DEV Community

Swipe Again for Swipe Again

Posted on

"Do Not Honor" Is Not an Answer. It Is Missing Evidence.

Every decline code is a piece of evidence about why a payment failed — except one. "Do not honor" is the issuer declining to testify. The bank looked at the transaction, decided no, and recorded the refusal in a code that deliberately carries no reason. Teams that treat it as a diagnosis build their recovery on a shrug. Teams that treat it as missing evidence build an evidence file instead, and the file is where the recovery actually comes from.

Start with what the code rules out, because that is genuinely useful. A do-not-honor is not an expired card — that has its own code. It is not insufficient funds, not a lost or stolen card, not an invalid number. The credential presented was real, current, and attached to an account the bank recognises. Something about this transaction, at this moment, tripped a general refusal: an issuer risk rule, a velocity check, a cardholder restriction nobody documented, a fraud model having a cautious day. The evidence is negative — a list of things the failure is not — and negative evidence narrows the response more than most teams realise.

It rules out, first of all, the aggressive response. Retrying a do-not-honor quickly and repeatedly is the one move the evidence argues against: if an issuer's risk system made the refusal, a burst of identical attempts reads as exactly the pattern that system exists to catch, and each attempt can harden the refusal. The file-based response is slower and more deliberate. Space the attempts. Change something real between them — the day, the time, the context — rather than resubmitting the identical request and hoping the bank has forgotten. And run the customer-contact path in parallel from the first refusal, because with this code the subscriber is not a bystander to the fix; they are frequently the only party who can make it, by confirming the charge with their bank or using a different card.

That is the working method set out in a do not honor decline read as missing evidence: because the code tells you nothing, the case file has to tell you everything — the card's age in the vault, the subscription's tenure, the amount's history, whether this issuer has refused before and later relented. Patterns live in the file that the code refuses to provide. One merchant I know found their do-not-honors clustered on a single issuer and a single plan tier; the fix was not retry logic at all but a conversation with customers on that tier before renewal week. None of that is discoverable if the code is filed as a verdict and the case closed.

There is a discipline point here that generalises past one decline code. A recovery operation is an evidence operation. Every case arrives with some amount of signal — a precise code, a vague code, a silent customer, a responsive one — and the craft is matching the response to the signal actually received, not to the signal you wish you had. "Do not honor" is simply the extreme case: zero signal, honestly labelled. Handled as an evidence problem, a fair share of these cases still recover — spaced retries land once the issuer's moment passes, customers confirm charges, backup cards appear. Handled as a verdict, they recover at the rate luck provides.

A final entry for the file: record what worked. When a do-not-honor case recovers, note which move preceded the recovery — the spaced retry, the customer's bank confirmation, the backup card — because a code that explains nothing makes your own outcome history the only dataset you will ever get on it. Six months of honest case notes will teach a team more about this code than the code itself ever will.

File the shrug. Work the file. The bank will not explain itself, but the case, read whole, usually does.


Published by Swipe Again — failed payment recovery at a flat $179/mo.

Top comments (0)