The Balance Sheet Reconciliation Review Nobody on Your Team Owns

Every close team I've worked with can tell me exactly who reconciles the prepaid insurance account. Ask them who reviews that reconciliation before the books lock, and you get silence. The reconciliation gets done, a checkbox gets ticked, and the schedule moves on. Nobody looks at it again until an auditor pulls it in October and asks why there's a $47,000 unexplained variance sitting in the reconciling items column that's been rolling forward since March.

This is the gap in most close processes. We treat "reconciled" as if it means "correct." It doesn't. It means someone tied a balance to a support schedule. Whether that support schedule makes sense is a separate question, and it's the question nobody owns.

Why reconciliation without review fails

A reconciliation can be technically complete and still be wrong. I've seen all of these clear the close untouched:

• An accrued liability reconciled to a schedule that simply repeated last month's balance because the preparer copied the tab and forgot to update it.
• A fixed asset rollforward that balanced perfectly but capitalized $12,000 of repairs that should have been expensed.
• An intercompany account reconciled on both sides to numbers that didn't match each other by $8,300, because each preparer only reconciled their own leg.
• A bank rec with a "deposit in transit" that had been in transit for four months. It was a duplicate deposit posted twice.

None of these get caught by the person who prepared them. That's the entire point. The preparer built the schedule around their own understanding of the account. If that understanding is flawed, the reconciliation inherits the flaw. You need a second set of eyes that wasn't anchored to the same assumptions.

What a review step actually checks

The mistake teams make when they try to add review is turning it into a full re-performance. That doubles your work and blows up your timeline. A reviewer shouldn't rebuild the reconciliation. They should interrogate it. There's a difference.

The four questions that catch most problems

When I review a reconciliation, I'm asking:

• Does the balance move the way I'd expect? If depreciation expense was steady at $18,000 a month all year and this month it's $6,000, I don't care that the schedule balances. Something changed and I need to know what.
• What's sitting in reconciling items, and how old is it? A reconciling item is an admission that the GL and the support disagree. Anything older than 60 days is a red flag. Anything older than a quarter is usually an error nobody wants to own.
• Is the support real, or is it a rollforward of a rollforward? Ask when someone last agreed the balance to an actual third-party document, not just the prior period's tab.
• Does the balance pass a materiality-adjusted smell test? A $200 variance in petty cash isn't worth a conversation. A $200 variance in a revenue accrual might be a systemic cutoff problem worth $50,000 across all accounts.

Those four questions take about three minutes per account on clean recs and flag the ones that need real digging. You don't review everything at the same depth. You triage.

Building the review without adding days

The fear is always the same: "We already close in five business days, we can't add a review layer." You can, if you structure it right.

Tier your accounts by risk, not alphabetically

Not every account deserves the same scrutiny. I put accounts into three buckets:

• High-risk, always reviewed: accruals, reserves, intercompany, anything requiring judgment, and anything over a dollar threshold you set (say $100K). These get the full four-question treatment.
• Medium-risk, reviewed on rotation: stable balance sheet accounts that rarely move. Review a third of them each month so every account gets looked at quarterly.
• Low-risk, spot-checked: tiny, dormant accounts. A quick glance is enough.

This is where the tooling matters. In Excel, tiering means a manual tracker that someone updates by hand, rotation schedules that live in someone's head, and a review status that's impossible to see across the whole close. You end up with a workbook where the "reviewer" column is filled in with initials that mean nothing because there's no enforcement behind them. I've inherited close binders where every rec was "reviewed by JM" and JM had left the company eight months earlier.

Make review a gate, not a suggestion

A review step only works if the account can't be considered closed until it's approved. This is the structural argument for a close platform over spreadsheets. In a system built for month-end close, the reconciliation and its review are linked states: preparer submits, reviewer approves or rejects with a comment, and the account status reflects reality. You can see, at any moment, that 84 of 112 accounts are reconciled but only 61 are reviewed, and exactly which ones are stuck.

Excel gives you none of that visibility. The reconciliation lives in one file, the review notes in an email thread, and the close checklist in a third document that's already out of date. When the controller asks "are we ready to lock?" nobody can answer with confidence because the state of the close is scattered across a dozen artifacts. A platform makes the answer a dashboard instead of a fire drill.

Push reviewer comments back to the same place

When a reviewer rejects a rec because the deferred rent schedule doesn't tie, that comment needs to land where the preparer will actually see it and where it stays attached to the account for next month. In a close platform, the rejection, the reason, and the eventual resolution live on the reconciliation itself. Next month's preparer sees the history. In Excel, that context evaporates the moment the file gets copied into a new period folder.

What good looks like after a quarter

A team that adds a real review layer notices the change within two or three closes. Reconciling items stop aging because someone is asking about them every month. The audit gets quieter because the questions auditors would have asked already got asked internally, and got answered while the details were fresh. Preparers get sharper because they know their work will be read by someone who understands the account.

The measurable win: fewer post-close adjusting entries. If you're passing three or four correcting journals every month after the books "close," that's the cost of skipping review, paid in rework and eroded trust in your numbers.

Start with your ten highest-risk accounts next month. Assign a named reviewer, give them the four questions, and require sign-off before those accounts count as closed. Don't try to review everything at once. Prove the step catches something real, then expand the tiers. The first material error you catch before the auditors do will justify the whole exercise.