Key takeaways
- A job costing system only works if costs are coded against the same cost codes used in the estimate, because uncoded invoices hide the variance until the job is finished.
- Weekly review of committed costs, not just paid invoices, is what shows where a job stands before the money is spent.
- Trades that carry high material content, subcontractor labour and site-specific conditions reliably produce the largest variances, so they deserve the closest tracking.
- Distinguishing a pricing error from a site problem determines whether you absorb the cost or seek a variation, and the answer depends on what the contract documents actually required.
- Closing the loop by feeding variances back into the next estimate is the only way job costing improves future margins rather than just recording past losses.
Most builders can tell you what a job cost once it is finished. Far fewer can tell you where it stands today, this week, while there is still time to do something about it. The difference is not better accounting software. It is the discipline of coding every cost against the estimate as it occurs, and reviewing the result on a fixed cycle.
Job costing is often treated as a finance function that happens after the fact. On a residential job, it is an operational control. The estimate is a prediction; the job costing system is how you find out whether the prediction was right, and why it was wrong, before practical completion removes your options.
This article covers the mechanics: how to structure cost codes so actuals and estimates can be compared, what to review weekly, which trades reliably produce the largest variances, and how to tell a pricing error from a site problem. The distinction matters commercially because one is yours to absorb and the other may be recoverable.
Why the estimate and the actuals must share the same cost codes
A cost code is a label you attach to a portion of the work — for example, structural steel, upper floor framing, wet area waterproofing, or electrical rough-in. The estimate is built up from those codes. The job costing system should receive costs under those same codes. If the two structures do not match, you cannot compare them, and the comparison is the entire point.
The common failure is a chart of accounts built for accounting rather than for estimating. The accountant wants costs sorted by expense type: materials, subcontractors, labour, plant. The estimator wants costs sorted by element of the building. If invoices are coded to the accounting structure, you can produce a tax return but you cannot tell whether the framing package is running over. Build one code structure and use it for both.
This is not a software problem, although software helps. It is a decision made at setup: the estimate is broken into codes, and those codes are loaded into the job costing system before the first invoice arrives. Retrofitting codes after costs have been posted is possible but slow, and it tends to happen only when a job is already in trouble.
Committed costs: the number most builders miss
Cost to date is what you have been invoiced for. Committed cost is what you have agreed to spend but not yet been invoiced for. A subcontractor engaged to frame the upper floor represents a commitment from the moment the contract is signed, even though no claim has been received. If your job costing only tracks paid invoices, the job looks healthier than it is for the entire duration of the subcontract.
The fix is to enter commitments when they are placed. When a purchase order is raised or a subcontract is executed, the value goes into the job costing system against the relevant code. The forecast cost for that code is then the higher of actual cost plus commitments, not just actual cost. This is what allows a meaningful review while the job runs.
Committed costs also expose the gap between what you allowed in the estimate and what you actually agreed to pay. If the estimate allowed a figure for the electrical rough-in and the subcontract was signed for more, that variance exists at the moment of signing. It does not need to wait for the invoice. Finding it early is the difference between adjusting your cash flow and discovering the problem at final claim.
What to review weekly, and why weekly
A monthly review is too slow on a residential job. By the time a month has passed, a trade that is running over may have finished and demobilised. Weekly review keeps the variance close to the event that caused it, which means the people who can explain it still remember the details.
The weekly review is not a full financial report. It is a scan of the cost codes where actual plus committed cost is drifting from the estimate. Set a threshold — a figure or a proportion of the code value — and investigate anything beyond it. The review should cover labour hours booked against hours allowed, subcontractor claims against scope, and the forecast cost to complete for each code.
What to review each week
- Cost to date plus commitments against the estimate for every cost code.
- Any code where the actual or committed cost exceeds the estimate by more than your agreed threshold.
- Labour hours booked against the hours allowed in the estimate for each task.
- Subcontractor claims paid and claimed but not yet paid, matched to their scope.
- Variations raised but not yet approved, and their effect on the forecast.
- Materials ordered and delivered but not yet invoiced, so the commitment is visible.
- The forecast cost to complete for each code, not just the cost incurred so far.
Forecast cost to complete is the part that inexperienced job costing misses. Knowing that a code has spent most of its allowance is only useful if you also know how much work remains. A code that has spent most of its budget and is nearly complete is fine. A code that has spent most of its budget and is half done is a problem, and the weekly review is where that becomes visible.
Keep the review short and consistent. A standing agenda item at the same time each week, with the same report, is more effective than an exhaustive monthly session. The goal is to build a habit of looking at the numbers while the job is live, not to produce a document.
Trades that reliably blow out, and why
Variances cluster. Certain trades carry characteristics that make them more likely to run over the estimate, and knowing which ones to watch means you can apply your attention where it pays.
Trades with high material content are exposed to price movement between tender and order. If the estimate is prepared well before the material is ordered, the rate used in the estimate may not be the rate you pay. This is a pricing risk, and it belongs in the estimate as an allowance or a clearly stated assumption.
Trades with a large subcontractor labour component are exposed to productivity assumptions. The estimate assumes a certain output per day. If the site access, the sequence or the condition of the work front differs from that assumption, the labour hours climb. This is a site problem more often than a pricing error, but it only becomes recoverable if the contract documents support the assumption you made.
Trades that depend on site conditions — excavation, drainage, waterproofing, and anything affected by weather — produce variances that are hard to predict at tender. The estimate should carry a stated allowance for these, and the job costing should track consumption of that allowance separately from the base estimate. If the allowance is absorbed into the trade code, you lose the ability to see when it has been used up.
Pricing error versus site problem: what to look for
| Pricing error | Site problem |
|---|---|
| The estimate omitted a task that the documents required. | The site conditions differ from what the documents described. |
| The rate used was too low for the actual work. | The work took longer because of access, weather or ground conditions. |
| The quantity was under-measured against the drawings. | The quantity was correct but the material behaved differently on site. |
| The same variance appears across multiple jobs. | The variance is isolated to this site and this trade. |
| The subcontractor priced the same scope higher than you allowed. | The subcontractor encountered a condition that was not visible at tender. |
Telling a pricing error from a site problem
When a code runs over, the first question is not how to recover the money. It is what caused the variance. The answer determines whether recovery is even available.
A pricing error is a mistake in the estimate. The work was required by the documents, the quantity was correct, but the rate or the allowance was wrong. The same error will appear on the next job of the same type. Pricing errors are absorbed by the builder, and the only remedy is to correct the estimating database so the next job is priced better.
A site problem is a difference between the conditions assumed in the estimate and the conditions actually encountered. The ground was rockier than the geotechnical report suggested, the access was tighter than the drawings showed, or the existing structure was in worse condition than the inspection revealed. Whether this is recoverable depends on the contract and on what the documents told you to expect. If the documents described the condition and you priced it wrong, it is still a pricing error. If the condition was not reasonably discoverable and the contract does not place the risk on you, there may be a basis for a variation.
The practical test is whether the same variance would appear on an identical job with identical documents. If it would, it is a pricing error. If it is specific to this site, it is a site problem. The two require different responses, and confusing them costs money in both directions: treating a pricing error as a claim wastes time and damages the relationship, while treating a recoverable site problem as a pricing error gives away margin you were entitled to.
Worked example: a builder in western Sydney
A builder in western Sydney is running a two-storey dwelling with a framed upper floor over a suspended slab. The estimate was prepared from the architectural and structural drawings, with allowances for excavation, piering, slab, framing, roofing, cladding, wet area waterproofing, and the services rough-ins. The job is at lock-up stage and the builder is preparing the claim for the next progress payment.
During the weekly review, one code stands out: the wet area waterproofing. The estimate allowed a figure for the waterproofing subcontractor based on the number of wet areas shown on the drawings. Actual plus committed cost for the code is already above the estimate, and the work is not complete.
The builder investigates. The subcontractor priced the work from the same drawings, but the scope of the waterproofing subcontract excludes the preparation of the substrate, which the estimate had assumed was included. The substrate preparation is now being carried out by the tiler at additional cost, and that cost is landing in the same code. The variance is a pricing error: the estimate misunderstood what the subcontractor’s price covered, and the same mistake would appear on any similar job. The builder absorbs the cost, corrects the estimating template so the next waterproofing estimate includes substrate preparation as a separate line, and moves on.
On the same job, a second code is running over: the piering. The estimate allowed for piers to a depth based on the geotechnical report. On site, the pier rig has encountered a layer of material that required deeper piers across part of the footprint. The builder checks the contract documents. The report described the expected conditions and recommended a design assumption, and the contract places the risk of subsurface conditions on the owner where they differ materially from the report. The builder documents the actual depths, the rig logs and the correspondence, and raises a variation for the additional piering. This is a site problem, and it is recoverable because the documents and the contract support it.
The two variances appeared in the same weekly review. One was absorbed and used to improve the estimating database. The other was documented and recovered. Neither would have been identified in a monthly review after the trades had left site.
The job costing cycle on a running job
- 1Set up cost codes from the estimateBreak the estimate into the same codes you will use to code invoices, so actuals land where the estimate expected them.
- 2Code every cost as it arrivesAllocate supplier invoices, subcontractor claims and wages to the correct code on the day they are processed, not at the end of the month.
- 3Record committed costsEnter subcontractor and purchase order commitments when they are placed, so the forecast includes money you have promised but not yet paid.
- 4Review actual versus estimate weeklyCompare the cost to date plus commitments against the estimate for each code and investigate any variance beyond your threshold.
- 5Decide and actDetermine whether the variance is a pricing error, a site problem or a scope change, and take the commercial action that follows.
- 6Feed back to the next estimateRecord the causes so the next job is priced with better knowledge of the real cost of that work.
Setting up the system so it survives a busy job
Job costing fails when it depends on someone remembering to do it at the end of a long day. The system has to make the correct behaviour the easy behaviour. That means cost codes loaded before the job starts, a coding rule that every invoice and claim is allocated on receipt, and a commitment register that is updated when purchase orders and subcontracts are raised.
It also means the person coding invoices needs to know which code applies. A short code list with clear descriptions, kept to a manageable number, is more likely to be used accurately than an exhaustive structure that requires judgement on every line. If your team struggles to adopt the software that supports this, the problem is usually the setup and the training rather than the tool. The same discipline applies to keeping drawings and revisions aligned with the costs you are tracking; if the drawing register is a mess, the cost codes will be too.
The commercial reality is that job costing costs time every week, and the return is a margin you can see and defend. A builder who reviews weekly catches a running trade while there is still scope to resequence, renegotiate or claim. A builder who reviews at the end catches nothing but a smaller profit.
For builders working through their continuing professional development, the skills that support this — reading contract documents, understanding cost structures, managing subcontractor scope — are covered across the PointCPD course catalogue. If you are new to the NSW CPD requirements, the guide to how NSW builder CPD works sets out what applies to your licence category.
The mistake that costs the most. Tracking only paid invoices and reviewing monthly means a trade can finish, demobilise and be paid before you see the variance. By then the only remaining option is to absorb it. Weekly review of actual plus committed costs is what keeps the decision in your hands.
Feeding variances back into the next estimate
A job costing system that records variances but never changes the estimate is an expensive diary. The value comes from the loop: variance identified, cause classified, estimating database updated. Over several jobs, the rates and allowances in the estimate move closer to reality, and the margin becomes more predictable.
The classification matters here too. Pricing errors should change the rate or the allowance in the database. Site problems should change the assumptions and the allowances carried for site risk, and the contract terms that determine who bears that risk. If both are lumped together as “the job ran over”, neither lesson is captured.
Keep the feedback specific. A note that framing ran over is not useful. A note that framing labour ran over because the upper floor joists required additional trimming around a services penetration that was not shown on the drawings is useful, because it tells the estimator to check for that detail next time and tells the site supervisor to look for it at set-out.
What to do on the next job
Before the job starts, break the estimate into cost codes and load them into the job costing system. Decide the review threshold and the weekly review time. Brief the person who codes invoices on the code list and the rule that every cost is coded on receipt. Set up the commitment register so purchase orders and subcontracts are entered when they are raised.
During the job, run the weekly review against the same report every week. Investigate variances beyond the threshold, classify each as a pricing error or a site problem, and take the action that follows. Document site problems as they occur, because the evidence is much harder to assemble after the work is covered up.
At the end of the job, update the estimating database with what you learned. The next estimate is the product of this job’s costing, and the margin on the next job depends on how honestly you close the loop.
Questions builders ask
What is job costing for builders?
Job costing is the practice of tracking the actual cost of a job against the estimate as the work proceeds. It involves coding every cost to the same cost codes used in the estimate and reviewing actual plus committed costs on a regular cycle. The purpose is to know where the job stands while there is still time to act, rather than discovering the result at practical completion.
Why do I need to track committed costs and not just paid invoices?
Paid invoices only show what you have already been billed for. Committed costs show what you have agreed to spend through subcontracts and purchase orders but have not yet been invoiced for. Without commitments, a job with a large subcontract looks healthier than it is until the claims arrive, which may be after the trade has finished.
How often should I review job costs?
Weekly is the practical cycle on a residential job. A monthly review is often too slow, because a trade that is running over can finish and demobilise before the variance is seen. A short weekly scan of the codes that are drifting from the estimate keeps the variance close to the event that caused it.
Which trades are most likely to run over the estimate?
Trades with high material content are exposed to price movement between tender and order. Trades with a large subcontractor labour component are exposed to productivity assumptions. Trades that depend on site conditions, such as excavation, drainage and waterproofing, are exposed to conditions that are difficult to predict at tender. Watch these codes most closely.
How do I tell a pricing error from a site problem?
A pricing error is a mistake in the estimate that would appear on any identical job with identical documents. A site problem is a difference between the conditions assumed in the estimate and the conditions actually encountered on this site. The distinction matters because a pricing error is absorbed by the builder, while a site problem may be recoverable depending on the contract.
Does job costing require specialised software?
Software helps, but the discipline matters more than the tool. What is required is a single cost code structure shared by the estimate and the job costing system, a rule that every invoice and claim is coded on receipt, and a commitment register updated when purchase orders and subcontracts are raised. These can be maintained in a spreadsheet if the habit is consistent.
Sources
- Small-scale technology certificates · Clean Energy Regulator
- National Construction Code 2025: what’s changed for residential builders · Housing Industry Association
Published 8 July 2026. General information for builders, not legal, financial or engineering advice for a specific job. Check current requirements with the relevant regulator or standard before relying on them.
