How duplicate payments actually happen
Duplicates rarely come from one big mistake. They come from small gaps between people, systems and suppliers.
Most duplicate payments are not fraud. They are the result of ordinary process gaps that line up on the same day.
The usual routes
The same invoice, keyed twice. A supplier emails a PDF and posts a paper copy. Two people process it, or one person processes it a week apart. If the invoice number is typed slightly differently — a missing leading zero, a slash instead of a dash — many systems do not see the match.
A supplier on the master file twice. "ABC Ltd" and "ABC Limited" are set up as separate vendors. Invoices go to both, and duplicate checks that only look within one vendor record miss them.
Statements treated as invoices. A supplier chases with a statement; it gets paid as if it were a new invoice.
Credit notes that never get applied. The overpayment is recognised, but the credit sits unused on the account.
What stops them
Duplicate controls work best when they look across vendors, normalise invoice references, and check amounts and dates within a window rather than requiring exact matches. Just as important is a clean vendor master: one supplier, one record, verified against Companies House.
Checking your own ledger
You can run the free Integrity Scan on a payments export to see which of these patterns appear in your data. Every finding shows the rule that triggered it, so you can check it against the original invoice.