How to run a residential development feasibility study

Before the money: the planning checks that come first

A feasibility study that starts with the spreadsheet is a study of the wrong thing. Before any dollar goes into a model, the site has to pass the planning gates, because a strong margin on a scheme the zone will not carry is a strong margin on nothing.

1. Permissible use. Can the thing you are modelling legally exist in this zone at all? A duplex margin means nothing in a zone that caps the site at one dwelling.
2. The envelope. Height limit, setbacks, site coverage. These decide the buildable volume, and the buildable volume decides yield before cost enters the picture.
3. The overlays. Flood, heritage, bushfire, vegetation. Any one of them can reshape the design, add conditions and cost, or take options off the table entirely.
4. Services and access. What sits under and beside the block. Easements, mains and stormwater paths shrink the envelope in ways no zoning summary shows.
5. The minimums. If the scheme involves subdivision: minimum lot size and minimum frontage. Miss either by a metre and the two-lot model becomes a one-lot reality.

Two of the five have guides of their own: flood overlays and heritage overlays. Which approval pathway the result lands in is covered by CDC vs DA.

The 7-step method

Step 1: Site and context analysis

Before any numbers, you need to understand the physical and legal reality of the site. Walk it (or study online mapping tools). Note the frontage, depth, fall, orientation, and access. A steep fall means retaining and expensive footings. A narrow frontage constrains driveway access to potential rear dwellings. Established trees, easements, and a single crossover can each quietly kill a yield assumption.

Download the title and check for easements, encumbrances, and covenants. An easement running down the middle of a block you were planning to build across is the kind of thing that turns a “six townhouses” site into a “four townhouses, if you’re lucky” site.

The output of Step 1 is a plain description of what the land is: dimensions, slope, access points, servicing, and any physical constraint that will shape what can go on it.

Step 2: Planning controls and what they permit

This is where most manual feasibility studies slow down, and if you’re new to development, it can be an overwhelming process finding the right information across government websites, maps and legislation.

Every site sits under a zone (what land can be used for) and usually one or more overlays (heritage, flooding, bushfire, significant landscape, tree protection). The zone and its associated controls set the envelope: how many dwellings, how tall, how far from each boundary, how much of the site you can cover, how much private open space and car parking each dwelling needs. You’re reading for the numbers that cap yield:

• Density / minimum lot size : how many dwellings the site can legally support
• Height : how many storeys
• Setbacks: front, side, rear
• Site coverage and open space: how much of the block can be built on
• Car parking : spaces per dwelling, plus visitor parking

The output of Step 2 is a defensible statement of what the planning framework permits.

(This is exactly the step Struqt is built to compress. Reading zones, overlays and constraints for an address by hand takes hours of navigating planning schemes and mapping portals. But you still have to know what the numbers mean, which is what the rest of this method is about.)

Step 3: Yield and scheme options

Now you turn the data into buildings. Yield is the number and type of dwellings the site can hold. Sketch two or three scheme options (a lower-density, lower-risk version and a more ambitious one) and test each against the Step 2 controls. Does it fit the setbacks? Does the parking work? Does each dwelling get its required open space?

Yield is the single most sensitive lever in the whole study. One extra saleable dwelling can swing a project from marginal to strong, because the land and many fixed costs are shared across more revenue. Equally, an over-optimistic yield that won’t survive the planning assessment is how developers talk themselves into bad sites.

Be honest here. Test the scheme you can actually get approved, not the one you wish for. The output is a realistic dwelling schedule: how many of what, at what size (measured in Gross Floor Area, GFA, and saleable area). In NSW, GFA is the figure the zone's floor space ratio caps directly.

Step 4: Cost estimation

With a scheme defined, you can estimate costs. Total Development Cost (TDC) is far more than construction. The elements that matter:

Construction. For a feasibility-stage estimate you use a cost-per-square-metre rate applied to the gross floor area (GFA). This is a ±15–20% approximation, so treat it as a range, not fact. As a public benchmark, current guides put multi-unit townhouse construction in the rough order of $2,200–$3,700/m² in Melbourne, and apartments $2,800–$4,500/m², with figures varying by city, storeys and specification. Cross-check a $/m² figure against a per-dwelling figure. RLB quotes Sydney townhouses at $425,000 to $775,000 per townhouse at late-2025 rates, excluding car parking and site works, and when the two disagree badly, your area assumption is usually the culprit.

Professional fees. The consultant team: architects, engineers, planner, surveyor, certifier, quantity surveyor, project manager, and often more. As a rule of thumb these commonly run from around 8 to 15 per cent of construction cost across the full team, and they’re easy to underestimate because they arrive as a dozen separate invoices.

Statutory costs and contributions. Council application fees, and infrastructure/development contributions that vary enormously by location, potentially up to around $40,000 per unit in some areas. Check the specific council; this line surprises people.

Finance costs. Interest on land and construction debt through the build period.

Contingency. Typically 5–10% of construction cost depending on design stage and complexity. Don’t skimp on it: this is what protects your margin. Issues arise during construction, from hitting rock during footing excavation to contaminated soil discovery, material costs rising before contract signing.

Selling costs. Agent commission, marketing, and legal costs. Agent commission commonly runs in the order of 1.5–2.5% of sales. Acquisition costs. Stamp duty and legals on the land purchase itself.

The output of Step 4 is the total cost of everything except the land, because the land price is the answer you want to arrive at.

Step 5: Revenue (Gross Realisation Value)

Gross Realisation Value (GRV), sometimes called GDV, is the total you expect the finished dwellings to sell for. Build it bottom-up from comparable sales — recent settled prices for similar new dwellings in the same area, not asking prices and not optimism. Then subtract selling costs to reach net realisation.

A discipline worth adopting: banks routinely haircut a developer’s GRV assumptions by 5–10%; if your feasibility only works at your expected GRV, the margin probably isn’t sufficient. Model the conservative number. The output of Step 5 is a defensible end value grounded in real comparable evidence.

Step 6: Residual land value, the verdict

This is the verdict the whole study builds toward. The logic is simple subtraction:

RLV = GRV − (all development costs) − (target profit)

Take the end value, remove every cost from Step 4, remove the profit you require for taking the risk, and what remains is the most you can pay for the land. Two conventions you need to set:

Target margin. The industry standard is a residual project margin of roughly 15 to 20 per cent, and major lenders typically require a minimum in that range for development finance approval. Note whether you’re expressing margin on cost or on revenue , and they’re different numbers. A 20% margin on cost is roughly equivalent to 16.67% profit on GDV, and mixing the two up is a classic way to make a bad deal look fine.

GST. In Australia the GST margin scheme can move land value by a large amount depending on how you acquire the site, so flag it and get advice. It is one of the most significant and commonly misunderstood factors in Australian development feasibility.

Now compare your RLV to the asking price. If the land is offered at or below your RLV, the deal works at your target margin. If the asking price is above it, either there are costs to be cut, another dwelling to add, or it’s not your site. The output of Step 6 is a single number: your maximum bid.

Step 7: Sensitivity and risk

One RLV number is a snapshot. A feasibility study earns its keep by showing how fragile that number is. Flex your two most sensitive inputs, construction cost and end sale value , by ±10% and watch what happens to the margin and the RLV. A robust project holds its margin under a 10% cost blowout or a 10% price softening. A fragile one only works if every unit sells at the top of the range and nothing goes wrong on site, which is another way of saying it doesn’t work.

Also stress the timeline: a project that runs six months long carries six more months of interest. Ask which single assumption, if wrong, breaks the deal. That is the thing to verify before you bid. The output of Step 7 is an honest picture of where the risk lives, and how much room the deal has to absorb bad news.

A worked example: three townhouses on a 700m² site

Numbers make the method concrete, so here’s a single site carried through all seven steps. Every figure below is an illustrative public benchmark, not a Struqt output or a real project . The point is to show how the arithmetic flows, not to price your site. Your own comparables and a QS cost plan will affect every element.

The site. A 700m² block, regular shape, gentle fall, single existing dwelling, no easements. Zoning and controls support three double-storey townhouses with compliant setbacks, parking and open space. That’s our Step 1–3 result: yield of 3 townhouses, roughly 160m² GFA each (480m² total).

Steps 5–6: revenue and the residual. Three townhouses at an illustrative $1,150,000 each (from recent settled comparables, not asking prices), with the target margin set at 18% on cost:

So on these illustrative numbers, the most you could pay for the land and still hit an 18% margin is roughly $915,000. If the agent’s asking $1.05M, the deal doesn’t work as drawn: you’d need a fourth dwelling, lower build cost, or higher end values. If it’s listed at $850k, there’s room.

Step 7: sensitivity. Flex the two big levers ±10%:

Construction +10% (→ ~$1.58M). RLV falls to roughly $760k. A cost blowout eats ~$155k of land budget.

Sale price −10% (→ $1.035M each). GRV drops $345k; RLV falls to around $600k. Softer sales hurt more than cost overruns here, because they hit revenue and the profit calculation.

The lesson the table teaches at a glance: this deal is more exposed to a price softening than to a build overrun, so the sale comparables in Step 5 are the assumption to pressure-test hardest before bidding.

Where manual feasibility studies go wrong

Recurring failure points are predictable:

Optimistic yield. Assuming a dwelling count the planning controls won't actually support.

Understating the cost base. Margin on cost is profit divided by total development cost, so understate the cost base and you overstate the margin. That is exactly the error that makes a marginal deal look safe. The usual culprits are missing contribution lines, thin contingency, and forgotten professional fees.

Treating a feasibility estimate as a firm budget. A ±15–20% feasibility number is a filter, not a commitment. The gap between a preliminary estimate and a detailed one is one of the most common causes of development failure.

Stale comparables. Using listing prices or year-old settlements for GRV. The end value is half the equation; get it from recent settled sales or don't trust the result.

Skipping sensitivity. A single-scenario study hides its own fragility. If you didn't flex the inputs, you don't yet know whether the deal is robust or just lucky on paper.

The one-page feasibility checklist

Done properly, a feasibility study is a decision tool, not a spreadsheet ritual. It tells you whether to walk the site, roughly what to bid, and where the deal will break if it breaks. The honest catch: done by hand, the first pass takes the better part of a day per site , and the slowest, most error-prone step is reading the planning controls in Step 2. Run the checklist below against any site before you make an offer.

Screening sites like this every week: for yourself or for clients? See how developers, architects and agents use Struqt to run the planning-controls pass across a whole shortlist.

Frequently asked questions