In this article
- 1Your P&L is not job costing
- 2Two halves of the same gap
- 3Where the money actually leaks
- 4What has to be counted — and what isn't worth counting
- 5Where accounting ends and production begins
- 6Purchase orders: a budget first, then a check and balance
- 7Labor: knowing the number during the job
- 8Decisions to make before you start
- 9Start with your last three jobs
- 10Common questions
Say you closed this month at a number you can live with. Now name the three jobs that paid for it, and the two that ate it. Most restoration companies cannot — and it is not because the books are wrong. It is because nobody ever asked the books that question.
Accounting tells you whether the company made money. Job costing tells you which work made it. Those are two different questions, and the second one is the one that changes what you bid, who you send, and what you quietly stop taking.
Your P&L is not job costing
A profit and loss statement answers a company-level question, once a month, after the fact. It is a good document and you need it. It is just not built to tell you that the commercial or residential water loss you were proud of lost money on labor, or that the small self-pay jobs nobody wants to run are your best margin in the building.
Answering that needs cost attached to a specific job while the job is still moving. And there is not much room to absorb the ones that go wrong. Restoration & Remediation, reporting on the industry Cost of Doing Business data, put 52% of restoration companies below ten percent net income in 2024.
Source: Anthony Nelson, “You Cannot Manage What You Refuse to Compare”, Restoration & Remediation, August 2026, citing the Cost of Doing Business Survey co-published by the Restoration Industry Association and KnowHow.
Operating that thin, the jobs that lose money are not something you can average away, and you cannot correct a pattern you cannot see. If you cannot rank the jobs you just ran, you cannot tell whether a good month came from how you run work or from the weather.
A good month made of jobs you cannot rank is luck you cannot repeat.
Two halves of the same gap
Two things decide whether a job made money, and they fail in completely different ways. The first is the estimate — what you are getting paid, and whether the scope was prepared properly for the work in front of you. The second is adherence: did the hours, materials and subcontractors you actually spent stay inside what that estimate allowed for?
So when a job comes in thin, there are two candidate explanations and you need to be able to tell them apart. Either you did not stick to the estimate — or the job was never estimated properly in the first place. Those call for opposite fixes. One is a purchasing and scheduling problem. The other is an estimating problem, and tightening purchase orders will not touch it.
Most of what follows is about the cost side, because it is the more mechanical half and the one more often left unmeasured. But keep the first in view — job cost is how you find out which problem you actually have.
Worth deciding early
Before any of this: know what margin you are aiming for, by cost type. You might budget to make 50% on materials and 30% on subcontracted work, or the other way round depending on your market and how you buy. Those are examples, not recommendations — the specific number matters less than having one, because a budget without a target margin is a guess with a decimal point in it.
Where the money actually leaks
On the cost side, the leak usually sits between what you planned to spend on a job and what actually got spent on it. And that gap is almost never found on the job. It is found on a vendor statement at the end of the month.
An example, and it is only an example
You have a $4,000 line for a framing package and you want to make half of it, so you budget $2,000. A week later somebody goes to buy it, puts it on the account against a purchase order number, and the number written down is $1,800. Good so far.
The bill arrives at the end of the month against that same PO number. It is $2,500. Look closer and the framing package is there — along with Gatorade, beef jerky, a new tool bag, and a replacement ladder for the one somebody forgot.
You can call that stealing. Plenty of owners would, and it would be hard to argue with them — someone put personal items on the company account. But run the more charitable version, where every line on that bill was legitimate and genuinely for that job, and you still have the same problem: $700 of margin left the job without a decision being made, and nobody knew for weeks.
We use that example on software walkthroughs, and it reliably stops the room. Recently an owner evaluating the system for his own restoration company interrupted to say that is exactly what happens at his shop. It gets that reaction because everyone has a version of it.
Job costing in QuickBooks works, and most companies are not even doing that much. But doing it properly there still requires purchase orders and time entries — accounting can only cost a job from records somebody else created. And underneath that sits a timing problem: payroll runs every two weeks, vendor bills post at month end. If job cost lives only in accounting, you learn what a job made two weeks to a month after you could have done anything about it.
What has to be counted — and what isn't worth counting
Job costing gets abandoned more often than it gets done badly, and usually because somebody tried to count everything. Three things carry almost all of the signal.
1
Labor — the biggest line, and the one most often remembered rather than recorded
On most mitigation work labor is the largest single cost, and it is the one companies reconstruct on Friday from memory. It has to come from people clocking in against the job they are actually on — a weekly guess spread across four jobs tells you nothing about any of them.
2
Materials — anything bought for one specific job
This is where the purchase order earns its keep. If a spend exists because of a specific job, it should carry that job's number from the moment somebody commits to it, not from the moment the bill gets reconciled.
3
Subcontractors — and permits, rentals, anything with an invoice attached to the job
Same discipline as materials, usually larger amounts, and often the difference between a reconstruction job working and not. Treat a sub's commitment like a purchase: budget it, approve it, attach it to the job.
4
Routine consumables — do not chase these
Nobody should be logging fifteen ounces of a cleaning chemical against a job. Allocate warehouse consumables as an overhead percentage instead — some companies assign a flat 10% and never think about it again. If a job genuinely eats an unusual quantity of product, cost it directly; set a threshold where that kicks in and otherwise leave it alone.
Worth remembering
Effort should match the value of the information. Direct purchases for a job deserve a purchase order. Routine consumables deserve a percentage. Decide where that line sits in your company deliberately, say it out loud, and then hold it — a rule everyone knows beats a stricter rule everyone quietly ignores.
Where accounting ends and production begins
This is not accounting versus production. It is the handoff between them, and the handoff is where it breaks.
Your accountant needs this information and cannot generate it. Every number that makes a job cost real is created on the production side — the estimate, the budget, what got purchased and by whom, hours worked, who managed the job. Accounting records it. If a technician clocked into the wrong job, or the framing package was bought without a job number, nobody downstream can repair that. They can only tidy it.
So job profitability has to happen on the production side, and it belongs on the job record where the revenue already lives. That is the reframe worth having. Not “we hired a good accountant, so job costing is handled” — but “the accounting is only ever as good as what production hands it.”
Purchase orders: a budget first, then a check and balance
A purchase order does two separate jobs, and most companies only think about the first one.
- It sets the budget before anyone spends. Even roughly, even before you know the vendor. You are recording an intention, and the intention is what makes an overrun visible later instead of invisible forever.
- It works as a check and balance on what actually gets bought. When the vendor bills against the PO number, you are not reconciling a pile of receipts against a memory — you are comparing one line to one line, and you can see whether what was purchased was in line with what was authorised.
- Set an approval threshold, and pick it honestly. Over a set amount, nothing gets bought without a project manager approving it — low enough to catch the ladder, high enough that nobody is asking permission to buy tape. You do not have to route every purchase through approval to get the benefit.
- You do not have to start with full purchase orders at all. If your team is not ready for a full approval and denial process, start by simply attributing receipts and purchases to the job they were for, then add a threshold on the purchases that actually move the number. Ease into it — a partial process that runs beats a complete one that gets abandoned in six weeks.
Labor: knowing the number during the job
Your payroll is your true labor cost, and it is accurate. It is also late. Knowing what a job cost in labor two weeks after the crew left is a historical fact; knowing it on day three is a decision you can still make.
That is the difference the production side makes. In Lever360, technicians clock in and out against the job rather than against the week, and you assign an hourly rate — the real one, or a rough one by position, whichever you will keep current. Labor accrues as the work happens, so gross profit moves while the job is open. Purchase orders, budgets and time land on the same job record, and the QuickBooks Online handoff creates a customer plus a sub-customer per job, so each job stays financially separate while rolling up under the customer.
Decisions to make before you start
Most job costing efforts stall on unmade decisions, not missing software. Settle these, write the answers down, and tell the team what they are.
Answer these first
Seven decisions, one afternoon
- Who owns job profitability? Name a role, not a department — and be specific about what the project manager is accountable for.
- What margin are you budgeting to, by cost type — materials, subs, labor?
- At what dollar amount does a purchase need approval, and whose approval is it?
- Which purchases get a full PO, and which just need a receipt attached to the job?
- How are labor rates tracked? By individual employee, or by position — technician, supervisor, project manager? Burdened, or raw wage plus an overhead percentage?
- What consumable overhead rate do you apply to every job?
- Above what amount do you cost consumables directly instead of leaving them in that percentage?
Start with your last three jobs
Do not start by choosing software, and do not start with thirty jobs. Three is enough to find out whether you can answer the question at all — and how much of the answer you have to guess.
Do this next
A one-afternoon job cost check
- Pull your last three completed jobs — different sizes and types, not your three favourites.
- For each one write down what the estimate was, and what you actually invoiced. Note whether you invoiced the full amount.
- Then the cost side: labor cost, and other direct costs — materials, subcontractors, permits, rentals, anything bought for that job.
- Write GUESS next to any number you had to estimate rather than look up. Do not skip this part.
- Use one labor rate method across all three, whichever you settled on above.
- Rank the three by gross margin, and note how the order differs from what you expected.
- Count the GUESSes. Then pull the same three jobs in your accounting system and see whether it gives you any more clarity than you just managed by hand.
- Score each estimate one to five on how accurate and thorough you now think it was.
- Compare best to worst and look for trends — job type, source of loss, which project manager ran it, who was on the crew.
- Put a date on the calendar to run the same exercise at ten jobs, then thirty, using the same method.
That scoring step matters more than it looks. The cost side is close to binary — hours went on the job or they did not, a purchase was attached or it was not. The estimate is qualitative and needs a judgement call. Two jobs can both land at eight points of margin, one because purchasing ran loose and one because the scope was light from the day it was written — and only one of those gets fixed by tightening purchase orders.
Done by hand this is an afternoon you will not want to repeat monthly, which is the honest argument for putting it in the system rather than a spreadsheet. When the estimate, purchase orders, clock-ins and invoice all sit on the same job record, the exercise stops being an exercise — the numbers are already there, on one page, while the job is still open.
Common questions
Isn't job costing what my accountant does?
Your accountant reports it; production creates it. They can only cost a job from records that reached the books attached to that job, so if a tech clocked into the wrong one, or the framing package was bought without a job number, there is nothing for them to work with and no way for them to know. Get the production habits right and the reporting gets easy.
Should equipment be a job cost?
Most companies treat equipment as overhead rather than a per-job cost, and that is a reasonable choice. A dehumidifier is capital that should pay for itself across many jobs, and how you handle depreciation varies enough that forcing it onto every job adds noise rather than clarity. Keep the equipment log on the job so you know where units are and can bill the days — that is worth doing for its own reasons — but you do not need equipment in your cost buckets to get useful job costing.
Do we really need a purchase order for everything?
No. Effort should match the value of the information. Anything bought for a specific job — materials, rentals, a subcontractor — is worth a PO, because that is where overruns hide and the amounts are meaningful. Warehouse consumables are not worth chasing; allocate them as a percentage. And you can start looser than that: attributing purchases to the job they were for, with no approval workflow at all, already puts you ahead of most.
Should labor be burdened, or do we use raw wage?
Practice legitimately varies. You can use each employee's actual rate, a rough rate by position, raw wage with an overhead percentage on top, or a fully burdened rate that already includes taxes, insurance and benefits. All of them work. What matters far more than which you choose is that you keep choosing the same one, because the value of job costing comes from comparing this job to the last thirty. Consistency beats precision here.
Lever360 Software
Know what the job made while you can still do something about it
Lever360 puts job cost where the job is: purchase orders carrying a budget before the spend and landing as cost on the job, approval thresholds on the purchases that matter, labor accruing from clock-ins against the job file, and a QuickBooks Online handoff with a sub-customer per job. Estimate, invoiced, cost and gross profit on one page — while the work is still moving, not two weeks after it closed.
See how job costing works
Want to see it against your own jobs? Book a 20-minute tour.
