot reach support or cannot cancel will go to their bank instead. Make cancellation simple, respond to support requests quickly and refund generously when a customer is clearly unhappy.

Fraud controls that do not destroy approvals

Fraud screening should be tuned to your actual fraud patterns. Useful controls include velocity limits per card, email, device and IP, blocking mismatched billing and IP countries for high-value orders, BIN-level rules for card types that show elevated fraud in your data, and selective 3DS2 for risky segments rather than on every transaction.

Be careful about over-tightening. Every legitimate order you block is lost revenue, and false declines rarely show up in dispute reports. Measure fraud rules by both the fraud they stop and the good orders they reject.

Pre-dispute alerts and order data sharing

Card networks and third parties offer services that notify merchants when a cardholder starts a dispute, before it becomes a formal chargeback, or that let issuers see order details when the cardholder queries a transaction. Refunding an alerted transaction quickly can prevent it from counting as a chargeback, and sharing order details can resolve confusion before a dispute is filed. Whether these alerts affect your ratio, and how, depends on the programme and the network, so confirm with your acquirer.

Fighting disputes you should win

Representment, contesting a chargeback with evidence, is worth doing for disputes where you have a strong case: delivery confirmations, proof of usage for digital products, logged acceptance of subscription terms, and prior undisputed transactions from the same customer. Winning a representment recovers the funds, though in most programmes the dispute still counts toward your ratio. Prevention is always more valuable than winning.

Monitor per account, not just in total

If you process across several MIDs, your blended ratio can hide a problem on one account. Monitor ratios per MID and per brand, and watch the trend week by week. When one account approaches its internal threshold, reduce its share of new volume while you fix the root cause, but never use multiple accounts to dilute or hide an excessive ratio. That practice is prohibited by card network rules and can lead to terminations and MATCH listings.

A simple operating routine

1. Review disputes weekly by reason code, brand, product, traffic source and MID.
2. Map each reason code family to an owner: fraud rules, offer clarity, support or fulfilment.
3. Check descriptors and confirmation emails whenever you launch a brand or offer.
4. Refund alerted transactions promptly where that prevents a chargeback.
5. Track ratios per account against your acquirer's internal thresholds, not only network ones.

LumniPay gives you the routing, risk controls and per-MID analytics to run this routine from one place: rule-based fraud screening, selective 3DS, volume shifting between accounts and a clear view of disputes by account, brand and BIN. The operational discipline, though, has to come from the team.