The most dangerous number in any construction project is the first one — the figure someone quotes off the top of their head before anyone has measured anything. It anchors the whole project. The owner remembers it, plans around it, and feels cheated when the real cost lands 40 percent higher. The number was never wrong because the building got expensive; it was wrong because it was never an estimate in the first place. It was a guess wearing an estimate's clothes.
The takeaway up front: a construction estimate is not one number, it is a number with a known confidence level, built up from defined cost categories, and it gets more accurate as the design gets more detailed. Estimate in layers — hard costs, soft costs, site work, contingency, and escalation — match the method to how much you actually know yet, and treat a suspiciously low figure as the warning it is. A realistic estimate that you can build to is worth far more than an optimistic one you cannot. This is the financial spine of the wider project; it sits alongside the phases covered in our construction project guide.
An estimate has a confidence level, not just a figure
Every estimate is really a range, and the honest version states how wide that range is. Early on, when all you have is an idea and a floor area, your number might be accurate to plus or minus 30 or 40 percent. By the time you have full construction documents and priced quotes, that range narrows to a few percent. The figure that matters is not the midpoint — it is the upper end you can live with, because construction rarely surprises you by coming in cheap.
This is why "how much will it cost?" has no single answer until you say at what stage. Treating a rough early figure as a fixed budget is the original sin of construction finance. Carry the range, tell whoever is funding the project which stage the estimate represents, and tighten it deliberately as the design firms up.
The four levels of estimate — match the method to what you know
You estimate differently depending on how much design exists. There are four broadly recognised levels, each more accurate and more expensive to produce than the last.
- Rough order of magnitude (ROM). A ballpark from almost nothing — a comparable project, a rate of cost per unit. Useful to decide whether the project is even plausible. Wide range, near-zero effort.
- Square-foot / area estimate. Floor area multiplied by a realistic local cost per square foot (or per square metre) for that building type and quality. The workhorse of early-stage budgeting, and the one most owners use.
- Assembly estimate. Cost built up from major building systems or assemblies — foundation, structure, envelope, services, fit-out — each priced as a block. Needs schematic design and gives a meaningfully tighter number.
- Detailed (quantity) estimate. A line-by-line takeoff of every material and labour quantity from full drawings, the basis of a contractor's firm bid. The most accurate and the most work, and usually the contractor's job, not yours.
The discipline is to use the right level for the decision in front of you. A ROM is fine to decide whether to keep going; you should not commit money on one. Before you sign a contract, you want at least an assembly estimate, and ideally a detailed bid against complete documents.
The square-foot trap
The area method is genuinely useful, but it fails in two predictable ways, and both burn owners. First, the rate is not universal — cost per square foot varies enormously by building type, finish quality, region, and site, so copying a rate from a basic warehouse onto a high-finish home, or from one city onto another, produces fiction. Source the rate from comparable projects in your own location and quality tier.
Second, a square-foot rate usually excludes the things that wreck budgets — difficult site conditions, demolition, long service runs, fees, and contingency. The rate prices a generic building; it does not price your site or your professional costs. Use the area method to get in the right ballpark, then build the full budget by category below. The rate is a starting point, never the whole answer.
The cost categories owners forget
A real budget is the contractor's price plus everything around it. Build it in layers so nothing hides:
- Hard costs — the physical construction: labour and materials for the structure, envelope, systems, and finishes. This is the part most people picture, and it is rarely the part that surprises them.
- Soft costs — design and engineering fees, permits, surveys, inspections, legal, insurance, and project management. Soft costs are commonly a meaningful fraction of hard costs and are the single most forgotten layer. They are real money and they come due whether you budgeted them or not.
- Site and enabling work — demolition, clearing, earthworks, utility connections, access, and anything to make the ground ready to build on. On a tight or sloping site, this line can dwarf expectations, which is exactly why a site assessment belongs before the budget is fixed.
- Contingency — a reserve for the genuine unknowns every project carries (discussed below).
- Escalation — an allowance for prices rising between today's estimate and the day you actually buy, which matters on anything that will not be procured for many months.
Miss any one of these and your "budget" is just the hard costs with a confident face on it. The forgotten layers do not disappear; they reappear as an overrun.
How much contingency to carry
Contingency is not padding, and it is not money you have failed if you spend. It is the planned reserve that absorbs the surprises you cannot foresee item by item but can predict in aggregate — because some surprise always happens. A common rule of thumb is in the range of 10 to 20 percent of construction cost: the lower end for straightforward new build on a known site, the higher end for renovations, older buildings, and sites with unknowns, because those carry far more hidden conditions.
Two cautions. Contingency covers unforeseen conditions, not scope you simply forgot to estimate — a missing soft-cost line is an estimating error, not a contingency draw. And contingency works only if someone watches it: track what you have spent against the reserve, so you know when it is getting thin and can make deliberate decisions rather than discovering at the end that it ran out weeks ago.
A worked example: building the number in layers
Take a single-storey extension with a base construction (hard-cost) estimate of £120,000 from an area-based figure. Built into a real budget, it looks like this:
- Hard costs (the build itself): £120,000
- Soft costs — design, engineering, permits, surveys, insurance: £18,000
- Site/enabling — demolition, drainage, a longer service run: £9,000
- Subtotal before reserves: £147,000
- Contingency at 15% (renovation-adjacent, some unknowns): £22,050
- Escalation, ~3% on a build procured several months out: £4,410
- Total realistic budget: ~£173,000
The headline number was £120,000; the number you can actually build to is closer to £173,000 — about 44 percent higher, and none of that gap is waste. It is the soft costs, site work, and reserves that the headline simply left out. An owner who funded £120,000 was never under-budget; they were under-informed. This is also why uncontrolled change orders are so corrosive: they spend a contingency that was sized for the unknown, not for optional upgrades.
Why a low estimate is the most expensive number
There is a persistent temptation — sometimes from an over-eager bidder, sometimes from wishful thinking — to chase the lowest figure. Resist it. A lowball estimate does not make a project cheaper; it makes it unfunded. The work still costs what it costs, and the difference surfaces mid-build as change orders, a stalled job waiting on money, or a contractor cutting corners to hit an impossible number. The cheapest bid frequently signals missing scope that returns later as extras.
A realistic estimate, by contrast, lets you decide deliberately: trim scope now while it is cheap, phase the work, or confirm you can fund the real figure before you start. The point of estimating is not to produce a small number — it is to produce a true one, early enough to act on. Because rates, fees, and conditions vary by location and project, confirm the specifics with a quantity surveyor, professional estimator, or your contractor before committing the money.
Frequently asked questions
How accurate is a cost-per-square-foot estimate?
Useful for an early ballpark, but only as good as the rate and only as complete as what the rate includes. Rates vary widely by building type, finish quality, and region, and they typically exclude soft costs, difficult site conditions, and contingency. Use a comparable local rate to get in range, then build the full budget by category rather than treating the area figure as the answer.
What's the difference between hard costs and soft costs?
Hard costs are the physical construction — labour and materials for the building itself. Soft costs are everything around it: design and engineering fees, permits, surveys, inspections, insurance, legal, and project management. Soft costs are commonly a significant fraction of hard costs and are the layer owners most often forget, which is why budgets that looked fine fall apart when the fees come due.
How much contingency should I include?
A reserve in the range of 10 to 20 percent of construction cost is common — the lower end for straightforward new build on a known site, the higher end for renovations, older buildings, and sites with unknowns. Contingency covers genuine unforeseen conditions, not scope you forgot to estimate, and it only works if you track spending against it so you know when it is running low.
Why do construction projects so often go over budget?
Usually because the original number was never a complete estimate — it captured hard costs but omitted soft costs, site work, contingency, or escalation, or it used an optimistic rate. The work then costs what it always would have, and the gap shows up as an "overrun." Estimating in full layers up front prevents most of it; controlling change orders during the build prevents the rest.
Next step
Before you fund anything, build your estimate the way it should be built: start with a realistic local area rate to get in range, then layer in soft costs, site and enabling work, contingency, and escalation to reach a number you can actually build to. Match the method to how much design you have, carry the range honestly, and confirm the local rates and fees with a quantity surveyor or professional estimator. A true estimate, produced early, is the cheapest insurance in the whole project — and always verify the specifics with a qualified professional and your local conditions.
