Two houses of the same floor area, built by the same builder in the same year, can differ in cost by an amount that makes people uncomfortable.
Neither owner did anything wrong. They just made different decisions at different points — and some of those decisions carried far more weight than they looked like at the time.
Here's what actually moves the number.
The drivers you mostly can't control
The site. Slope is the big one — a fall across the block means retaining, deeper footings, more excavation and often crane or truck access problems. Then soil classification, which is discovered by testing rather than chosen. Then access: a narrow frontage or a tight corner adds cost to every single delivery for the length of the build.
Services. Where the sewer, water, power and stormwater actually sit, and what has to happen if they're not where you'd want them.
The planning path. Overlays, neighbourhood character controls, heritage, vegetation protection. These don't only add report costs — they add time, and time has a cost of its own.
Market conditions. Trade availability and material pricing at the moment you go to tender. Nobody controls this. It's why a cost estimate has a date on it.
You can't change these. You can find out about them early, which is a different and very valuable thing.
The drivers you can control
Floor area. Still the single most powerful lever. Every additional square metre carries structure, roof, finishes and services. The most effective cost reduction in residential architecture is almost always a well-designed smaller plan rather than a cheaper large one.
Wet areas. Bathrooms, kitchens and laundries cost several times what a bedroom costs per square metre. The number of them, and how far apart they are, matters.
Shape. A house with lots of corners, steps and projections has more external wall for the same internal floor area — more structure, more waterproofing, more junctions, more labour. Articulation is a design choice with a price on it.
Structural ambition. Long spans, cantilevers, large unsupported openings and curves are all achievable and all cost money. Sometimes they're worth it. They should be a deliberate purchase, not an accident of the sketch.
Level changes and roof complexity. Both quietly expensive, both easy to add without noticing.
Glazing. Area, size of individual panes, and performance requirements. Large single panes escalate quickly.
Finishes and joinery. The category everyone assumes is the main driver. It matters — but by the time you're choosing tiles, most of the cost is already locked.
The point most people miss
The decisions that move the number most are made in the first stage of the project, not the last.
By the time you're selecting finishes, the floor area, the shape, the structure and the site strategy are all settled. That's most of the cost, decided before anything looked like a real house.
Which is also why documentation depth matters. A well-resolved set of drawings gets you a tender you can trust. A thin set gets you a low number followed by variations — and variations are where budgets are lost, not in the original contract sum.
What to do next
A feasibility study exists for exactly this. Three sessions and a cost estimate, before the design is committed — so the expensive decisions get made with information rather than optimism.