Sampling invoices when you cannot test every payment

When a client asks for payment verification across thousands of remittances, we rarely test the full population in a short engagement. Sampling is not a shortcut; it is a way to spend attention where risk concentrates.

How we choose a sample

• All payments above a threshold the client sets
• Every remittance tied to a bank-detail change in the period
• A random slice of mid-value invoices so ordinary work is not ignored
• Any vendor already on an internal watch list

What sampling cannot claim

A clean sample is not a certificate that every other payment is fine. We state coverage clearly in the findings pack so management does not over-read the result. If the population is small enough, full testing may cost less than designing a sample — we say so in the fee letter.

For supplier-side questions rather than remittance populations, see our vendor risk review.