Reconcile failed SaaS payments through retries and recovery

One invoice can produce several gateway attempts and only one successful collection. Matching by amount alone can make a retry look like a duplicate sale.

A usable SaaS payment failures and retries file identifies the entity, period, currency, source systems, cutoff, preparer, and reviewer. Those details sound routine. They are also what let someone reopen the file next month without relying on memory.

Export every in-scope record before filtering. Save the report name, filters, extraction time, row count, and amount total. Compare that population with the prior open-item list and note any account, location, channel, or job that is missing.

Keep the source file untouched. Put corrections in the workpaper so the reviewer can see what changed and why.

Collect the invoice register, gateway event log, retry report, customer credit, processor payout, and receivable ledger. Assign each item a short reference and capture its source, date, amount, currency, status, and owner. Label a missing record as missing. A guess, an estimate, or an unapproved chat screenshot does not replace it.

Start with the last reviewed balance or status. Show additions, applications, releases, refunds, transfers, and corrections on separate lines. Tie the result to the current ledger or operating report. Gross movement often exposes two offsetting mistakes that a net total hides.

Use plain labels: missing record, timing difference, amount mismatch, duplicate, stale item, or judgment question. Add an owner, requested action, due date, age, and known financial effect. Close the item only when the stated evidence or approval arrives. If it remains open at cutoff, carry it forward visibly.

The preparer can compare observable fields, check arithmetic, and draft a focused question. The preparer should not decide customer access, write-offs, and revenue policy. Send that question to the billing operations owner with the evidence reference, the amount involved, and the response date needed. Save the answer in the workpaper rather than leaving it in a private message.

Reperform large entries, manual adjustments, old exceptions, and records whose status changed. Trace some items from source to schedule and others from schedule back to source. A repeated mismatch may point to a weak written procedure, a stale mapping, or a missing system feed. Record what you observed without guessing at the cause.

The final packet should include the untouched exports, reconciliation, evidence index, exception queue, approval record, and a short status note. State what tied, what did not, who owns the next action, and which decisions remain open. That is enough for the reviewer to continue the work without reconstructing it.