An owner can generally tell you what the business made last year. Ask which kinds of job produced it and the answer becomes an impression rather than a number, usually flattering toward whichever work is most enjoyable.
Job costing fixes that, and it does not require software or an accountant. It requires attaching four numbers to every job as it happens, and then reading them after a season.
What to record on each job
- What you charged. The invoiced total, not the estimate.
- Labor hours actually spent, including drive time, the supply run, and the two hours spent on the phone with the customer beforehand.
- Materials, at what you paid, including the fittings you already had on the truck.
- Anything subcontracted or rented for that job specifically.
That is the whole record. Four numbers on the back of the work order, entered into a spreadsheet once a week during the same fifteen minutes you use for receipts.
The critical instruction is to record hours honestly rather than as estimated. The gap between the hours you quoted and the hours you spent is exactly the information you are trying to surface, and rounding it away defeats the exercise.
The costs people forget to attach
Three of them, and they are the reason a job that looks profitable is not.
Drive time first. A job forty minutes away consumes more than an hour of the day before anybody picks up a tool, and for small jobs the driving can exceed the work. Owners routinely price by the task and pay for the distance.
Return trips second. The second visit to fix something, collect a check, or bring the part you did not have is a real cost and it is almost never recorded, because it does not feel like part of the job.
The unbilled front end third. Estimating, site visits, phone calls, and the emails answered at nine at night. That time is spread across the jobs you win, so a customer who needed four visits to decide has consumed a day whether they hired you or not.
What to do with three months of it
Sort the list by what you charged divided by the hours you actually spent. That single column is the whole analysis and it will not match your instincts.
Then group the jobs by type. Service calls, small repairs, larger installs, whatever the natural categories are in your work. Compare the groups. You are looking for the kind of job that shows up frequently and lands near the bottom of that column.
The two patterns it usually reveals
The first is that the small jobs are losing money, or close to it. This is nearly universal in trades and it happens because the fixed portion of a job, the driving, the setup, the paperwork, and the invoice, is the same whether the work takes an hour or six. Priced by the hour, a one hour job is being sold at a loss before anybody starts.
The fix is a minimum charge or a trip fee, set high enough to cover the fixed portion. Owners resist this because it feels like penalizing small customers. It is more accurate to say the small customers were being subsidized by the large ones, and the large ones were paying for it.
The second pattern is a customer rather than a job type. Somewhere in the list there is a name that appears repeatedly with hours far above the quoted figure. That is usually a customer who changes their mind, requires several visits, or is simply pleasant enough that nobody minds the extra hour. Now that it has a number attached, you can decide what to do about it, which is generally to price them differently rather than to lose them.
What it changes about estimating
The last and most useful step. Once you have a season of real hours, your next estimate stops being a guess and becomes a comparison. This is like that job, that job took nine hours rather than the six I quoted, so this one is nine.
That is the whole payoff, and it compounds. An operation that has been costing jobs for two years quotes more accurately than one that has been in business for twenty without recording anything, because the twenty year old business is still working from an impression while the two year old one has a record.