Every finance team has a story about a journal that should never have posted. The usual response is a reversal, a note, and a promise to be more careful. A better response is a step that makes the posting impossible in the first place.
Reversals record problems; approvals prevent them
A reversal is an honest correction, and there is nothing wrong with it. But a ledger full of reversals is a ledger that is documenting its own rework. Each one costs a posting, an explanation, and a small amount of trust in the numbers between the error and the fix.
What a configurable approver step actually costs
The objection is always speed. In practice, a well-scoped approval adds seconds to routine work, because most journals do not need approval at all. The value comes from scoping it narrowly.
Scope by risk, not by habit
- Approve journals above a value threshold.
- Approve anything touching equity, tax or intercompany accounts.
- Approve manual journals; leave system-generated postings alone.
- Approve entries posted to a closed or closing period.
Approving everything is how approval workflows get switched off within a quarter.
Make review meaningful
An approver who sees only a total is a rubber stamp. The review screen should show the full entry, the supporting document, and who prepared it. Segregation matters too: the preparer should not be able to approve their own work, including by proxy.
The measure that matters
Track the number of reversals per period before and after. If reversals fall and the close does not lengthen, the control is working. If reversals fall because people stopped posting anything at all until the last day, the scope is too wide — narrow it.