Glossary · Payments

What is Buyer reference?

A buyer reference is the identifier a payer puts on a bank-rail transfer so the recipient can match it to an obligation: the UPI transaction reference, the SEPA end-to-end ID, the remittance information on a wire. On manual rails it is the reconciliation key: without it, a received payment is money without a meaning.

Crypto settlement reproduces the same idea differently: the transaction ID plus the checkout's address derivation does the matching automatically, which is why crypto legs do not need buyers to type references at all.

How Buyer reference works

A buyer reference is the unique identifier a payment carries so it can be matched to the obligation behind it: the invoice, the order, the subscription period. It rides the payment metadata wherever a field exists, and the payment record preserves it regardless, so reconciliation is a join rather than a hunt.

The problem it solves is the anonymous deposit. On-chain settlement delivers value from addresses that carry no memo, and without a reference, the ledger entry is money that arrived from nobody for nothing. The reference closes that gap: each checkout generates one per order, and the payment record binds it to the transaction identifier, amount, asset, and timing.

The design rule for references is uniqueness with meaning: opaque enough never to collide, structured enough that a human can eyeball what it belongs to. Systems that generate them per order remove the judgment call entirely, which is the point.

One subtlety worth naming: references belong to the obligation, not the payer. The same customer across ten orders carries ten references, and any customer-level identifier lives separately in the merchant's own records. Collapsing the two is how duplicate-payment confusion gets manufactured.

Buyer reference in practice

An invoice payment arrives with reference INV-1042. The accounting system finds the invoice by that key, sees the transaction hash beside it, and marks the pair settled without a human opening an explorer.

Buyer reference on U.CASH

Every bank-rail settlement on U.CASH produces the rail's native references in the order view, and crypto legs reconcile by transaction and address. One export covers both vocabularies.

This is the small field that decides whether payments reconciliation is automated or archaeological. U.CASH generates per-order references on every checkout surface, and the exports carry them next to the chain references, so the books join cleanly at both ends. On integration: put the reference in the webhook handler's ledger key rather than in a display field. References that live only on screens get retyped; references that are join keys get used. The export structure supports both, but the key is where the value is.

Related payments terms

Payment railFiat on-rampCheckout embed

See also

Payment railBatch payout

Buyer reference: FAQ

Why do bank rails need references when crypto does not?
Because a bank credit arrives as money plus a text field, while a crypto payment arrives as money plus a transaction on a public ledger. The ledger self-describes; the text field only says what the sender typed into it.
What if a payer forgets the reference?
Checkout-generated references attach automatically, so the forgetful-payer case mostly disappears. For manually shared details, the fallback is matching on amount and timing, which is exactly the archaeology references exist to avoid.
Can references be reused?
They should not be. Per-order uniqueness is what makes the join sound; standing identifiers belong to the customer, not the payment.
Where does the reference live afterward?
In the payment record beside the transaction hash and amounts, and in every export. The chain keeps the payment; the reference keeps the meaning.
Keep reading

The glossary, A to Z

Every term behind the rails, the assets, and the settlement.