Closing a Job
Closing is where a job's cost becomes final and where most of what could be learned is lost. The six-step close and why it takes five minutes.
Closing a job is the last moment its cost can be made true and the last moment anything can be learned from it. In most operations it is a financial act — the invoice went out, the job is closed — and the costing consequences are not considered at all.
The commercial issue in “Closing a Job” becomes easier to audit when estimated work, recorded hours and approved changes remain distinct. Teams considering view the provider’s detailed notes for how to monitor employees without being intrusive can strengthen the project-time evidence without treating a software report as a substitute for the quote, contract or accepted invoice.
A five-minute close turns a job from a billing record into evidence.
Additional context for “Closing a Job” is available from CPA Canada business resources. Current legal, payroll, tax and accounting requirements should be confirmed for the relevant jurisdiction.
The six steps
- Confirm the deliverable was accepted.
- Pull in any hours booked elsewhere that belong here.
- Record the scope changes that happened.
- Compare quoted with incurred, in hours.
- Write one line on the cause of the difference.
- Set the cost state: closed for billing, open for tail.
Five minutes. Steps five and six are the ones that get skipped.
Why step one matters
Because a job that is closed before acceptance will reopen, and the hours that arrive will have nowhere to go.
Acceptance is also the only unambiguous definition of finished that does not depend on anybody's judgement.
Why step two matters
Hours leak to standing codes and to adjacent jobs during delivery. Pulling them back at close is the only time anybody will remember where they belong.
A week later nobody can reconstruct it.
Why step five matters most
The one line on the cause is the entire learning. "Over by 13 hours: the client's data arrived in a format we had not seen, which we have now seen three times."
Three months of those lines, read together, is an estimating improvement nobody had to be told to make.
What the line should not be
"Underestimated." That is a restatement of the number, not a cause, and it produces no change.
Where the cause genuinely is not known, writing "cause not identified" is more useful, because a run of them is itself a finding.
The cost state
Two states, not one. Closed for billing on the invoice date; open for cost until the tail has run.
This is the single change that makes job costs true, and it is a policy decision rather than a system one in most cases.
Who closes
The person who ran the job, with the costing data in front of them. Not finance, who cannot supply the cause, and not automatically on invoice, which supplies nothing.
The jobs that never close
Standing jobs, support codes, anything open-ended. These need period closes instead: a monthly or quarterly line on what the period cost and what it was for.
Without them they run for years and nobody has ever looked at the total.
The close that finds a loss
Will happen, and the response decides whether closes keep happening honestly. A close that produces a loss and a blameless conversation produces another honest close next month.
One that produces a search for who approved the quote produces optimistic closes forever.
Batching closes
Weekly, in one sitting, with the data prepared. Ten jobs in an hour. Closing individually as they finish is theoretically better and practically does not happen.
What matters is that it happens within days rather than at quarter end.
The record it builds
One row per job: type, quoted, incurred, tail, scope changes, cause. Six fields.
After a year it is the most valuable document the operation has about its own work, and it was assembled five minutes at a time by people who were going to close the job anyway.
Closing a job nobody owns
Happens when the person who ran it has left or moved. The close produces no cause line and the job teaches nothing.
A nominated owner at the point of opening, carried through handovers, is the only reliable fix. It takes a field and it is the difference between a year of closes with causes and a year of closes without.
The close in a bad quarter
The first thing dropped when people are busy, and the thing whose absence is felt a year later when nobody can explain why estimates are poor.
Keeping it to five minutes is what makes it survivable; any close that takes half an hour will be skipped in exactly the quarter whose data would have been most informative.
What the cause lines look like after a year
Thirty or forty sentences, most of them specific and small. Read in one sitting, the repetitions are obvious in a way they never are one at a time.
That reading — an hour, once a year — is where the estimating basis actually changes, and it is the return on every five-minute close in the preceding twelve months.
What is lost without it
Everything this section was for. The hours were recorded, the invoice went out, the money was collected, and the organisation knows no more about its own work than it did before the job started.