How to read budget vs actual on a construction project
A variance report is only useful if you know which numbers lie. A working guide to committed cost, earned value and the three variances worth acting on.
Every cost report says the same thing at first glance: we have spent this much of that much. That single comparison is the most misleading number in construction finance, because it says nothing about how much you have actually built.
Spending 60 percent is not news
A project 60 percent through its budget is doing brilliantly if it is 75 percent built and in serious trouble if it is 40 percent built. The budget line alone cannot tell you which, so pair it with earned value: the budgeted value of the work actually completed.
| Spent | Earned | Reading |
|---|---|---|
| 60% | 75% | Ahead on cost. Check the quality of the progress claim before celebrating. |
| 60% | 60% | On plan. Watch the trend, not the point. |
| 60% | 40% | Overrunning. Find the line, do not average across the project. |
Committed beats invoiced
Most cost reports show what has been invoiced. That is a lagging number by anything from a fortnight to a quarter. Committed cost includes what you have already ordered and cannot cancel: purchase orders raised, subcontracts signed, labour rostered.
A line can be comfortably under budget on invoices and already overspent on commitments. If you only watch invoices, you find out when the invoice arrives, which is after the concrete has been poured.
Practical rule: control against committed cost, report profit against invoiced cost. They answer different questions and both belong in the system.
The three variances worth naming
Rate variance
You bought the right quantity at the wrong price. Steel moved, the supplier changed terms, or the estimate used last year rate. The fix is commercial: renegotiate, resequence purchases, or claim it if the contract allows fluctuation.
Quantity variance
You bought the right price at the wrong quantity. Either the take-off was wrong or the site consumed more than it should have. The fix is operational, and the two causes need different conversations: one with estimating, one with the site.
Timing variance
Nothing is actually wrong. The cost landed in a different period than the plan expected. This is the variance that generates the most panic and needs the least action, which is exactly why it should be labelled rather than debated.
If your report cannot separate these three, every variance meeting starts by re-deriving what happened. That is the meeting, gone.
A monthly routine that takes an hour
- 1
Close the site log
Attendance, materials and progress up to the cut-off date. No new entries after this point for the period.
- 2
Pull committed cost
Open purchase orders and signed subcontract values, not just what has been billed to you.
- 3
Sort variance by value, not percentage
A 40 percent overrun on a small line is noise. A 4 percent overrun on the biggest line is your month.
- 4
Label the top five
Rate, quantity or timing, one word each. Only the first two get an action.
- 5
Forecast to completion
Take the current rate of consumption on the overrunning lines and extend it. This number, not the current variance, is the one that decides whether you intervene.
The one report to keep
If you keep a single page, make it: budgeted value, earned value, committed cost and forecast to completion, per BOQ section, sorted by variance value. Everything else is a drill-down from those four columns.
Questions this raises
How often should we run variance?+
Monthly for the formal report, weekly for the top three lines on any project already showing overrun. Weekly on everything is a job nobody sustains past month two.
Do we need earned value on small projects?+
You need progress recorded per line, which is all earned value is. On a small job that is a few minutes of logging per week.