What estimating is, and where it goes wrong.
Most builders learn estimating on the job, under time pressure, from whoever taught them. It is worth setting out properly, because almost every bad building outcome in Australia starts with a number that was never measured.
An estimate is a forecast, not a price.
A construction estimate forecasts what a project will cost to build: materials, labour, plant, subcontractors, preliminaries, overheads and margin. It is built before anyone breaks ground, from drawings that are never quite complete, about a site nobody has fully investigated yet.
That is why the discipline matters. An estimate is a set of assumptions with numbers attached. The good ones are honest about which is which.
A quantity take-off is the measuring half: cubic metres of concrete, square metres of plasterboard, lineal metres of gutter. A bill of quantities is that take-off written out line by line so it can be priced, compared and checked. Rates are the pricing half, and they belong to the builder.
The eight steps of a proper estimate.
This is the sequence a quantity surveyor follows. Red Dot does steps one to four and hands you the rest with the working shown.
Read the documents before measuring anything
Architectural, structural, hydraulic, BASIX, geotechnical, consent conditions. The quantities that cost you are rarely on the floor plans. They are in the schedules, the sections and the engineer's notes — the pages that get skimmed when a tender closes Friday.
Establish the scope, in writing
What is in the contract and what is the owner supplying? Is the kitchen included? How far does the landscaping go? What is a provisional sum and what is a prime cost item? Most disputes about money are actually disputes about scope.
Measure the quantities
Areas, volumes, lineal metres, counts. This is the part that takes the hours, and the part that gets skipped when a tender closes Friday. It is also the only part where an error is entirely yours — a wrong rate is the subbie's problem, a wrong quantity is not.
Audit what the drawings claim
Printed areas and volumes are claims, not facts. Consultants work from different models at different revisions. The architect's gross floor area, the energy assessor's room schedule and the hydraulic engineer's impervious area routinely disagree on the same job.
Decide what is measured and what is a package
Only about a dozen trades are genuinely priced on a unit rate — concrete, framing, plasterboard, tiling, painting and a few more. The rest are lump sums, because the subbie sells a complete outcome, or the item is fabricated off a schedule, or site conditions set the price.
Apply rates, and be honest about their grade
A rate from a real invoice last month is not the same thing as a rate from a handbook, which is not the same thing as an educated guess. Estimates go wrong when all three are printed in the same font and nobody can tell them apart afterwards.
Size the preliminaries to the site you actually have
Preliminaries are commonly ten to fifteen per cent of a residential job and they scale with time, not floor area. Supervision, scaffold, site sheds, craneage, traffic management and waste all cost more on a steep block with poor access, and more again if the programme slips.
Add contingency, then say what it is for
The convention is five to ten per cent, and a contingency with no stated purpose is just padding. A useful one names the risks: the rock volume nobody has measured, the trade with no drawings issued, the long-lead item that is not a catalogue product.
Where estimates actually go wrong.
It is rarely the rates. It is almost always one of five things, and all five are avoidable.
Quantities that were never measured
Someone eyeballed a plan at 10pm after a day on site because the tender closed Friday. It is not a skill problem. It is that measuring properly takes hours nobody had.
A printed figure taken on trust
Drawings carry a "do not scale" note for a reason, and their stated areas and volumes are no more reliable than the model they came from.
Scope assumed rather than confirmed
Two people read the same drawings and price two different jobs. Is the kitchen in the contract or owner-supplied? Where does the landscaping stop? Most arguments about money later are arguments about scope now.
Preliminaries sized off floor area
Supervision, scaffold, site sheds, craneage and waste scale with how long the job runs and how hard the site is to work on, not with how big the house is. Two jobs of identical floor area can be thousands apart on prelims alone.
A missing document nobody chased
No geotechnical report means the rock is a guess, and rock is often the largest single unknown on a sloping site.
Where the drawings will not carry a price, we do not invent one.
Joinery with no elevations, finishes with no schedule, a trade with nothing issued yet. That line becomes a provisional sum, says so on its face, and comes with the schedule your subcontractor needs to price it properly. A named provisional sum you can send out for a real quote beats a confident number nobody can stand behind.
Why this is worth paying for.
Open tendering is a numbers game with a low win rate and a real cost per attempt. Every set you price costs you hours whether you win it or not. If the take-off is the slow part — and it usually is — then buying the take-off lets you price more tenders in the same week, on measured quantities instead of a rushed guess.
That is the whole commercial argument, and it is why builders who already employ an estimator still use us for overflow.