Bookkeeping Cleanup

Almost every failed cleanup fails for the same reason: the steps were done in the order they were noticed rather than the order they depend on. Use this as your bookkeeping cleanup checklist, because each step assumes the ones above it are finished.

The order is load bearing at three points in particular. Duplicates have to go before anything is reconciled, or you will reconcile against a total that is wrong. Opening balances have to be established before the first period is closed, or every period after it inherits the error. And the periods have to be locked as they are finished, or a later edit silently reopens work you have already paid for.

Books that need cleaning usually have all of these at once, and they compound. A duplicate in January makes February's reconciliation fail, which makes March's opening balance wrong, which means the error you are actually chasing in June was created five months earlier by something unrelated to June.

The order is not a preference. Each period's opening balance is the previous period's closing balance, so a reconciliation attempted out of order fails against a number that has not been fixed yet. Starting at the most recent month feels faster because it is the month you care about, and it is the usual reason a cleanup stalls halfway through with more open questions than it started with.

Meinry pulls the full history from the bank feed rather than trusting the existing ledger, because the ledger is the thing that is wrong. Each transaction is categorised against what the business actually did, with the counterparty, the amount and the surrounding pattern as evidence. Duplicates are matched and collapsed. Then each account is reconciled month by month, oldest first, and where a period will not reconcile it says which period and by how much rather than carrying the difference forward into the next one.

Some of it genuinely cannot be recovered from a bank feed. A $4,000 payment to a vendor you used once. A deposit with no invoice behind it. A transfer that looks like income on one side and an expense on the other. A payment to a person who might be a contractor, might be a refund, and might be you.

Those come to you as a list with the amount, the date, the counterparty and what the evidence suggests, and nothing is decided on your behalf. Everything else is already done before you are asked anything, which is the difference between a cleanup that takes an afternoon of your time and one that takes a fortnight of it.

Cleanup is traditionally billed hourly, commonly $50 to $100 an hour, because a firm cannot price it until they have looked, and looking is a large part of the work. That is why every cleanup quote starts with a discovery call, and why the number you are given depends on what the person on that call thinks they saw.

The range that results is wide enough to be almost useless as a guide. A few hundred dollars for a quarter of a low-volume business, several thousand for a year of a busy one, and the difference is transaction count rather than anything you would notice from the outside. Two businesses with the same revenue can be a factor of ten apart.

Three things make a quote go up more than people expect. Accounts that were never connected, because each one is a separate reconciliation and its history has to be found. Missing documentation, because a transaction with no evidence becomes a question rather than a decision. And a prior period that was already filed on the old numbers, because now the correction has consequences beyond the ledger and somebody has to decide what to do about them.

Meinry does the looking first. Connect the accounts and it works the backlog, so you see the actual scope, meaning how many months, how many transactions and how many exceptions, before you decide anything. The cleanup itself is on the ladder above rather than billed by the hour, because the work is volume work and volume is what the bands are drawn on.

Most books that need cleaning are in QuickBooks, so a few of the failure modes are particular to it rather than general.

Opening Balance Equity is the first thing to look at. It is a holding account QuickBooks creates during setup, and a balance sitting in it long after setup means something was entered without a counterpart. It should be zero on a clean set of books, and a figure there is usually the fastest way to find what went wrong at the beginning.

Undeposited Funds is the second. Payments recorded against invoices but never matched to the deposit that actually landed accumulate there, and the balance grows quietly while the bank feed shows the deposits arriving separately, which is one of the more common ways revenue ends up counted twice.

Then the bank rules. A rule written to save time on one vendor will happily miscategorise two years of transactions before anybody notices, and rules are usually the reason a set of books is wrong in a consistent, systematic way rather than a random one. Reviewing the rules is often faster than reviewing the transactions they produced.

The last one is the chart of accounts itself. Accounts created by typing a name instead of selecting an existing one leave you with three variations of the same category, split across periods, so no report adds up to what you expect. Merging them is straightforward and has to happen before the categorising, not after.

Most of the delay in a cleanup is not the work. It is waiting for access and for documents, and both are things you can have ready before anyone starts billing.

The access matters more than the paperwork. Read access to every bank and card account for the whole period, not just the current ones, and the same for every payment processor: a closed account nobody mentioned is the single most common reason a cleanup reopens a month later. Then the accounting file itself, with a note on who else has been editing it.

The documents matter in a narrower way than people assume. Nobody needs a receipt for every transaction. They need them for the ones that are ambiguous, large, or unusual, which is a small fraction of the total and which you cannot identify in advance. So the useful preparation is to know where the receipts are and be able to find one quickly, rather than to assemble them all first.

The last piece is the one people leave out and it is the most valuable: a short list of anything unusual that happened in the period. A loan taken, an asset bought, an owner drawing, a refund to a customer, a grant, a payment made from a personal account. Six lines of that saves more time than a folder of receipts, because every one of those items is otherwise a transaction that gets categorised wrongly and then has to be found again.

A cleanup that is not followed by maintenance is a cleanup you will pay for again, usually eighteen months later and for more. The reason books get into this state is almost never one bad month. It is that nothing was running between the months, and a cleanup does not change that by itself.

This is the honest argument for continuous bookkeeping over a periodic service: the state you are paying to fix is the state that a monthly-touch cadence produces by design.