Many cost forecasts are built backward from invoices. Spend to date, plus budget remaining, equals the estimate at completion. That arithmetic only works if the work done matches the money spent. On a troubled job, it rarely does.
Spent is not earned
Invoices record what was billed. They do not record what was installed. Front-loaded pay applications, stored materials and change work billed before it is approved all push spending ahead of progress. A forecast built on spending inherits every one of those distortions.
Start with installed quantities
For each major cost account, measure what is actually in place: cubic yards poured, tons erected, linear feet of pipe, rooms finished. Compare that with the quantity budgeted. That ratio is your real percent complete, and it is the basis of earned value.
Then compare cost with progress
- CPI by account. Earned value divided by actual cost, account by account. A low CPI in one trade is a cost problem you can name.
- Productivity trend. Units installed per labor hour, period over period. A falling rate predicts overruns before the invoices show them.
- Remaining work at current performance. Price the remaining quantity at the rate actually achieved, not the rate in the estimate.
Present more than one EAC
Show leadership a range: remaining work at budget, at current CPI and from a bottom-up estimate of what is left. The gap between them is the risk. Explain it in a sentence.
Tie it to the schedule
A later finish costs money: general conditions, escalation, extended equipment. A cost forecast that does not move when the schedule moves is not a forecast. Put both on the same page.
Test the numbers in the Earned Value Pulse, or ask about cost forecasting built from your field progress.
