Sensitivity: the three inputs worth testing

A feasibility produces one number and invites you to treat it as the answer. It is really the answer to one particular set of assumptions, all of which are forecasts.

The useful question is not what the margin is, it is which assumption has to be wrong, and by how much, before the project stops working.

Three inputs carry nearly all of the risk

Every feasibility has dozens of inputs and almost all of them are noise. Three of them move the answer.

1. Revenue. The achieved sale price per unit, which is a forecast about a market twenty months away.
2. Build cost. Usually the largest single line, and the one most exposed to a trade shortage or a material price.
3. Timeline. Every month of delay carries interest, holding costs and a later sale into a different market.

Test those three properly and you have tested the project. Fine tuning the legal budget is a way of feeling thorough without learning anything.

Move one at a time, then move them together

Take each input and run it at ten per cent worse than your base case, one at a time, keeping the others still. The three answers tell you which input the project is most exposed to, and that ranking is often a surprise.

Then run all three at once, because that is what a bad market actually looks like. Prices soften, trades get expensive, and approvals slow down, and they tend to happen together rather than politely one at a time.

The combined case is the one that changes decisions

Single input tests tend to leave a project looking robust, because a ten per cent move in one line rarely breaks anything. The combined case is where a project either holds or does not.

A scheme that survives all three moving against it is a scheme you can commit to. One that only works while everything holds is not necessarily a bad project, but it is a project that needs a contingency and a plan for the day it goes wrong, and that decision should be made now rather than later.

Work backwards from the margin you need

Sensitivity tells you how fragile a price is. The other direction is more useful when you are still buying: name the margin you need, and solve for the land price that delivers it.

In Diract you type the deal's numbers and the margin you need, and the solver works back to the highest land price that still gives that margin. That figure is your walk away number, and having it before the negotiation is worth more than any amount of analysis during one.

Write down the assumptions beside the answer

A feasibility six months old is unreadable unless the assumptions travelled with it. What sale price, what build rate, what interest rate, what programme, and who provided each.

That record is what lets you update rather than rebuild when the quantity surveyor's figure lands. It also shows which assumptions came from evidence and which came from the room.

Then check the assumptions against the actuals

The feasibility stops being a forecast once the project starts and becomes a budget you are measured against. Holding the actual costs against the original lines is what tells you whether your assumptions are reliable, on this project and on the next one.

Budget against actual on a development covers which columns show an overrun while there is still time to act on it.

What to do first

• Test revenue, build cost and timeline, and leave the small lines alone
• Run each at ten per cent worse, one at a time, to rank the exposure
• Run all three together, because that is what a downturn does
• Solve backwards for the land price your required margin allows
• Record every assumption and its source beside the answer
• Hold the actuals against those lines once the project starts