Building companies rarely fail because the jobs were unprofitable. They fail in the month a profitable job needed more money going out than was coming in, and there was nothing left to bridge it with. Growth makes this worse, not better — the more work you win, the further ahead of your certificates you have to fund.
The gap that does it
Under a JBCC contract the employer pays you within fourteen calendar days of a payment certificate being issued. You must pay your subcontractors within twenty-one calendar days of that same certificate. Those two clocks start together, which sounds comfortable until a certificate is issued late, or paid late, and the twenty-one days runs regardless.
Your obligation to your subcontractors is tied to the certificate, not to when you were actually paid. A client who pays on day thirty has not extended your deadline; he has taken nine days out of your bank account.
On one job that is survivable. Across four jobs with staggered certificate dates it is a genuine forecasting problem, and it is not one you can hold in your head.
What a useful forecast contains
- Money in, by certificate date plus the payment period — not by the month the work was done. Work done in March that is certified on 5 April and paid fourteen days later is April money.
- Retention held out of each certificate, and the release dates sitting somewhere in the future. See the retention register.
- Money out, split by how fixed it is. Wages happen weekly whatever else does. Material orders can sometimes move. Subcontractor payments have a contractual date.
- Preliminaries running whether the job moves or not — site establishment, supervision, plant hire. The cost of three rained-out weeks is mostly here.
- The closing balance for each month, and a line showing the worst point within the month. A month that ends positive can still have been negative on the 12th, and the 12th is when the wages ran.
Three things that reliably break a forecast
- Forecasting income by work done. The most common error by a distance. Work done is not money; a certificate is.
- Forgetting the deposit is not income. It is money you have already spent on materials that have not arrived yet.
- Leaving variations out. An approved variation moves both the contract sum and the cost. A forecast that only knows the original sum will be wrong in a direction you do not expect.
How far ahead to look
Three months is enough to be useful and short enough to be kept current. Twelve-month forecasts on a residential contractor’s desk are almost always fiction after week six, because the work that fills months seven to twelve has not been won yet.
The number worth watching is not the twelve-month total. It is the lowest balance in the next ninety days, and whether it is above zero.