The month end close process accounting firms run for clients is not the one in the guides. You already know the checklist: reconcile the bank, accrue, post depreciation, run the reports. Every guide on the first page of Google says the same thing, and most are written for one company with one finance team.
That is not your problem. Your problem is 25 closes landing in the same week, and one senior person reading all of them.
This piece argues that the bottleneck in a multi-client close is review, not speed. Fix review and the calendar takes care of itself.
What the top-ranking guides cover (and skip)
Search the phrase and you get a familiar set: Ramp's step-by-step month-end close guide, Brex's seven-step checklist, and Xenett's ten-step guide for accounting firms. They are decent. They are also a list of tasks.
Three things are missing:
- The multi-client view. A checklist assumes one set of books. Firms have a portfolio, with different systems, different year-ends and different levels of client tidiness.
- Where the time actually goes. Guides list "review" as one step among many. For a firm, review is often the longest step and the least designed.
- When problems are found. Every guide treats the close as the moment you discover what happened last month.
The pressure is real, but speed is the wrong target
Brex cites a Sage Intacct poll in which 93% of accounting and finance professionals say they are under pressure to close the books faster. Nobody doubts that.
But "faster" is a blunt instruction. Bank feeds and auto-coding have already taken a lot of the posting work out of the close. If your controller still loses the last week of the month, the time is going somewhere else.
It is going into review.
Review is where multi-client closes stall
Xenett reports, from more than 1,000 firms using its software, that firms without a standardised review process take 40% longer to close each month. That is a vendor's figure from its own user base, so treat it as directional. But it matches what most practice managers will tell you over coffee.
Review stalls for three reasons:
- It is serial. One senior person reviews client after client, so the queue grows as the client count does.
- It is uniform. The tidy client with a clean ledger gets the same hour as the messy one.
- It is late. Review happens after the books are "done", so any error found means rework in the busiest week.
Here is an illustrative example (our numbers, not a benchmark). A controller reviews 20 clients at 90 minutes each. That is 30 hours, nearly a full working week. If 15 of those reviews find nothing, about 22 hours went on confirming that things were fine.
Move review from the ledger to the exception
The fix is not a faster reviewer. It is to change what gets reviewed.
Decide, per client, what "wrong" looks like. Then watch for it all month, not only at close.
Examples of exceptions worth defining:
- Gross margin moving more than a set number of points against the trailing average.
- A supplier invoice that looks like a duplicate.
- A customer paying well outside their usual days.
- A bank account with unreconciled items older than a set threshold.
- A balance sheet account that has not moved when it normally does.
- Payroll costs out of line with headcount.
When these are monitored continuously, the controller's review changes shape. Instead of reading the whole ledger to find the two things that matter, they open a short list of flags, each with the transaction behind it.
The close then confirms rather than discovers. By day three, most of the surprises have already been dealt with, usually by the bookkeeper, in week two.
How to set this up across a portfolio
You do not need to rebuild everything. Start small.
1. Tier your clients. Group them by complexity and risk, such as clean and simple, standard, and messy. Review depth follows the tier.
2. Write the exception rules once, then adjust per client. A default set of five or six rules covers most clients. Tune thresholds for the ones with seasonal or lumpy trading.
3. Give flags an owner and a clock. An exception nobody owns is just another report. Assign it to the bookkeeper on the account and set a mid-month deadline.
4. Standardise what review means. This is the part the 40% figure points at. Same accounts, same questions, same sign-off, every client, every month. Then a second person can pick up a review when the first is off sick.
5. Track review-found-nothing. Count how many reviews turned up no changes. That number tells you how much time exception-based review can give back.
What changes for the firm
The gains show up in three places.
- Capacity. If review hours drop, the same controller can cover more closes. We covered the arithmetic in how many clients one accountant can handle.
- Quality. Errors are caught while the transaction is fresh and the client can still remember it.
- Client conversations. A flag in week two ("your margin dropped four points, here is why") is advisory work. A correction in week five is an apology.
This is also where tooling earns its place. Pastel, for example, watches client books continuously and flags what needs attention, rather than waiting for the monthly close. It is built to cut the manual, repetitive work of bookkeepers, payroll and controllers by up to 80%. You can build a lighter version of the same discipline with spreadsheets and alerts in your accounting software, but it takes more upkeep as client numbers grow.
A realistic first step
Do not start with all clients. Pick five, ideally a mix of tidy and messy. Write exception rules for each. Run one month where review is flags first, ledger second.
Then compare: how long review took, what the flags caught, and what the full ledger read-through found that the flags missed. If the gap is small, you have your case for rolling it out.
The close checklist will always matter. But the firms that get their weeks back are not the ones with the shortest checklist. They are the ones that stopped treating the close as the moment of discovery.