Cost control
Earned value management (EVM), explained with one example
Earned value answers two questions with three numbers: are we getting the work we paid for, and what will the project cost at the end? Here are the formulas, one worked example and how to read the results.
The three numbers
Everything in earned value comes from three figures measured at the same date:
| PV | Planned value. The budget for the work that should be done by now. |
|---|---|
| EV | Earned value. The budget for the work that is actually done. |
| AC | Actual cost. What that work has cost. |
| BAC | Budget at completion. The whole approved budget. |
The trap is comparing PV with AC, "we planned to spend 400 and spent 420". That says nothing about how much work was done. Earned value puts the work in the middle.
The formulas
| CV = EV − AC | cost variance: negative means over budget |
|---|---|
| SV = EV − PV | schedule variance, in money |
| CPI = EV / AC | value earned per unit of money spent |
| SPI = EV / PV | work done compared with work planned |
| EAC = BAC / CPI | forecast cost if efficiency stays the same |
| EAC = AC + (BAC − EV) | forecast if the rest goes to plan |
| VAC = BAC − EAC | expected overrun or saving at the end |
| TCPI = (BAC − EV) / (BAC − AC) | efficiency needed from now on to finish on budget |
One worked example
A project has a budget of $1,000,000. At the end of month four the plan said $400,000 of work should be done. The team has completed work budgeted at $350,000 and has spent $420,000.
| CV | 350,000 − 420,000 = −70,000 |
|---|---|
| SV | 350,000 − 400,000 = −50,000 |
| CPI | 350,000 / 420,000 = 0.83 |
| SPI | 350,000 / 400,000 = 0.88 |
| EAC | 1,000,000 / 0.833 = 1,200,000 |
| VAC | 1,000,000 − 1,200,000 = −200,000 |
| TCPI | 650,000 / 580,000 = 1.12 |
Read it like this. Every dollar spent has bought 83 cents of planned work. If that continues, the project ends $200,000 over. To finish on the original budget, the team would have to work at 1.12 from now on, a third better than it has managed so far. A sponsor should treat that as unlikely and ask for a recovery plan or a revised budget.
How to read the indices
- Above 1.0 is better than plan, below 1.0 is worse. Most organisations flag anything under 0.95 and escalate under 0.90.
- CPI settles early. Once a project is about a fifth complete, its CPI rarely recovers by much. That is the best-known finding in the field, and the reason early numbers matter.
- SPI is in money, not days. It also drifts back to 1.0 at the end of every project, late or not. For time, use earned schedule.
- Pick the EAC that matches the cause. If the overrun was a one-off, use AC + (BAC − EV). If it reflects how the work is going, use BAC / CPI.
What you need to start
Three things: a budget spread over time (a cost-loaded schedule or a simple monthly plan), an honest measure of progress, and actual costs by period. The second is the hard one. Use rules that cannot be argued with: 0/100 for short tasks, 50/50 for medium ones, and measured quantities for long ones. "90% complete" for three months in a row is the classic failure.
In Critova, the budget and progress come from the schedule, cost entries are typed in or uploaded from a spreadsheet, and each index shows the formula behind it. See cost and earned value.
Common questions
Is earned value only for large government projects?
No. The full ANSI/EIA-748 system is heavy, but the three numbers and two indices work on any project with a budget and a plan.
What is a good CPI?
1.0 or above. Between 0.95 and 1.0 deserves a look, and below 0.90 usually needs a recovery plan or a re-baseline.
Can CPI and SPI disagree?
Yes. A project can be ahead of schedule and over budget, for example when overtime buys speed. Read them together.
Bring one schedule. See your critical path in an hour.
Free during our launch until 31 March 2027.
