A total is not an estimate. What a funding body actually reviews is the structure — and an estimate whose structure does not match the one they are obliged to report against will fail, even when the number at the bottom is right.
The Queensland Project Cost Estimating Manual (PCEM) splits a project estimate at the top into two categories that answer a single question: who is going to spend this money. Construction contractor’s costs cover everything the contractor must price to build and hand over the physical works, paid through the construction contract as progress claims, and generally running 60–75 per cent of total project cost. Principal’s costs cover everything the owner spends to plan, investigate, design, approve, procure, supervise and commission the same asset, funded directly or through separate consultant engagements, and generally running 25–40 per cent.
That split is not bookkeeping. It sets who is accountable for each dollar, it defines the boundary of what goes out to tender, it drives cash-flow forecasting because the two halves are spent on completely different timelines, and it is the only basis on which one project can honestly be benchmarked against another. A project whose estimate does not separate them cannot be compared with anything.
Below that first split, the contractor’s half divides again into three levels, each with a different business purpose: costs that attach to physical work, costs that support the site without attaching to any item, and the business-level recovery that keeps the contracting entity solvent. Getting those three in the right boxes is what the rest of this page is about.
TX1:Trinity keeps directs, overheads, markup and margin as separate objects and lets you carry several classification trees over the same costs at once — so one estimate reports in your structure and in theirs.
Start a free 14-day trialDirect job costs are the costs that vary with the quantity of work performed and can be attributed to a physical item: labour with its allowances, on-costs, superannuation, workers compensation and non-productive time; plant with hire or ownership, fuel, maintenance and mobilisation; materials, both permanent and the temporary works consumed getting them in place; and subcontract packages, which arrive already carrying somebody else’s overhead and margin. On civil infrastructure, labour commonly falls in the 30–45 per cent range of direct costs, plant 20–30 per cent on mechanised work, and materials 30–40 per cent, running higher on structures. These are the numbers produced by building rates from first principles.
Indirect job costs support the site without belonging to any one item, and they divide by behaviour rather than by discipline. Recurring costs run with time: project and works management, site administration, quality, safety and environmental staff, site vehicles, offices, amenities and security. Fixed costs happen once: establishment, mobilisation, survey control and setting out, temporary works design, insurances and performance security, then demobilisation and clean-up. That distinction matters because the two pools should not be distributed by the same rule.
Off-site overheads and margin is the business layer, calculated as a markup on the sum of direct and indirect costs. PCEM identifies four components with typical ranges: business unit overheads at 3–6 per cent, corporate overheads at 2–4 per cent, the contractor’s own contingency at 2–5 per cent, and profit at 3–8 per cent — 10–23 per cent combined. Those ranges move with the market. In a busy market with few bidders, contractors price at the top; in a thin one they cut into margin to hold their workforce together.
Attaches to an item and scales with its quantity. Labour, plant, materials, subcontract.
Supports the site, attaches to no item. Splits into time-related recurring and one-off fixed.
Business recovery and profit, applied as markup over the two layers below. Typically 10–23 per cent.
The reason this structure transfers cleanly into an estimating tool is that PCEM is describing the same object model an estimator builds anyway. Each cost category corresponds to a distinct kind of object with its own behaviour, and the correspondence is close to one for one.
| PCEM cost element | What it holds | Where it lives in TX1:Trinity |
|---|---|---|
| Direct job costs | Labour, plant, materials and subcontract consumed by a physical work item. | Resources typed L. P. M. S. built into the item, resolving to that item’s direct rate. |
| Crews and plant spreads | The assembled gang that actually performs the work, priced as one hourly figure. | Group resources (G.), whose published rate is the sum of their component rows. |
| Indirect job costs — recurring | Supervision, site administration, quality, safety, environmental, facilities. Runs with time. | An overhead pool allocated across items, appearing in each rate as the allocated component. |
| Indirect job costs — fixed | Establishment, mobilisation, setting out, temporary works design, insurances, demobilisation. | Separate overhead entries with their own allocations, so one-off costs are not distributed like time-related ones. |
| Business unit and corporate overheads | Regional office, estimating department, head office, corporate services. | Markup, set per resource type so subcontract packages are not marked up like own labour. |
| Contractor’s contingency | Retained risk on minor estimating variance and productivity. | The risk component of the margin layer, above total cost and separable from it. |
| Profit margin | Return to the business for carrying the work. | The profit component of the margin layer, applied after risk and corporate recovery. |
| Principal’s costs and obligations | Phase costs, property acquisition, public utility plant relocations. | Priced as items like any other, then separated by the branch of the code set they allocate to. |
| Estimate and scope contingency | Quantified risk exposure held by the principal, not by the contractor. | Held outside the base estimate; the build-up carries the most likely cost only. |
The one thing to take from this page: a reviewer is not checking your total — they are checking whether your structure lets them find, and test, the one number their own governance makes them responsible for.
Assume for a moment that your estimate is correct to the dollar. It will still be rejected if a reviewer cannot take it apart along the lines their framework requires. This is not bureaucratic pedantry; it is the only way the review can be done at all. The reviewer has a template, a benchmark set and a reporting obligation, and every one of those is expressed in the categories above.
Consider what each structural failure removes from them. Bury the off-site overhead and margin inside item rates and there is no way to test whether the markup falls in the expected 10–23 per cent band; the reviewer cannot tell a competitive price from a generous one. Mix principal’s costs into the contract sum and the contractor share is inflated, so every benchmark comparison the agency holds returns nonsense. Leave contingency inside the base estimate and the number presented as a most likely cost is already risk-loaded, which means the quantitative risk analysis run on top of it double counts, and the confidence level reported to the investment committee is not the confidence level they think they are approving.
None of those defects change the total. All of them end the review. And the cost is rarely a polite request for restructuring — it is a gate missed, a funding round waited out, and an estimate that has to be rebuilt by someone who was not there when the assumptions were made.
The principal’s half is the one most often estimated as a percentage and then forgotten, which is unfortunate, because its shape across the project life is well understood and quite unlike the construction spend. A modest planning allowance supports the decision to proceed at all. Concept work steps up sharply to fund the survey, geotechnical investigation, environmental studies and options analysis a business case needs. Development is the heavy phase, carrying design to tender-ready through two stages: S1D takes design to roughly 30–50 per cent with an estimate at about ±20 per cent, and S2D takes it to 80–100 per cent with a pre-tender estimate at about ±10 per cent. Implementation is dominated by contract administration and construction surveillance, and finalisation covers as-built documentation, defects and handover.
| Phase | What the principal is buying | Share of principal’s costs |
|---|---|---|
| Planning | Network and corridor studies, needs analysis, early stakeholder engagement. | ~5% |
| Concept | Survey, geotechnical investigation, environmental studies, options analysis, business case. | ~10% |
| Development (S1D + S2D) | Detailed design, approvals, tender documentation, procurement to contract award. | ~30% |
| Implementation | Superintendence, contract administration, surveillance, testing, stakeholder liaison. | ~50% |
| Finalisation | As-built records, asset system updates, defects management, close-out. | ~5% |
Two obligations sit outside that profile because they behave differently: property resumption and acquisition, and public utility plant relocations. Both are volatile. Property outcomes turn on valuation, negotiation and occasionally the Land Court; utility costs are difficult to pin down until investigation is complete, and undocumented services remain one of the most reliable sources of overrun in Australian civil work. Estimating either as a neat percentage of construction cost is a decision to be surprised later.
Two more structural disciplines finish the anatomy. Contingency belongs outside the base estimate and must be identified by type — contractor’s, estimate, scope and programme — because each is held by a different party and determined a different way. And every estimate needs a base date, with escalation applied against the cash-flow profile using construction-specific indices rather than headline CPI, shown as its own line. Whether all of this is even expected of you depends on the class of estimate you are producing, which is the subject of the next chapter.
The layers of the PCEM structure are separate objects in TX1:Trinity rather than columns in a spreadsheet, which is what makes them reportable independently. Direct costs come from typed resources built into items. Overheads are pools that are allocated rather than smeared, and the allocated portion stays visible in the rate: CalculatedSellRate = DirectRate + (TotalAllocated / Qty). Above that, the pipeline runs direct costs plus overheads to a total cost, then markup applied per resource type, then margins split into risk, corporate and profit, to a sell price — the same four-part off-site recovery PCEM describes, held as four things instead of one blended percentage.
The part that matters most for review is code sets. A project can carry several parallel classification trees over the same costs at once — your own work breakdown structure, a contract schedule such as AS 2124, and the client’s chart of accounts — with percentage-split allocations where a cost belongs partly to more than one branch. That is what lets a single estimate report in the structure you build in and the structure the funding body reviews in, without maintaining two versions that drift apart. It is also how principal’s costs and obligations stay distinguishable from the contract sum: not by living in a different file, but by allocating to a different branch.
Carbon follows the same skeleton. Emission factors attach to the same resources that carry cost, scoped to the GHG Protocol’s Scope 1, 2 and 3, with data-quality ratings from 1 to 5, a verification workflow and snapshots — so the emissions report and the cost report are two views of one structure rather than two exercises. Libraries and whole projects move as .tx1 packages, compressed and checksum-validated, with AES-256 encryption available.
Contractor costs are everything the construction contractor must price to design, build and hand over the physical works, and they are paid through the construction contract as progress claims. Principal costs are everything the project owner spends to plan, investigate, design, approve, procure, supervise and commission the project, plus obligations such as property acquisition and utility relocations. Under PCEM, contractor costs generally run 60 to 75 per cent of total project cost and principal costs 25 to 40 per cent, varying with delivery model and complexity.
Because a reviewer is not verifying your arithmetic, they are reconciling your estimate against the structure their own governance requires. If off-site overhead and margin is buried inside item rates, the reviewer cannot test it against the expected 10 to 23 per cent band. If principal costs are mixed into the contract sum, the contractor share cannot be benchmarked. If contingency sits inside the base estimate, the confidence level being reported is not what it claims to be. Each of those is a structural failure and none of them changes the total.
Indirect job costs support construction without being attributable to any single work item: site management, supervision, quality and safety staff, environmental controls, site facilities, and the one-off costs of establishment, setting out, insurances and demobilisation. They divide into recurring costs that run with time and fixed costs that occur once. Allocate the recurring pool on a basis that reflects duration and the fixed pool on a basis that reflects what triggered it, rather than spreading both as a flat percentage over every item.
Outside the base estimate, and identified by type. The base estimate should represent the most likely cost if the defined scope is delivered and no risk event occurs. Contractor contingency is a markup the contractor carries for its own retained risks. Estimate contingency is determined by the principal through quantitative risk analysis against the risk register. Scope contingency covers scope not yet defined, and programme contingency sits above the project. Contingency is never a substitute for estimating effort or investigation.
State a base date, then escalate against the cash flow rather than uniformly. Every rate in the estimate is expressed in base date dollars, so the first requirement is to say what that date is. Escalation is then applied using construction-specific indices rather than headline CPI, weighted by when the expenditure actually occurs, and shown as a separate line so it stays visible. Forecasts more than two or three years out carry real uncertainty and should be treated as a risk in their own right.
Build it once, classify it twice. Try TX1:Trinity on a real schedule and see the same costs reported through your breakdown and through theirs.