Reconciliations You Can Trust
The $300,000 difference that was not fraud
The largest reconciliation gap I have ever chased was a $300,000 difference between a trial balance and a bank statement. When a number that size shows up, everyone's first fear is fraud, and the room gets quiet fast. It was not fraud. When I finally traced it to the bottom, the gap was three ordinary process failures stacked on top of each other. First, timing differences that nobody had ever written down, so the same deposits in transit were re-explained from scratch every month, sometimes correctly and sometimes not. Second, activity posted to the wrong fund, which meant two accounts looked wrong instead of one and every fix in one place created a new question somewhere else. Third, a set of unresolved prior-period items that had been rolled forward month after month because no one owned the job of clearing them, and each roll made the pile look a little more normal.
No single one of those failures would have produced a six-figure difference. Compounding did. In about ten years of university finance operations, that is the pattern I saw over and over: the scary number is almost never one big mistake. It is several small, undocumented gaps that were each survivable on their own and were allowed to sit long enough to braid together.
The lesson I took from that $300,000 hunt was not "check the bank rec harder." It was that a reconciliation done as a monthly act of heroism will eventually fail, no matter how sharp the person doing it is. A reconciliation done as a repeatable system is boring, and boring is exactly what you want from this part of the business.
This is general educational information, not formal accounting, tax, or legal advice. I work non-attest, alongside your CPA, not in place of them.
What "reconciled" has to mean
Most small teams use "reconciled" to mean "matched." The software ticked off the items that agree, someone listed the difference at the bottom, the total tied, done. That is matching, and matching is the easy 90 percent. It tells you the two balances can be forced to agree. It does not tell you the balance is right.
Reconciled has to mean three things about every difference. It is identified: you know the specific items that make up the gap, not only the net amount. It is explained: each item has a reason attached, in writing, that says why it exists and when it should clear. And it is tracked by age: each open item carries the date it first appeared, so you can see at a glance whether it is a normal two-day timing lag or a problem that has been quietly celebrating birthdays.
A difference without an explanation is an open question. An explanation without an age is a hiding place. The prior-period items in my $300,000 story had explanations, of a sort. What they never had was a date, so nobody could see that "pending research" had meant the same thing for seven months.
The standard a reconciliation should meet
Here is the test I hold every reconciliation to: it has to survive a cold read. Hand the file to a colleague who has never touched the account, with no verbal walkthrough, and they can get from the ledger balance to the bank balance using only what is on the page. Every reconciling item has a description a stranger can understand, a dollar amount, a date, and a status.
If the reconciliation only makes sense with the preparer sitting next to it, it is not a control. It is a memory aid for one person, and it dies the day that person goes on vacation, changes jobs, or simply forgets which shorthand meant what. The cold-read standard is also what makes review real. A reviewer who has to ask the preparer what each line means is not reviewing anything; they are being given a tour.
Four small controls that make it boring
None of this requires new software. It requires four decisions, made once, calmly, outside the pressure of the close.
A fixed account list with owners. Write down every account that gets reconciled each month and give each one a single named owner. A team is not an owner; a person is. New accounts join the list the day they are opened, not the day they blow up. The first time you build this list, expect to find at least one account everybody assumed somebody else was covering.
A materiality threshold agreed in advance. Decide, before the month starts, what size of difference gets investigated to the penny and what gets noted and monitored. A team might say anything over $250 gets a full explanation and anything under it gets logged with a one-line note. The exact number matters less than when you pick it. A threshold chosen in advance is a policy; one chosen at 6 p.m. on close day is a rationalization.
An aging column for unresolved items. Every open item carries its first-appearance date, and the reconciliation shows it. Then set a tripwire: anything older than one full cycle gets escalated to whoever reviews the account, no exceptions. This one column is the direct antidote to the silent roll-forward that fed my $300,000 gap. Those prior-period items would have set off alarms at 60 days instead of compounding for seven months.
A second set of eyes on high-risk accounts. Not every account needs review, but cash, payroll clearing, and anything where money moves in or out of the organization does. The reviewer signs and dates the reconciliation, and the cold-read standard is what makes their signature mean something. On a three-person team this can be a fifteen-minute monthly task. It is the cheapest fraud and error control you will ever install.
Aim for boring
When these four controls are in place, reconciliation stops being a monthly scramble and becomes a checklist item that takes a predictable amount of time and produces a file anyone can defend. The measure of success is how uneventful it feels. A reconciliation that surprises you every month is telling you the process is broken upstream, and the surprises are compounding somewhere you cannot see yet, the way three small gaps once compounded into $300,000.
Start this month with the account list and the aging column. Those two alone will surface most of what has been rolling forward unexamined. If you pull that thread and find more open items than anyone expected, and you cannot tell which ones matter, that is the kind of untangling a Financial Operations Assessment is designed for. The review sorts out which accounts carry real risk, where the differences come from, and the order to clear them in.