Visa RDR Rule Scoping: Stop Auto-Refunding Winnable Cases

On this page

What RDR actually does
The cost of a rule that is too broad
The rule-scoping checklist
What to keep outside the rules
The review rhythm that keeps rules honest
Frequently asked

Visa Rapid Dispute Resolution (RDR) is auto-refunding cases you could have won — not because the program is broken, but because it does exactly what your rules tell it to do. RDR checks each Visa pre-dispute against rules you wrote in advance, and every match is refunded with no human review and no second chance to fight it. The fix is not to switch RDR off; it is to narrow the rules until they cover only the cases where a refund is cheaper than a fight, and to leave every winnable case outside them.

What RDR actually does

RDR is a Visa pre-dispute service, run by Verifi, that resolves a dispute before it becomes a chargeback. When a cardholder dispute is submitted into Visa Resolve Online (VROL), the RDR decision engine checks it against seller-defined rules and attributes — such as dispute category, transaction amount, issuer BIN, currency and transaction date — and a qualifying dispute is resolved automatically, with funds moved from the seller's acquirer to the issuer and no manual review or action by the seller.

Two consequences follow, and both are the point of this guide. First, a case resolved through RDR never posts as a formal chargeback, so it does not add to your chargeback count — the reason RDR is popular with merchants approaching a monitoring threshold. Second, the judgment in an RDR case happened when you wrote the rule, not when the dispute arrived. There is no representment after an RDR refund. If your rule matches a friendly-fraud case where you held delivery proof, usage logs or prior undisputed transactions from the same customer, that evidence is never used, because the case never reaches a human.

The cost of a rule that is too broad

A rule that is too broad costs you the transaction value on cases you would have won, plus the refund itself. The typical failure mode is a rule written in a hurry during a volume spike — auto-refund everything under a round amount, or every dispute in a whole category — and then never revisited. Volume makes it worse: a broad rule fires silently, case by case, with no single moment that looks like a mistake, so the loss shows up only as a rising refund line that nobody connects to the rule.

Run the comparison before you widen any rule. On one side, the cost of fighting: the processor dispute fee, staff time to assemble evidence, and the risk of losing anyway. On the other, the cost of auto-refunding: the full transaction value, surrendered on every match, including the matches you would have won. The Fight-or-Refund Decision Calculator runs that comparison per case type, and the Chargeback Cost Calculator shows what a lost dispute really costs once product, shipping, fees and staff time are counted. If the fight side is smaller — and for any case with solid evidence it usually is — that case type does not belong in RDR.

The rule-scoping checklist

Scope RDR rules to low-dollar, high-friendly-fraud-risk cases only — and even then, only to the slice where you have no evidence worth using. Work through this checklist rule by rule:

Cap the amount low, and justify the cap. Set the amount ceiling at the point where assembling evidence costs more than the order value. A ceiling chosen because it is a round number is not a policy; a ceiling tied to your real cost per representment is.
Scope by dispute category, not by everything. Separate fraud categories from non-fraud categories, and write a rule per category. A rule that covers all categories at once is a rule you have not thought about.
Start with clear merchant errors and criminal fraud. A double charge you can see in your own records, or a fraud case with no usable evidence, is a refund whatever route it takes. These are the cases RDR exists for.
Exclude any category where your evidence wins. If delivery confirmation, signed receipt, usage logs or Compelling Evidence 3.0-style prior transactions regularly overturn a category, that category stays out — however small the amounts.
Use issuer BIN and transaction attributes sparingly. BIN or date conditions can carve out a genuinely unwinnable slice, but each extra condition is a rule that needs reviewing. Keep the rule set small enough that one person can explain every rule in it.
Write the exclusion list next to the rule. For every rule, record which case types it deliberately does not cover and why. The exclusion list is what stops the next person from "simplifying" the rules into a blanket refund.

What to keep outside the rules

Keep friendly-fraud cases with winnable evidence outside RDR, always. Friendly fraud — a customer disputing a purchase they made and received — is the category where representment most often pays, because the evidence already exists in your systems: tracking with delivery confirmation, download or usage records, prior undisputed transactions from the same account or device, and the customer's own communications. Auto-refunding those cases does not just lose one sale; it teaches the customer, and any ring watching, that disputing after delivery pays twice over — once in the goods, once in the refund.

Also keep high-value cases out regardless of category. Above your evidence-cost ceiling, every case deserves a human look, because the saving from one win outweighs the cost of reviewing many losses. And keep anything unusual out: a new product line, a new traffic source, or a sudden cluster of disputes in one category is a signal to investigate, not a pattern to auto-refund. RDR rules should cover the routine and the unwinnable; anything that looks like a new problem goes to a person first.

The review rhythm that keeps rules honest

Review your RDR rules monthly, against your own RDR-resolved cases, not against the rule document. Pull last month's auto-refunded cases and ask, for a sample of each rule: would we have won this with the evidence we held? If the answer is yes for a meaningful share of a rule's matches, narrow that rule — lower its ceiling, drop a category, or add an exclusion — and record the change. Then check the other direction: are disputes that should have been auto-refunded still reaching you as chargebacks? A rule that fires too rarely is also mis-scoped, and the fix is the same discipline in reverse.

Finally, check your position, not just your refund line. Run your monthly figures through the Chargeback Ratio Calculator and the Merchant Monitoring Program Checker: RDR keeps resolved cases out of your chargeback count, which helps a ratio, but a refund-heavy month still costs revenue, and a rising refund rate is itself a signal worth investigating. RDR is a valve for the cases you choose to lose cheaply. Scoped tightly, it protects your ratio at a fair price. Left broad, it sells your winnable cases for you.

Price the fight before you auto-refund it

Run the fight-or-refund comparison for each case type before it goes into an RDR rule.

Open the fight-or-refund calculator

Frequently asked

The short answers to the questions merchants ask most about scoping RDR rules.

RDR is a Visa pre-dispute service run by Verifi. When a cardholder dispute is submitted into Visa Resolve Online, the RDR decision engine checks it against rules the merchant set in advance, and a matching case is resolved with an automatic refund before it becomes a chargeback.

No. Once a dispute matches your RDR rules, the refund is automatic and there is no representment on that case. The only control is the rule itself, written before the dispute arrives.

No. Because it is resolved at the pre-dispute stage, it never posts as a formal chargeback. That is the ratio benefit of RDR — and also why a broad rule quietly gives away cases you could have won.

Cases where fighting costs more than the refund: low-dollar disputes, clear merchant errors, and criminal-fraud cases with no usable evidence. Friendly-fraud cases where you hold delivery, usage or prior-transaction evidence belong outside the rules.

No. RDR is a Visa program run through Verifi. Mastercard disputes need their own pre-dispute and alert arrangements through your processor.