Flux & Variance

Commentary written against a balance that then moved

You explained a $50,000 variance. Then a late entry landed and the account is now $759,000 off. The commentary you already wrote didn't get worse. It just stopped being complete.

This happens near the end of every close, and it happens to the accounts you'd already finished. A journal entry that should have posted earlier finally lands — an adjustment someone was waiting on, a correction, something that surfaces in the last day or two when there's the least time left to deal with it.

The account you explained yesterday isn't wrong today. It's incomplete, and nothing about the file tells you that on its own.

What actually has to happen

The instinct is to treat this as a re-review — open the account back up, read the commentary again, decide if it still sounds right. That's not quite it, and it's not what actually gets checked.

What matters is whether the drivers you already wrote still cover the account's current variance. If you explained $50,000 and the account is now off by $759,000 after the entry, the question isn't whether your original explanation was good. It's whether there's now a $709,000 gap between what you explained and what the account actually shows.

The residual test doesn't change because the entry was late. It just has to be run a second time, against a number that moved out from under commentary that was already finished.

Practically, that means: pull the variance again, subtract what you've already explained, and see what's left. If the remaining amount doesn't clear your threshold, you're done and the account can stand as written. If it does, something has to be added — the late entry itself, most likely, named as its own driver with its own source, not folded silently into the existing explanation.

Who actually catches this

In practice this tends to be informal rather than a defined step. The preparer and whoever reviews the account go back and forth as late entries land, and the check happens as a conversation — not a re-run of the whole file, but a spot check: does this account still make sense given what just posted.

There's usually a second layer on top of that. Whoever reviews the preparer's explanations is also, at least implicitly, re-checking that the spot check happened and held. Two people converging on the same informal question, under the least slack in the schedule, because the last day of close is exactly when this comes up.

That process mostly works. It's also mostly invisible — there's no artifact that shows it happened, which means when it doesn't work, the person most likely to notice isn't the preparer or the reviewer who already agreed the account looked fine. It's someone outside that conversation entirely: a different reviewer, an auditor, someone reading the package cold who doesn't share the context that made the spot check feel sufficient at the time.

What actually needs re-checking, and what doesn't

Not every late entry threatens every account it touches equally. A few patterns are worth knowing before you start re-opening things.

An isolated entry is cheap to check. If it hits one account and nothing downstream of it, the residual test on that single account is the whole job.

An entry that feeds an allocation or a calculation is not isolated. An adjustment to something that flows into a cost allocation, an accrual calculation, or any account whose balance is derived from another one can move several accounts at once, not just the one the entry was booked to. The instinct to check only the account the entry touched misses this.

A sign or classification change is worth a second look regardless of size. The same structural checks that apply during the original close — whether an account's relationship to something else still makes sense — apply again here, because a late entry can just as easily break a relationship as change a balance.

Where this could actually be faster

The part of this that's genuinely slow isn't the judgment. It's finding out, quickly, which already-written commentary lines are even affected by a given late entry — scanning back through everything you finished to remember which accounts the entry touches, directly or through something it feeds.

That's a compilation problem, not a judgment one: given an entry, which accounts does it hit, and which of those already have commentary written against them. A tool can do that reliably, because it's retrieval, not interpretation. What it can't do is decide whether the resulting gap is worth explaining or whether the new driver is the right one — that's still the conversation between preparer and reviewer, just handed a shorter list to work from instead of the whole file.

The short version

A late entry doesn't invalidate commentary that was already correct. It changes the number the commentary has to cover. The residual test still applies — it just has to run twice, and the second time is the one nobody remembers to check on purpose.

Most of this surfaces again at flux review. The free flux template flags which balances moved enough to explain. The CloseOps Flux & Variance System ($79) starts from your trial balance instead: confirm each account's classification and it builds the income statement and balance sheet flux statements, then ranks what to investigate.

Found something wrong or missing? Tell me here — anonymous, thirty seconds.