You tap your crypto card at a coffee shop abroad, the terminal beeps, and you walk out with your latte. Behind the scenes, your card’s issuer was asked, “Is this okay?” — and for a moment, nobody answered.
That silence is not a decline. In card networks, a decline is a deliberate “no.” A timeout means something broke: the issuer’s servers are overloaded, a release went wrong, or a network link flapped. The network can’t wait, so it has two choices: decline everyone, or let someone else decide. Most of the time it chooses the second. That “someone else” is the card scheme, and the mechanism is Stand-In Processing, or STIP.
If you’re comparing crypto cards, understanding STIP and its cleanup partner — the reversal message — explains a lot. For the normal flow, see inside card authorization; for the timeout scenario, try the auth flow demo.
What STIP does when the issuer goes quiet
STIP is the scheme’s backup brain. When the issuer doesn’t respond within the timeout window — usually a few hundred milliseconds — the scheme steps in and approves or declines the transaction on the issuer’s behalf, using rules the issuer configured in advance.
Why allow this? Because availability is the whole point of a card network. One failed tap and most people reach for a different card. STIP decouples the issuer’s technical problems from your checkout: the issuer may be down, but your latte still gets paid for.
The catch is that the scheme doesn’t know your real available balance, your latest account status, or your issuer’s full fraud model. So STIP is a controlled risk trade: the network stays usable, but the issuer accepts bounded exposure. To keep that exposure tight, STIP uses layered rules — card-status lists, single-transaction limits, per-card cumulative limits, and a lighter fraud screen.
The cleanup: advice replay and weird balances
STIP is only the first half. Every transaction the scheme approved while the issuer was down must be posted back to the issuer’s books once it recovers. The messages that do this are called advice messages. Unlike an authorization request, which asks for a decision, advice says, “This already happened, please record it.”
This replay step is where things get messy. Advice can arrive minutes or hours after recovery, so the “when you tapped” and “when it posted” timestamps can fall on different clearing days. If advice gets retried and the issuer doesn’t handle it idempotently, the same transaction can post twice or create a duplicate hold.
For crypto card users, this matters because crypto-backed cards often rely on partner banks or program managers as the issuer. The more hops between the card scheme and the ledger holding your USDT or USDC, the more places advice replay can drift.
Reversals: the message that gives your limit back
Authorizations don’t just spend money — they hold part of your limit. That hold is released when the final clearing arrives. But many authorizations never clear as-is: a cashier voids the wrong amount, a hotel pre-auth is larger than the final bill, or a network timeout means the terminal never got the approval response. In each case, the hold must be released or it will sit there, shrinking your available balance.
The release message is the reversal. In ISO 8583 terms — the message standard most card networks trace back to — a common reversal is the 0400 message. It tells the issuer: “Cancel that authorization and give the held limit back.”
There are a few everyday flavors. A timeout reversal happens when the terminal sends a request and hears nothing back; not knowing whether the issuer approved, it proactively backs out any hold that might exist. This is the mirror image of STIP: silence can trigger a stand-in approval on one side and a timeout reversal on the other. A partial reversal releases only the unused portion — fuel pumps and hotels are classic examples. A full reversal cancels the whole thing.
Dangling holds: when the release message gets lost
Reversals travel the same network as everything else, so they can also be lost, delayed, or processed twice. When a reversal fails to reach the issuer, the original hold stays in place even though the transaction will never settle. Your available spending power quietly shrinks, and there may be no obvious purchase to explain it.
The fixes are simple in concept: retry lost reversals reliably, process them idempotently, and sweep for old holds that never got a matching clearing.
What this means if you’re using a crypto card
Most users will never know STIP saved their tap. But the exceptions show up: a declined purchase even though you have funds may be a cumulative stand-in limit during an issuer outage; a balance drop after a hotel or fuel pump can be a lost reversal; and a charge that ghosts in and out of your list is often advice replay after a partial outage.
If you’re choosing a crypto card, ask how the issuer handles holds and reversals — not just fees and rewards. Our card comparison tool and fee calculator can help model costs, but operational reliability is harder to see on a features table. Cards with clearer hold-release policies tend to feel smoother in real use.
The bottom line
STIP and reversals are the unsung plumbing of card payments. STIP keeps your card working when the issuer stumbles; reversals make sure authorization holds don’t become permanent. Both depend on three ideas: respond in tiers by severity, detect duplicates within a time window, and process retries idempotently. When any of those breaks, the symptom is usually a balance that doesn’t quite match what you remember tapping.
For more on card authorization in crypto-linked payments, see our payment auth wiki. And if you want to see the timeout-and-reversal dance yourself, the auth flow demo lets you step through what happens when the issuer doesn’t answer.