Because the permit answers a different question than the feasibility does. A permit says the council will allow this. A feasibility says these numbers work. Neither one says that anyone will pay for the result, that it can be built for the figure in the model, or that the money to build it exists.
Owners usually put the question this way: Why do commercial feasibilities fail after DA approval? Victoria issues planning permits rather than development applications, so the label changes at the border, but the failure pattern does not. In the deals we screen, what gives way after an approval is the evidence behind the numbers, not the planning.
What the software is genuinely good at
Feasibility software earns its place. Give it a site, a set of controls and a list of assumptions. It will resolve the envelope, the staging, the yield and the land value faster than a spreadsheet you build yourself, and it will hold the sums together when an input changes. On a complex site it can run through combinations you would not have the time to test by hand.
What it does not do is audit its own inputs. A land value is a sum worked on the numbers you gave it. Move the rent assumption, or the build rate, and the answer moves with it — often by more than the person who typed it expected. The tool is not wrong when that happens. It answered the question asked: what fits, given these figures. Whether the figures are evidence or hope is a different question, and not one software is built to ask.
What a permit is, and what it is not
A planning permit is a consent. It records that a responsible authority will allow a described use or development on that land, usually with conditions attached. It does not say the project stacks up, that the market wants it, or that the cost plan is sound. Council does not assess your margin.
Conditions are where the gap opens. Where a permit arrives with conditions attached, the "before works" conditions must be satisfied and the plans endorsed before construction can begin. A condition that adds a parking bay, upgrades a crossover or requires acoustic treatment is a planning outcome and a cost event at the same time — and it lands after the feasibility was signed. Our guide to the Victorian permit process walks that sequence end to end.
Time is the other quiet cost. In Victoria a development permit generally defaults to two years to start works and four to finish, unless the permit says otherwise. Every month spent resolving conditions is a month of holding cost, and where the model assumed a start date the conditions have since pushed back, the carry it never budgeted for is already running against you.
The four things an approval does not carry
Rent evidence. A cost plan. Pre-commitment. Finance.
We screened a deal that held a granted, current planning approval, and the screen still would not clear it. Not because the planning was wrong — the planning was the part that had been done well. The approval carried none of those four, and none had been assembled since. The planning questions had been answered. The money questions had not been asked.
That is the shape the failure takes in the cases we screen. The permit becomes the proof, and the proof is only ever about permission — which is how a project can pass every planning test put to it and still be undeliverable on the day somebody finally prices the thing.
Where judgment enters
The fix is not a better model. It is grading what the model is standing on. Every input behind a decision is either a primary document, something credible but second-hand, or an assumption. An executed lease is the first. A market rent someone quoted you is the second. A build rate borrowed from a comparable project is the third — and an assumption is a task, not a fact. It is either upgraded to a document or it breaks the deal.
Run that grading over a feasibility after approval and the weak points show up fast. They also cluster. The cost plan rests on rates nobody has quoted. The income rests on a rent nobody has signed. Both were fair when the model was built, and neither was revisited once the permit arrived. The five decision gates exist to make that check routine rather than occasional, so what the model is standing on gets tested at every point where the answer could reasonably have moved.
Re-run it at the gate, not at the start
A feasibility is not a document you produce once. It is a position you hold, and it should be re-tested at the points where the project's facts actually change: when the permit and its conditions land, when the design is documented, and when tenders come back. Those are the moments when assumptions become prices. A model still running on pre-approval inputs at contract award is not a feasibility any more; it is a memory of one.
An Axis Quick Scan tests a single deal against its archetype, from A$49.95, and our team confirms every report before it ships.
This article is independent decision support — not financial, legal, tax or planning advice. Cost plans belong with a quantity surveyor, valuations with a certified practising valuer, and funding through licensed partners.
Verdict: a permit proves you are allowed to build it; only graded evidence tells you whether you should — and that check belongs after the approval, not before it.