Keep refunds linked to the original purchase evidence

Link the refund to the purchase evidence chain

Keep the two pieces of evidence connected so the amount and reason can be checked without searching the whole client folder.

If your evidence is a supplier invoice PDF, ParseMill can turn its invoice details and line items into a reviewable Excel workbook. Compare the source and correct the result before export. Receipt photos, missing-document requests and transaction matching belong in your existing receipt and accounting tools.

The relationship needs review before treatment is decided. A refund can be partial, delayed, combined with another adjustment or unrelated to the initially suspected purchase. The intake task is to assemble the chain and expose differences, leaving accounting decisions to the authorized practitioner.

Build a chain with distinct source roles

Use references for original purchase evidence, original payment, return or cancellation communication, refund document and observed incoming payment. Record which links are confirmed and which are candidates. A credit document and a bank refund can describe different stages, so do not treat one as proof that every stage occurred.

Intuit: Supplier credits and refunds explains that recording a supplier refund depends on the original purchase record. Intuit: Transaction attachments describes storing supporting files with eligible transactions. The proposed evidence chain helps the reviewer locate that history before choosing a treatment. It does not prescribe a journal entry or a universal refund procedure.

Compare a partial refund

In this fictional example, a EUR 200.00 purchase is followed by a EUR 75.00 refund. The relationship is documented, but the remaining purchase value and any further expected refund still require review.

Ask for the supplier's refund detail if needed to explain the relationship. Do not subtract EUR 75.00 and assume the remainder describes the correct accounting result. The arithmetic can help frame the question, while the reviewer determines the meaning of the underlying events.

Keep timing and completeness separate

Record the date a refund was promised, the date any refund document was issued and the date money was observed. If the expected payment has not arrived, keep that dependency open. If the refund amount differs from the document, create a focused difference question rather than overwriting one amount with the other.

• Preserve the original purchase and payment references.
• Link return or cancellation context.
• Distinguish promised refunds from observed payments.
• Show partial and unexplained amounts explicitly.
• Ask the reviewer to confirm the relationship and treatment.

If several purchases share one refund, create a reviewed allocation map rather than attaching the refund to an arbitrary transaction. If currencies differ, retain original units and the relevant conversion evidence. A plausible net amount should not hide an uncertain relationship.

Verify the reviewer can follow the chain

Before handoff, trace from refund to original purchase and back again. The record should show what changed, which sources support the change and what remains unresolved. A standalone refund query with no purchase context forces the next reviewer to repeat the investigation.

Use the related workflow guide for the wider evidence process and the workflow audit checklist to inspect recurring refund questions. A useful test case includes a partial refund and a promised refund that has not arrived, so the workflow preserves both evidence relationships and outstanding actions.