A contractor finishes a nine-month subdivision job, sends the final claim, and waits. Four months later the accountant produces the year’s figures and the job made about half what it was priced to make. Everyone agrees it was the wet winter, or the rock in stage two, or the subcontractor who let them down. Nobody can say which, or how much, or when it started going wrong.
That is not a bookkeeping problem. It is the absence of job costing and cost control — the discipline of measuring, every month, what a job has actually cost against what it was priced to cost, and forecasting where it will land. Without it a contractor is flying on instruments that only work after landing.
The number that arrives too late
Most civil SMEs have exactly two moments of financial truth: the estimate, and the annual accounts. Our guide to preparing civil works cost estimates covers the first properly, and the plant hours underneath it are the subject of our guide to plant maintenance and availability. This guide is about everything between the two, and about why the gap matters more than either.
Three things follow from having no measurement in between:
- You cannot intervene. A job losing money in month two can often be recovered — by changing method, resequencing, renegotiating a supply, or pursuing a variation that was being absorbed. A job discovered to have lost money in month twelve cannot.
- You cannot learn. If you do not know which activities beat their rate and which did not, next year’s estimate repeats this year’s errors. The feedback loop that should make an estimator better never closes.
- You cannot tell a good year from a lucky one. Profit at the business level nets winners against losers. Two jobs at plus fifteen and minus twelve look identical to one job at plus three, and they are completely different businesses.
The last point deserves emphasis because it is the one that compounds. A contractor who knows which kinds of work make money can bid more of them and decline the rest. That is a durable advantage, and it is only available to a business that measures at job level.
What cost control actually is
Worth clearing up, because the term is used for three different activities.
| Activity | What it answers | Who usually does it |
|---|---|---|
| Bookkeeping | What have we spent, and is it coded correctly for tax? | Bookkeeper or accountant, monthly or quarterly |
| Job costing | What has this job cost so far, broken into activities? | Whoever codes the invoices and timesheets |
| Cost control | Where will this job finish, and what do we do about it? | The person who priced it, with the person running it |
Most contractors have the first. Some have the second. The third is where the value is, and it is not a bookkeeping function — it is a forecasting conversation between the estimator and the supervisor, held on a schedule. The numbers are inputs to that conversation, not the point of it.
The three numbers
Everything reduces to three figures per activity, and the discipline is keeping them separate.
| Number | Where it comes from | What it is for |
|---|---|---|
| Budget | The estimate, split to the same activities you will measure | The benchmark. Fixed at award and only changed by an approved variation |
| Cost to date | Actual committed and incurred cost, coded to the activity | What has happened. Historic and, on its own, misleading |
| Cost to complete | A judgement made by the people doing the work | The only forward-looking number, and the one that matters |
Cost to date plus cost to complete gives forecast final cost. Forecast final cost against budget gives your forecast result. That single comparison, per activity, updated monthly, is the whole system. Everything else is plumbing.
The mistake that makes cost reports useless is treating percentage spent as percentage complete. Sixty per cent of the drainage budget spent does not mean the drainage is sixty per cent done — it may be forty per cent done in bad ground, in which case the job is already losing and nobody has said so. Physical progress and cost consumed are independent measurements and must be taken independently.
Setting the job up so it can be measured
Most cost control failures are set up at the start, when nobody decides what will be measured.
The requirement is a cost code structure — a list of activities that costs get coded against. Three rules make it work:
- It must match how you estimated. If the estimate has a rate for bulk earthworks and another for detailed excavation, the cost codes must separate them too. A budget you cannot compare to a cost is not a budget.
- It must match how work actually happens on site. A code nobody can allocate to at the end of a shift will be allocated wrongly. If your crew moves between drainage and pavement in a day, either they record the split or you accept that those two codes are one code.
- It must be short. Ten to twenty codes on a typical civil job is plenty. Fifty codes produces guessing, and guessed allocations are worse than coarse ones because they look precise.
Where the contract is a schedule of rates, the schedule gives you the structure for free — code to the schedule items and your cost report reconciles directly to your claim. On a lump sum, you have to impose the structure yourself, and it is worth doing at award rather than in month three. Our guide to contract award and mobilisation covers what else belongs in that first fortnight.
Capturing cost as it happens
Four cost streams, each with its own failure mode.
| Stream | How it is captured | Where it goes wrong |
|---|---|---|
| Labour | Timesheets coded to activity, at a charge-out rate including on-costs | Coded to the job but not the activity, or costed at bare wage rates |
| Plant | Internal hire rate per hour for owned plant; invoices for hired plant | Owned plant treated as free because there is no invoice |
| Materials | Supplier invoices and delivery dockets | Recorded when invoiced rather than when delivered, so a month looks cheap and the next looks terrible |
| Subcontractors | Their claims, plus an accrual for work done and not yet claimed | Same timing problem, at larger amounts |
The timing problem in the last two rows is the single most common distortion in SME cost reports. A cost report built from invoices received is a report about your suppliers’ invoicing habits, not about your job. The fix is to accrue: at cut-off, add the value of material delivered and subcontract work done that has not yet been invoiced. It takes twenty minutes with the site diary and the delivery dockets, and without it the numbers cannot be trusted.
The site diary earns its keep here as well as in claims. Our guide to contract administration for civil SMEs covers the records that protect an entitlement; the same records tell you what was delivered and what was done, which is exactly what an accrual needs.
Cost to complete — the forecast that matters
This is the part that cannot be automated, and it is the part most contractors skip.
Cost to complete is a judgement about the remaining work, made by the people who can see it. It is not budget minus spend — that arithmetic assumes the budget was right, which is precisely the thing you are testing. The question to ask, activity by activity, is: from where we are now, what will it cost to finish this?
What makes the answer honest:
- Ask for quantities and rates, not a dollar figure. “Eleven hundred metres of pipe left at about the rate we have been achieving” is checkable. “Another two hundred thousand” is not.
- Use achieved production, not tendered production. If you have been laying sixty metres a day against a tendered eighty, the remaining work takes the time sixty metres implies. Contractors routinely forecast the rest of the job at the tendered rate they have never once hit.
- Include what you know is coming. The wet season, the shutdown, the difficult section everyone is quietly dreading.
- Do not let optimism close the gap. A forecast that always shows the job recovering to budget in the final month is not a forecast. This pattern has a name on larger projects and it is worth knowing you are doing it.
The most valuable output is not the number but the conversation. A supervisor asked monthly what it will cost to finish starts thinking about it continuously, which is the actual behaviour change.
Reconciling the claim against the cost
Cost is half the picture. The other half is what you have earned, and the two have to be compared on the same basis at the same date. This is the step that turns a cost report into a result.
The comparison is between the value of work completed and the cost of completing it. Value is not the same as what you claimed, and the difference is where a specific and dangerous illusion lives.
- Over-claiming hides losses. If you claim ahead of the work — front-loading rates, claiming for materials on site, or simply claiming optimistically — cash looks healthy while the job is losing. The loss surfaces at the end, when there is nothing left to claim and the remaining work still has to be paid for. What to do once it has surfaced is covered in our guide to managing a loss-making job.
- Under-claiming looks like a loss that is not there. Work done and not claimed makes a good job look bad, and prompts corrective action that is not needed.
- Both are common, and they cancel out at completion, which is exactly why the problem is invisible until then.
The discipline is to value the work at the contract rates for what is genuinely complete, independently of what was claimed, and compare that to cost. Where your claim and that valuation differ, you know your cash position is running ahead of or behind your earnings — which is a different problem from a margin problem and needs a different response. Our guide to cash flow in civil construction contracts covers the cash side; the point here is that healthy cash and a losing job can coexist for months, and only this reconciliation separates them.
The early warning signs
What to look for in a monthly report, roughly in order of how early they appear.
| Signal | What it usually means |
|---|---|
| Achieved production below tendered production | The earliest signal of all, visible before any cost report. Fix the method or reprice the remainder |
| Cost consumed running ahead of physical progress | The rate is wrong, the conditions are worse, or the resourcing is heavier than priced |
| One activity consistently over while others hold | An estimating error or a genuinely different condition — and the most useful thing you will learn all year |
| Growing gap between claim and valued work | You are claiming ahead of earnings; the cash is borrowed from your own future |
| Rising unclaimed variation value | Work is being absorbed. Every month it ages, the entitlement gets weaker |
| Plant hours far above allowance | Double-handling, standing time, or a method that is not working |
| Subcontract commitments above budget before the work is done | The package was bought above the estimate and nobody re-forecast |
| Forecast that recovers to budget every month | Optimism, not forecasting. Ask what specifically changes to produce the recovery |
Variations and the cost of unclaimed work
Cost control and contract administration meet here, and the meeting point is expensive.
Work outside the contract scope costs money the moment it is performed. It earns money only when it is notified, valued and approved — and in most contracts, only if it was notified within a time bar. A cost report that shows an overrun caused by extra work is telling you that the contract administration failed, not that the estimate did. Our guides to variations in civil construction contracts and extension of time and delay claims cover the entitlement mechanics.
Two practical rules keep the two systems connected:
- Give every variation its own cost code the day it is instructed, approved or not. Cost that lands in a base activity code disappears into the overrun and can never be extracted to support a claim.
- Report unclaimed and unapproved variation value as a separate line every month. It is the clearest measure of money sitting on the table, and it forces someone to chase it while the evidence is fresh.
Plant: the cost nobody allocates properly
Owned plant is the biggest systematic distortion in SME job costing, because it generates no invoice. A machine that is already paid for feels free, so its hours get left out — and the job it worked on looks more profitable than it was.
The fix is an internal hire rate: a rate per hour charged to the job for every hour a machine works, covering finance or depreciation, maintenance and repairs, tyres and wear items, insurance and registration, and a share of the fitter and the workshop. Fuel is usually charged separately at actual cost — which should be the cost net of fuel tax credits rather than the price at the pump, or every hour of off-road work is overstated. The rate does not need to be exact — it needs to exist and to be applied consistently, because its purpose is comparison rather than tax accounting.
Three things fall out of doing this, and all three are valuable:
- Jobs get compared honestly. A job that used three machines for four months is no longer flattered against one that hired everything in.
- Utilisation becomes visible. Total hours charged out against hours available tells you whether a machine is earning its keep — the input the hire-versus-buy decision needs, covered in our guide to plant hire agreements.
- Your estimating rates get validated. If your tendered plant rate and your internal rate disagree, one of them is wrong, and it is worth knowing which.
Standing time deserves its own treatment. Hours where a machine was on site and not working are a cost with no production against them, and they are usually caused by something — access, weather, a late delivery, waiting on another trade. Recording them separately turns a vague overrun into a specific, sometimes claimable, cause.
Labour and the on-cost trap
The equivalent distortion on the labour side is costing people at their bare hourly rate.
The real cost of an hour worked includes the wage, plus superannuation, workers compensation premium, leave in all its forms, public holidays, allowances, training and tickets, protective equipment, and the non-productive time that is paid but not charged — travel, inductions, toolbox talks, wet days on pay. Depending on the arrangement, the loaded rate can be a long way above the bare rate. Our guides to awards, enterprise agreements and labour rates and workers compensation and injury management cover the components; the point for cost control is that the rate you cost at must be the same rate you estimated at, and both must be loaded.
A related trap: supervision. On a small job the supervisor may also operate a machine, and their time gets coded to production. That makes the job look efficient and leaves supervision unbudgeted on the next one. Decide how supervision is treated and apply it consistently.
The monthly rhythm
Cost control works when it is a routine rather than an investigation. A workable monthly cycle for a civil SME:
| When | What happens |
|---|---|
| Cut-off day | Same date every month. Timesheets in, dockets in, plant hours in, subcontract work assessed |
| Day 1–2 | Accruals added for material delivered and subcontract work done but not invoiced |
| Day 2–3 | Quantities measured on site — physical progress, independent of cost |
| Day 3–4 | Cost to complete estimated per activity by supervisor and estimator together |
| Day 4–5 | Report produced: budget, cost to date, cost to complete, forecast result, movement since last month |
| Day 5 | The meeting. Half an hour. What moved, why, and what we are doing about it |
| Same week | Actions: reprice remaining work, chase variations, change method, or accept and adjust the forecast |
The movement column is the most useful part of the report and the one most often left out. A job forecasting a small loss is information; a job that has moved from a healthy margin to a small loss in one month is an emergency, and the two look identical without the comparison.
Making it survive contact with a small business
Everything above is standard practice on large projects, delivered by people employed to do it. A twenty-person civil contractor has no commercial manager, and the owner is already estimating, selling and running the jobs. So the honest question is what the minimum viable version looks like.
- Start with one job. The biggest or longest current one. Prove the routine there before applying it everywhere.
- Use ten cost codes, not fifty. Coarse and consistent beats detailed and guessed.
- A spreadsheet is enough to begin. Budget, cost to date, cost to complete, forecast, movement. Accounting software with job costing helps once the routine exists, but buying software does not create the discipline — plenty of contractors own the module and have never run a cost report from it.
- Do it monthly, not weekly. Weekly is unsustainable for a small team and the noise exceeds the signal.
- Protect the meeting. Half an hour, same time, with the person who priced it and the person running it. If only one attends, it becomes a reporting exercise and dies within three months.
- Close the loop to estimating. At completion, compare final cost per activity against the original rates and write down what you learned. This is the step that pays for the whole system, and it takes an hour.
That last point connects cost control to everything else in this library. Measured production rates and validated plant and labour rates make the next estimate better, which makes the next bid sharper — the mechanism our guide to pricing strategies for government tenders depends on. A contractor who knows their real costs can price close to the market with confidence. One who does not is guessing, and eventually guesses low on the wrong job.
It also has a direct tendering benefit that is easy to miss. Clients and prequalification schemes assess financial management, and a contractor who can produce job-level cost reporting is demonstrating a control environment that most SME competitors cannot. Our guide to demonstrating financial capacity in tenders covers what assessors look for; cost control is one of the few things that improves both the assessment and the underlying reality.
Checklist
- Does every job have a budget split into the same activities you can measure?
- Do your cost codes match both how you estimated and how work happens on site?
- Are there ten to twenty codes rather than fifty?
- Is labour costed at a loaded rate, and is it the same rate you estimated at?
- Is owned plant charged to jobs at an internal hire rate?
- Is standing time recorded separately from working hours?
- Do you accrue material delivered and subcontract work done but not yet invoiced?
- Is physical progress measured independently of cost consumed?
- Is cost to complete estimated by the people doing the work, using achieved production rather than tendered production?
- Does every variation get its own cost code the day it is instructed?
- Is unclaimed and unapproved variation value reported as a separate line?
- Do you compare the value of work completed against cost, independently of what you claimed?
- Does the report show movement since last month, not just position?
- Is there a fixed monthly meeting with both the estimator and the supervisor?
- At completion, do you compare final cost per activity against the original rates and record what you learned?
The short version
- Most civil SMEs have two moments of financial truth — the estimate and the annual accounts — and everything that matters happens in between.
- Three numbers per activity: budget, cost to date, cost to complete. The third is the only forward-looking one and the one most often skipped.
- Percentage spent is not percentage complete. Measure physical progress independently of cost.
- Cost to complete is a judgement made from achieved production, not budget minus spend.
- A cost report built from invoices received describes your suppliers’ invoicing habits. Accrue what has been delivered and done.
- Compare the value of completed work against cost, independently of what you claimed — over-claiming hides losses for months.
- Owned plant with no invoice is not free. Charge an internal hire rate or every job that used it looks better than it was.
- Give every variation its own cost code the day it is instructed, and report unclaimed variation value monthly.
- Report movement, not just position. A forecast that recovers to budget every month is optimism, not forecasting.
- Start with one job, ten cost codes and a spreadsheet. Software does not create the discipline.
- Close the loop to estimating at completion. That hour is what makes the whole system pay.
Sources and further reading
This guide is general information for Australian civil construction businesses and is not accounting, taxation, audit or financial advice. The approach described here is management reporting for internal decision-making; it is not a financial reporting framework and does not follow the recognition and measurement requirements of the Australian Accounting Standards, which govern how revenue and costs on construction contracts must be reported in statutory accounts. Internal plant rates, labour on-cost loadings and accrual practices described here are management conventions, not tax positions. Always work from your executed contract, and take advice from your accountant on statutory reporting, revenue recognition and tax treatment.
- Standard construction cost management practice as applied in Australian civil contracting — the budget, cost-to-date and cost-to-complete structure described in §03, the accrual treatment in §05, and the reconciliation of work value against cost in §07. These are established management accounting conventions rather than a regulated framework, and the specific structure a business adopts is a matter for it and its accountant.
- Australian Accounting Standards governing revenue recognition and the reporting of construction contracts in statutory financial statements, referenced in the disclaimer above. These determine how contract revenue and costs appear in published accounts and operate independently of the internal management reporting described in this guide.
- Related TenderBuilt guides carrying the primary-source detail referenced above: preparing civil works cost estimates (how the budget is built), schedule of rates vs lump sum, cash flow in civil construction contracts, variations, extension of time and delay claims, contract administration for civil SMEs, contract award and mobilisation, awards, enterprise agreements and labour rates, plant hire agreements, demonstrating financial capacity and pricing strategies for government tenders. See also winning work in a downturn.