Ask the controller of a services firm how long the month-end close takes and the answer is usually a number between seven and twelve working days, said with the tone of someone who has stopped expecting it to change. Ask what happens in those days and a pattern appears that has nothing to do with accounting. The first six days are not closing the books. They are finding out what happened during the month, because the month did not record itself.
The month-end close process in a services firm is slow for one reason: month-end is the first time anyone reconciles the time that was worked to the projects it was worked on, and the projects to the invoices that went out. Three facts that could have been recorded weekly are reconstructed monthly, by hand, by the people who are then blamed for how long it takes.
The close starts with timesheets, and timesheets start late. On the first working day of the new month, a share of last month's time is still unsubmitted, a larger share is submitted and unapproved, and some of it is on the wrong project because the consultant picked the first code that looked right. The finance team chases the stragglers, the project managers approve in a batch without reading, and the entries on the wrong code are found later, if at all, by the client who queries the invoice.
Three days, and the output is a time ledger that is now final enough to bill from. Nothing in those three days was accounting. It was the collection of a fact that existed on the day the work was done and was not written down until the month was over.
With the time final, billing can start, and this is where the second gap opens. Fixed-fee projects are billed on milestones, and the milestone was reached on the twelfth but nobody told finance, so it goes on this month's invoice three weeks late. Time-and-materials work is billed from the approved hours, and the hours on the wrong code are billed to the wrong client or not at all. Change orders that were agreed in an email in week two are discovered in week six of the next month when the project manager is asked why the project is over budget.
The firm has been doing work it did not bill, and month-end is when it finds out. We described the shape of this in the piece on what breaks between the proposal, the project and the invoice: the three records are kept in three places, and the close is the monthly attempt to make them agree.
Three more days, and the invoices go out. Cash that should have moved on the fifteenth moves on the sixth of the following month, which is the working-capital cost of the close that nobody puts on the same page as the finance team's hours.
Now the accounting begins. Revenue recognition for a services firm means deciding how much of each fixed-fee project is earned this month, and the usual tool is a spreadsheet: a row per project, percentage complete typed by the project manager, multiplied by the fee. The percentage is a judgment, the spreadsheet is a copy of last month's with the numbers changed, and the two or three projects that were closed early or extended late are the ones that get recognised twice or not at all.
Three days to produce a number that the leadership then compares with the utilisation number, which was produced from the same unreliable time ledger, and the two disagree, and a meeting is held about which one is right. We wrote about why that comparison fails in the utilisation you cannot see; the close is the same problem seen from the finance side.
Put the nine days together and the shape is clear. The length of a close is the number of facts that were not recorded when they happened, times the effort of recovering each one. Time not approved in the week it was worked. Milestones not marked on the day they were reached. Change orders agreed in email. Revenue percentages invented at the end rather than derived during. None of those is an accounting problem. Each is an operational fact that had a moment when it was true and easy to record, and the firm let the moment pass.
A three-day close is not achieved by a faster finance team or by close-management software that tracks the reconstruction more neatly. It is achieved by a month in which the facts were recorded, so that the close is a confirmation of what the systems already say.
First, weekly timesheet approval with a lock. Time is submitted by Friday, approved by the project manager by Monday noon, and locked. An unapproved week blocks the manager's own report, so approval happens. An entry on the wrong code is corrected in the week it was made, by someone who remembers the work. Day one to three of the close disappears, because the time ledger was final every Monday.
Second, a billable-but-unbilled report, read every Monday by one named person. It lists approved time not yet on an invoice, milestones marked reached and not yet billed, and change orders recorded against a project with no corresponding fee line. Each row is a decision: bill it, or write down why not. Days four to six disappear, and the cash moves in the month the work was done.
Third, revenue recognition computed from the records, not typed into a spreadsheet. For time-and-materials work, earned revenue is approved hours times rate, which the system already holds. For fixed-fee work, percentage complete is derived from approved hours against the budgeted hours, or from milestones reached, and the project manager overrides it with a reason only where the derivation is wrong. Days seven to nine become a report that was available every week of the month.
The three checks only work when time, projects and invoices are in systems that share the project as a key, which is the design question underneath all of this. In the Zoho suite, time logged in Zoho Projects against a task carries the project and the billable flag, the approval and the lock are configuration, and an invoice in Zoho Books can be raised from approved time or from a milestone without re-keying. The billable-but-unbilled report is a join across the two, and the revenue computation is a report over approved time and milestones. The work is the three checks, the weekly rhythm, and the one person who reads the report.
A firm on a timesheet tool, a separate project tool and an accounting system that do not share the project key will build the three checks as a monthly spreadsheet, which is the nine-day close with a nicer name.
The close was never slow. The month was silent. Discovery is no-risk: we read your last three closes and show you which days were reconstruction, and you pay only if you go ahead.
In a no-risk discovery we look at your CRM, finance and operations systems and who owns each part, and show what a connected system would do differently. You pay only if you proceed. Or see how we approach it.
More on the same problem:
"Something isn't working." "We're outgrowing our tools." "We're afraid to make the wrong move." No-Risk Discovery is a short, practical conversation that gets you clarity before you commit to anything big. We'll tell you if we're a fit. If we're not, we'll tell you that too.
Say who you are and ask. A person replies within one business day.
