Reading BIN intelligence to route smarter

Routing decisions are usually made at the level of the scheme and the currency: Visa in the EU goes here, Mastercard in the US goes there. That granularity leaves most of the available approval-rate difference on the table, because approval is decided by the issuer, and the issuer is identifiable from the first digits of the card.

What the BIN actually tells you

The issuer identification number resolves to more than a bank name. A well-maintained BIN table gives you the issuing institution, the issuing country, the product type — consumer debit, consumer credit, commercial, prepaid — and often the specific product tier. Each of those changes the economics and the likely outcome of an authorization.

• Issuer country vs. acquirer country determines whether the transaction is domestic or cross-border, which moves both interchange and the issuer's risk posture.
• Debit vs. credit changes which decline codes are likely — balance-driven declines cluster on debit.
• Commercial cards carry different interchange and frequently require Level 2/3 data to avoid a downgrade.
• Prepaid behaves like debit with lower balances and higher decline rates on larger amounts.

The same card, different acquirers, different answers

The observation that makes BIN-level routing worth building: an issuer's approval rate for a given BIN range is not constant across acquirers. Presented by one acquirer a transaction is approved; presented by another the identical amount, card, and MCC is declined.

The reasons are structural rather than mysterious. Acquirers differ in whether they have local acquiring in the cardholder's market, in the quality and completeness of the data they forward, in their fraud reputation with that issuer, and in how their MCC assignment matches the issuer's own rules. Those differences are stable enough to learn.

Building the matrix

The practical artefact is a success-rate matrix over issuer × acquirer × MCC × amount band, with enough volume per cell to be meaningful and a decay so that stale behaviour ages out. Cells with thin volume fall back to the issuer-level or country-level average rather than making a confident decision on six observations.

Two disciplines keep it honest. First, the matrix must be built on attempts, not on settled transactions, or it inherits a survivorship bias toward whatever the current routing already prefers. Second, some share of traffic has to be routed against the matrix's preference — otherwise the alternatives never accumulate data and the matrix slowly becomes a record of a decision it can no longer question.

Where it pays most

Cross-border traffic, where local acquiring options exist; markets with domestic routing schemes; and portfolios with wide issuer diversity. A merchant whose volume concentrates on a handful of issuers in one country will find less headroom here than one spread across dozens of markets — which is worth knowing before the work is scoped.